Translating “net financial assets”

Steve Roth at Asymptosis offers a remarkable, detailed discussion of Modern Monetary Theory’s notion of “private sector surplus” with an emphasis on aggregate accounting. Roth’s core point is well taken: “Private sector surplus” (equivalently the increase in “private sector net financial assets”) should not be conflated with the economic saving of households. As Roth points out, household sector saving is the difference between household sector income and household sector (noninvestment) expenditure. “Private sector surplus” is likely to increase household sector income, and so in that sense it forms a component of household sector saving, but it is quantitatively small relative to total household income — especially, as Roth emphasizes, if you use comprehensive measures of income that include capital gains and losses. [1]

Roth is right about all this. But I think he is talking past MMT economists a bit. Roth invites us to think about comprehensive saving by households. But that’s very far from what MMT’s baseline sectoral balances decomposition claims to capture. Instead “net financial assets” capture only the financial position of the aggregated (domestic) private sector, including both households and businesses. MMT enthusiasts sometimes mix these things up, and when that happens it should be called out. But this confusion has been called out a lot over the years, and I think for the most part MMT economists have become pretty precise in expressing themselves. There is a great deal of value in the MMT decomposition, not as a measure of household saving, but of something else entirely. Let’s try to understand it.

I’m going to steal from my own, old post, a derivation of the MMT-standard decomposition (usually attributed originally to Wynne Godley). We start with a tautology:

Every financial asset is also some entity’s liability. The sum of all financial positions is by definition zero. So we can write:

NET_WORLD_FINANCIAL_POSITION = 0 [0]

Suppose that, quite arbitrarily, we divide the world into a “foreign” and a “domestic” sector. Then we have:

NET_FOREIGN_FINANCIAL_POSITION + NET_DOMESTIC_FINANCIAL_POSITION = NET_WORLD_FINANCIAL_POSITION = 0 [1]
NET_FOREIGN_FINANCIAL_POSITION + NET_DOMESTIC_FINANCIAL_POSITION = 0 [2]

Suppose that, again arbitrarily, we decompose the domestic economy into a public and private sector:

NET_PRIVATE_DOMESTIC_FINANCIAL_POSITION + NET_PUBLIC_DOMESTIC_FINANCIAL_POSITION = NET_DOMESTIC_FINANCIAL_POSITION [3]

Substituting into our previous expression, we get

NET_FOREIGN_FINANCIAL_POSITION + NET_PRIVATE_DOMESTIC_FINANCIAL_POSITION + NET_PUBLIC_DOMESTIC_FINANCIAL_POSITION = 0 [4]

We can also write this in terms of changes or flows. Since the sum above must always be zero, it must be true that any changes in one sector are balanced by changes in another:

ΔNET_FOREIGN_FINANCIAL_POSITION + ΔNET_PRIVATE_DOMESTIC_FINANCIAL_POSITION + ΔNET_PUBLIC_DOMESTIC_FINANCIAL_POSITION = 0 [5]

Two of the flows in the equation above have conventional names, so we can rewrite:

CURRENT_ACCOUNT_DEFICIT + ΔNET_PRIVATE_DOMESTIC_FINANCIAL_POSITION + CONSOLIDATED_GOVERNMENT_SURPLUS = 0 [6]

Rearranging…

ΔNET_PRIVATE_DOMESTIC_FINANCIAL_POSITION = -CURRENT_ACCOUNT_DEFICIT + -CONSOLIDATED_GOVERNMENT_SURPLUS [7]
ΔNET_PRIVATE_DOMESTIC_FINANCIAL_POSITION = CURRENT_ACCOUNT_SURPLUS + CONSOLIDATED_GOVERNMENT_DEFICIT [8]

The highlighted Equation 8 is where the action is. NET_PRIVATE_DOMESTIC_FINANCIAL_POSITION is often described as net financial assets of the domestic private sector. MMTers (domestic) “private sector surplus” is precisely ΔNET_PRIVATE_DOMESTIC_FINANCIAL_POSITION.

The crucial thing to understand is what the net means in net financial assets. It is precisely financial savings net of domestic real investment by the private sector. It is the farthest possible thing from a comprehensive measure of household savings. It is private sector savings excluding the vast preponderance of household savings, which is backed by private sector assets (whether owned by households directly or owned by businesses who then issue financial claims to households). “Ordinary” private sector savings either doesn’t show up as financial assets at all (a home without a mortgage is just a real asset owned by a family, like a television or a baseball card), or else they “net out” when we aggregate, because one private sector entity’s asset is precisely extinguished by another private sector entity’s liability. If a household is “long” a share of stock, a firm is “short” that same position, and owes the household whatever that claim represents. The aggregate financial position of the private sector combines the financial positions of businesses and households, so the financial claims of households against firms are matched by mirror image liabilities of firms to households. They annihilate one another like matter and antimatter.

So why do we care about this odd sliver of savings? Why do MMT economists make it so central to their analysis? Private sector net financial assets are “special” precisely because they are not backed by domestic real assets, but instead by promises that are credibly independent of domestic real asset values, especially promises of states. Saving that takes the form of real stuff, whether that stuff is directly held or hidden behind financial claims, is inherently risky. House prices fall. If you own a factory, or shares in a firm that owns a factory, the factory can burn down. Even if you hold a diversified stock portfolio, you will find it subject to wild swings in value. If you own private sector debt, you expose yourself to credit risk. If you own a diversified portfolio of domestic stocks and bonds, your own circumstances and that of your investment portfolio will be correlated in an unpleasant way. The times when you lose your job and need to draw on savings are likely to be the same times when stocks have crashed and people are defaulting on their debts. People desperately covet assets that are divorced from the risks of the domestic real economy. And that is precisely what “net financial assets” are.

Net financial assets are special, because they serve insurance functions that assets produced by the domestic private sector simply cannot provide. When households are risk-averse, they covet these assets especially. For firms, these assets offer protection against insolvency risk that real assets, whose values both fluctuate idiosyncratically and covary with the real economy, cannot provide. MMT economists often suggest that if the public sector fails to accommodate the private sector’s appetite for net financial assets, recession and financial instability will result. That makes sense. It’s conventional, if a bit vapid, to describe recessions as times when “animal spirits” are low, when people are risk averse. But what matters is not the courage in people’s hearts (or lack thereof). What matters is how people behave. If people’s behavior is counterproductively risk averse, you can encourage greater risk-taking by offering insurance. That’s precisely what injections of “net financial assets” into an economy provide. If firms are teetering on the brink of bankruptcy, you can flood the economy with safe assets they can use to shore up their balance sheets to reduce their risk of default. That’s precisely how the United States saved its banks in 2008 (for better or for worse). The headline bailouts and TARPs and accounting forbearance were all expedients to keep those firms alive until a flood of assets immune to correlated private sector collapse could find their way onto bank balance sheets (with the help of opaque subsidies). Those special assets are “net financial assets”.

“Net financial assets” are a heterogeneous category. They include both claims against the domestic state and claims on foreign public and private sectors. A claim on a foreign firm in foreign currency does not provide the same insurance as claim against the domestic government in domestic currency. Nevertheless, claims on the foreign sector do provide insurance against domestic shocks that do not impair the foreign counterparty. And note that contrary to naive financial theory, which predicts developed economies will net-accumulate claims on emerging economies to invest in their growth, in practice emerging economies tend to net-accumulate claims on developed economies. The insurance function of safer foreign assets outweighs the investment function of accepting foreign capital (or at least it has since the Asian Financial Crisis). For firms and households in an emerging economy, foreign claims and claims on government are both useful insurance. In developed as well as emerging economies, negative positions with respect to foreign creditors increase the domestic private sector’s exposure to risk as surely as indebtedness to the state would, assuming debt contracts are uniformly enforced.

All this terminology — private sector surplus, net financial assets, etc. — is associated with heterodox, lefty MMT, but it maps very nicely to discussion of “safe asset shortages” in the mainstream financial press or Gary Gorton’s schtick on the importance of “informationally insensitive” assets. The main difference has to do with whether we can or ought to rely upon the domestic private sector to produce these kinds of assets. The MMT analysis, by construction, excludes private sector “triple-A” assets, where people like Gorton emphasize a role of private sector in producing assets that might provide this sort of insurance. The MMTers have it right. The domestic private sector simply cannot produce assets that provide insurance against systematic risks of the domestic economy without the help of the state. (Gorton tacitly recognizes this when he suggests the state should supervise and guarantee assets produced by shadow banks like it insures bank deposits. No thank you.)

The insurance function of “net financial assets” is not unambiguously a good thing. Net financial assets are special precisely because they provide insurance against systematic risk. When net financial assets are claims on foreign debtors, they are not so problematic, they just represent a form of diversification that can insure against domestically (but not globally) systematic shocks. Claims against the domestic state, however, offer safety to their holders in a manner that can be quite dangerous to the rest of us. “Insurance” against a truly systematic shock is necessarily a zero-sum game. If we are all collectively poorer, the only way the state can make some claimants whole is by shifting their share of the aggregate loss to people who don’t hold the government’s promises. We’ve experienced this very painfully over the past decade, as both the European and American policymakers refused to accept any risk of inflation (thereby prioritizing the value of past promises). Policymakers chose to make absolutely sure that holders of state assets would be made whole in real-terms, and imposed severe costs on debtors and the marginally employed to do so. (I think policymakers overshot the inherent zero-sum-ness of providing insurance during a systematic shock and have played a sharply negative-sum game.) It would be better, I think, if states downgraded the insurance they provide by weakening the promise they make to asset holders from price stability to an NGDP path target. And I worry much more than I think most MMT economists do about the unjust distribution of risk-bearing that might accompany a large stock of net financial assets very unequally distributed. (Unusually, I’m with Greg Mankiw on this one.) I think the economy includes people who are already overinsured by their stock of net financial assets, and those people tend disproportionately to accumulate new issues. So we should think more about how we can accommodate private sector entities’ need for some degree of insurance by redistributing existing net financial assets rather than creating new ones.

This sentence is a pithy conclusion.


[1] Should you? Or should you use a NIPA-style accounting that “looks through” capital gains? That’s a complicated question. Looking through capital gains entirely is clearly unsuitable, because household capital gains include revaluations of shares due to e.g. retained earnings, which represent increases in the quantity of real assets that households’ shares lay claim upon. This kind of capital gain is economic saving. But what about price appreciation of the existing housing stock? On the one hand, this is certainly perceived by individual households as real wealth and a form of savings. On the other hand, housing price increases also represent a kind of liability on the part of households and households-to-be who are not yet homeowners. If our object is to study distributional questions, differences between households, capital gains of this sort should obviously take center stage. They are real wealth to the households who enjoy them, and often represent costs in some form to households who don’t. But if we are aggregating, it’s not as clear that mere repricings of existing assets should be included in household saving. Unfortunately, it’s almost impossible in practice to disentangle gains due to real growth in the assets backing claims from repricings of existing assets. Consider our prototypical example of “mere repricing”, price appreciation of an existing home. If a home remains entirely unchanged in an unchanged neighborhood in an unchanged city, then that is mere repricing. But perfect stasis is impossible outside of a thought experiment. If an “existing home” is renovated, and its price appreciates, how much of the price appreciation is “mere repricing” and how much of it reflects the change in real assets? If the neighborhood surrounding a largely unchanged home improves dramatically, then the house is a more efficiently deployed real asset, and a change in price may reflect new real value, which does constitute a form of economic saving — a claim on real growth — rather than mere repricing. Similarly, while stock appreciation can reflect an increase in the quantity or real-economic usefulness of the assets backing shares, it can also result from a mere revaluation of largely unchanged firms. When the Fed eases, stock prices rise, but the real assets that back them are not meaningfully improved. I think on the whole Roth’s choice to include capital gains in his analysis of aggregate household saving is right, much better than the alternative choice of ignoring it. But neither choice can be “correct”.

Update History:

  • 24-Jul-2015, 1:15 p.m. PDT: “Note that the The insurance function of “net financial assets”…”
 
 

408 Responses to “Translating “net financial assets””

  1. Steve Roth writes:

    As always, so much better than mine. Thanks.

    But I’m still struggling, for two reasons:

    1. This doesn’t relate private surplus to net worth, which measure is sort of my anchor and touchstone for understanding this stuff in balance-sheet terms.

    What’s the accounting relationship between Private Surplus and change in HH Net Worth? (For simplicity assume a closed or world economy, with gov but no ROW.)

    Starting HH Net Worth + Private Surplus +/- ?? = Ending HH Net Worth

    What specific labeled measures, on which tables, in which statement of accounts, should I add or subtract there?

    This assuming that:

    HH Net Worth = Private Net Worth (Is this safe?)

    and:

    Private Surplus = Gov Deficit Spending

    I totally get the latter. It’s straightforward. But what else does Private Surplus equal (in/on the private-sector income statement/balance sheet)? Absent that understanding, it remains a tautologically-defined term for me. Help!

    This is a big way that I try to understand things, what these accounting terms/measures mean (relative to each other), actually adding them up in spreadsheets to see what equals what. I’m seeking my identity. ;-)

    2. What does Net Financial Assets mean, as a stock? Defined in terms of labeled balance-sheet entries in the FOFAs or the IMAs. I totally understand Net Acquisition of Financial Assets, the flow, but the stock confuses me. I assume it has to do with assigning certain (all?) balance-sheet liabilities against (only) financial assets? Help some more?

    Thanks…

  2. Steve Roth writes:

    By the way, I’ve spent quite a bit of time trying to figure out #2 (well, #1 also…).

    I googled it again this morning — “net financial assets=” — and my second hit is…this post! The first hit is to a Mike Norman post pointing to a rogueeconomistrants post…

    http://rogueeconomistrants.blogspot.com/2011/10/what-is-net-financial-asset.html

    …which tells us this:

    “What Tom [Hickey] means with ‘net financial assets’ is net private sector savings,”

    Oh god. “Savings” (the plural, the stock) is completely nonexistent hence undefined in the national accounts, though economists (and wannabes like me [but not me]) toss it around constantly.

    Which leads me to point once again to Nick Rowe’s great post headline, “Why ‘saving’ should be abolished”:

    http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/01/why-saving-should-be-banned.html

  3. Ramanan writes:

    SR,

    “Starting HH Net Worth + Private Surplus +/- ?? = Ending HH Net Worth”

    Saving is not just private surplus. In fact “private surplus” is the financial balance of the whole private sector of which the household sector is just one.

    So instead of private surplus, use household surplus. In modern national accounting terminology it’s called “net lending”.

    But that’s not saving either. Saving (disposable income less consumption) is identically equal to gross capital formation (investment) plus net lending.

    So you get saving. Now, there’s also this thing called “other changes” where things such as disaster losses etc make an appearance.

    Then you find revaluations and add.

    So, in summary, you add up household net lending, investment, other changes and revaluations to get end of period net worth from the beginning of the period net worth.

    But different countries have different formats. So “which tables” can only be answered on a case by case basis.

  4. Steve Roth writes:

    @Ramanan: “you add up household net lending, investment, other changes and revaluations to get end of period net worth from the beginning of the period net worth.’

    Right, I know that. But that word-equation doesn’t include “Private Surplus,” so doesn’t help me understand.

    Can you complete that private surplus/net worth equation for me, replace the question marks? I’m trying to understand what private surplus is (besides government deficit), how it relates to (change in) HH (private) net worth, using terms/measures from the national accounts.

    ??

    Thanks.

  5. Dan Kervick writes:

    I find the concept of “private sector net financial assets” to be of dubious economic significance. Some reasons:

    1. The proposition that financial claims on government are inherently less risky than privately owned real assets or financial claims on privately owned financial assets is an imprecise claim that is true only at some times and in some places and for some governments. Some governments are less stable than some private enterprises. And their policies might be less economically reliable, so that the real values of their nominal promises fluctuate unpredictably, even if they reliably meet those nominal obligations.

    2. The net wealth of the private sector can grow substantially during some period even if its net financial assets decline during that period. It’s wealth can even grow during a period in which it is a net debtor to the government sector throughout the period. Similarly, it’s net wealth could deteriorate even if its net financial wealth increases.

    3. There are all sorts of informal obligations to and from government which are not reflected in the usual accounting.

  6. Ramanan writes:

    Dan,

    Good points. Commented at Asymptosis as well … reposting it here …

    About the point on NFA … the primary aim is not NFA. Which is to not say that NFA is not important … it is important of course. But the way fiscal policy works is not via NFA but by boosting domestic demand and hence output. This results in people having higher incomes and they accumulate assets.

    The reason this is important is because one can have the private sector behave in a way in which private expenditure is higher than private income (for whatever reasons … great expectations of the future and so on). And if the government simply expands fiscal policy, it is not given that NFA will rise. It might fall as well (because of higher expectations of the future). But output will rise, income and expenditure both rise with NFA falling.

    So fiscal expansion shouldn’t be used synonymous with rising NFA.

  7. Ramanan writes:

    SR,

    Should have HH surplus and not private surplus. Apart from that.

    Starting HH Net Worth + Household Surplus + HH Investment + Other Changes + Revaluations = Ending HH Net Worth

  8. Dan Kervick writes:

    A few other observations:

    1. The “S-I” term in the Godley equation cannot be equated with any kind of private sector “surplus”. The fact that the non-government sector surplus/deficit is equal to the public sector deficit/surplus is an elementary flow of funds result that has little to do with the Godley equation. To tell the truth, why that equation (the usual derivations of which are badly fallacious) has come to seem so significant to people is mysterious to me.

    2. The change in net financial asset position of the non-government sector cannot be equated with the size of the government deficit/surplus. It also depends on other central bank operations that can change the composition and net equity position of the consolidated government balance sheet even if the fiscal arms of government are running a balanced budget.

    3. We live in a globalized economy in which many entities have financial claims on or obligations to various governments around the world. Even if we focus on the net accumulation of financial assets alone, the ability of the domestic US private sector to net accumulate such assets does not depend on the actions of the US government alone.

  9. Ramanan writes:

    Dan,

    Agree with 3.

    Not sure what 2 means. Central bank profits sent to the government affects the government’s budget balance. So I don’t know what you mean.

    About point 1, I don’t know why you say “S-I” is not the private sector surplus. It is the financial surplus, although the terminology surplus is less common.

  10. Steve Roth writes:

    @Ramanan:

    “Should have HH surplus and not private surplus. Apart from that.

    Starting HH Net Worth + Household Surplus + HH Investment + Other Changes + Revaluations = Ending HH Net Worth”

    Again, that equality doesn’t include Private Surplus, so it doesn’t help me understand what Private Surplus is.

  11. Steve Roth writes:

    @Ramanan:

    And: I’m pretty certain that MMT’s Private Surplus has a meaning that’s completely orthoganal to “Operating Surplus” in the national accounts. But…?

  12. Ramanan writes:

    SR,

    Yes operating surplus is a different thing. But just like the government’s budget deficit and surplus, other sectors can also have deficits and surplus, no?

  13. Steve Roth writes:

    @Dan Kervick: “I find the concept of “private sector net financial assets” to be of dubious economic significance.”

    I also, though I’m not nearly read to reject its usefulness. But I do question whether this statement:

    Nonfinancial Assets + (Financial Assets – Liabilities)

    carries a great deal more import than this one:

    Nonfinancial Assets + Financial Assets – Liabilities

    Or even this one:

    (Nonfinancial Assets – Liabilities) + Financial Assets

  14. Steve Randy Waldman writes:

    SR —

    First I’m going to steal an equation from my old post. See that post for the derivation.

    NET_HOUSEHOLD_FINANCIAL_INCOME = CURRENT_ACCOUNT_SURPLUS + CONSOLIDATED_GOVERNMENT_DEFICIT + ΔBUSINESS_NONFINANCIAL_ASSETS

    So, substituting the definition of PRIVATE_SECTOR_SURPLUS

    NET_HOUSEHOLD_FINANCIAL_INCOME = PRIVATE_SECTOR_SURPLUS + ΔBUSINESS_NONFINANCIAL_ASSETS

    Note that this measure of “financial income” is also a measure of “financial savings”, which may seem off since normally savings in income minus consumption. These equations reflect an exhaustive analysis of financial assets, so any consumption would be reflected as a reduction of some stock of assets. However, other things besides consumption would also change this measure of “financial savings”. For example, if a shop is liquidated and its assets transferred to households, NET_HOUSEHOLD_FINANCIAL_INCOME declines but there is not actual loss of savings or net worth. (There is a loss of “financial savings” or “financial net worth”.)

    Suppose we then define something called NET_HOUSEHOLD_NONFINANCIAL_INCOME such that

    TOTAL_HOUSEHOLD_INCOME = NET_HOUSEHOLD_NONFINANCIAL_INCOME + NET_HOUSEHOLD_FINANCIAL_INCOME

    by definition.

    Now we have

    TOTAL_HOUSEHOLD_INCOME = NET_HOUSEHOLD_NONFINANCIAL_INCOME + PRIVATE_SECTOR_SURPLUS + ΔBUSINESS_NONFINANCIAL_ASSETS

    if we define, naturally

    defining, naturally

    NET_HOUSEHOLD_NONFINANCIAL_INCOME = ΔHOUSEHOLD_NONFINANCIAL_ASSETS
    TOTAL_HOUSEHOLD_INCOME = ΔHOUSEHOLD_NET_WORTH

    we can substitute and get

    ΔHOUSEHOLD_NET_WORTH = PRIVATE_SECTOR_SURPLUS + ΔHOUSEHOLD_NONFINANCIAL_ASSETS + ΔBUSINESS_NONFINANCIAL_ASSETS

    This looks like a nice equation that gets at what you want, how PRIVATE_SECTOR_SURPLUS affects ΔHOUSEHOLD_NET_WORTH. It is comprehensive: any changes in the value, as well as quantity of assets gets included, as does the MMT private sector surplus.

    Unfortunately, as is often the case with pure accounting identities, it tells us less than we think. There is no guarantee that PRIVATE_SECTOR_SURPLUS, ΔHOUSEHOLD_NONFINANCIAL_ASSETS, and ΔBUSINESS_NONFINANCIAL_ASSETS are independent of one another, no certain story about what causes what. HOUSEHOLD_NONFINANCIAL_ASSETS is defined tautologically as anything that affects HH net worth that is not a financial asset, but this equationplay gives us no guidance about what that includes. For example, if you believe in “Ricardian equivalence”, then an increase in PRIVATE_SECTOR_SURPLUS caused by a government deficit is precisely offset by a reduction in HOUSEHOLD_NONFINANCIAL_ASSETS. This exercise cannot support a claim that an increase in PRIVATE_SECTOR_SURPLUS translates to an increase in comprehensive household net worth. That kind of claim is beyond the capacity of an accounting arrangement. We need to impose models about how these variables interact, and contestable mappings of these variables to real-world things.

    When MMTers’s use the standard sectoral balance decomposition, they usually do impose models. Often they presume that the consolidated private sector has an independent desire for net financial assets, which governments can accommodate smoothy by running sufficient intentional deficits or else will have forced upon them as private sector risk aversion provokes a recession that creates the required deficit as taxes fall more sharply than expenditures. Those are defensible claims, although as Dan Kervick points out, they are necessarily contingent: only certain kinds of government deficit spend in downturns near automatically, only certain kinds of governments issue claims craved by private sector entities as insurance. My main quibble with common MMT stories has to do with the treatment of private sector requirement of net financial assets as independent and external. I think that disruptive private sector risk aversion is largely a distributional phenomenon, and so the private sectors requirement of new net financial assets is conditional on the distribution of wealth and income.

    Blah!

  15. Marko writes:

    I think the most valuable point in SR’s post is in danger of being buried in accounting trivia : that we should be having open and frequent discussions in the public sphere about H-S (or “comprehensive”) income as it relates to economic performance and policy.

    Consider this question , if posed to the average citizen , or , for that matter , to the average Congressman :

    How many dollars of net worth is created in the U.S. economy for every dollar of GDP ( or GDI ) ?

    I bet the answers would be all over the place , even though Piketty’s work has flooded the media with simple pictorial representations that give us the answer , like this one :

    http://gabriel-zucman.eu/files/capitalisback/PikettyZucman2014Fig1.png

    In round numbers ( and ignoring bubbly distortions ) you could say that we generate $4 in net worth for every $1 of gdp , thus our H-S income would be 4+1 = 5 x gdp. For Japan it would be 7 x gdp. For the UK , say 6 x gdp. For every country shown , H-S incomes are multiples of gdp , and these multiples have been growing over time ( while gdp growth has been slowing ). Even back in 1970 there would have been in the range of 2.5-3.5 dollars of wealth created for every dollar of gdp. This ain’t a new , new thing. And a dollar from the sale of an asset spends just like a dollar from a paycheck , last time I checked.

    We should have been talking about H-S income decades ago.So why weren’t we , and why aren’t we now ?

    Ahh , never mind , carry on with the accounting trivia……

  16. Marko writes:

    I should have included the following label for the graph linked above :

    “Private wealth / national income ratios, 1970-2010″

  17. Dan Kervick writes:

    Not sure what 2 means. Central bank profits sent to the government affects the government’s budget balance. So I don’t know what you mean.

    Ramanan, imagine a hypothetical system where there is only a single bank, the central bank. Every household and every business holds it’s deposits at that bank. The bank makes and collects on loans to and from it’s depositors. The depositors also sometimes make loans to the bank (via the usual mechanism of shifting funds into interest-paying term deposit accounts.) There is no physical currency in this system: all payments are processed through depositor balances at the bank.

    Suppose the government has the usual fiscal operations, but maintains a strictly balanced budget. (One mechanism we might imagine: there are proportional and automatic tax deductions made from the various private sector accounts at pre-determined rates, in response to every single treasury expenditure.) Since there is never a deficit (or surplus), the fiscal side of the government issues no debt instruments.

    In a system in which there is such a budget balancing mechanism, there is still the possibility of the public having positive or negative net financial assets, and period-over-period changes in net financial asset position. In such a system that is all a matter of central bank interest rate and balance sheet policy.

    Obviously, also, in such a system the money supply can grow continually (whether or not net financial assets are growing) as the government gradually expands it balance sheet on both sides as the economy grows. There doesn’t seem to be any necessary connection, in fact, between the growth of the economy and changes in net financial assets. For example, during a period of high growth when policy is to slow down the rate of growth, but still keep it robustly positive, the central bank can reduce the interest in pays on term deposits and increase the interest it charges for loans. This will likely reduce the net financial assets of the private sector, even as that sector continues to grow. (Of course, since the economy is growing and saving its income, it’s real assets and real wealth can continue to grow even while its net financial assets are declining.)

    What is true of this system is true of ours as well. It’s just that we have an intervening layer of commercial banks in our two-tiered, but centralized system. Commercial bank lending does not in itself contribute to the net financial asset position of the private sector, since those banks are part of the private sector. But central bank lending to commercial banks and other institutions holding deposits at the central bank can and do influence the private sector net financial asset position. This is the case whether or not the fiscal branches are running a deficit, surplus or balanced budget.

    Whether or not a central bank returns net profits to the treasury, after expenses and dividend payments to shareholders) is a policy choice. In the United States, the Fed is not required to do so (although it has done so as a matter of Fed policy for many decades. But I don’t think that affects the overall conclusion anyway.

  18. Dan Kervick writes:

    Marko, I don’t that that chart answers the question you asked. That chart shows wealth-to-income ratios, which represent the aggregate result of long term wealth accumulation in relation to current national income. You can’t use that number to determine how much net wealth is generated in a given period by each dollar of GDP in that period.

    Every year, we consume the vast majority of what we produce, and save only a fraction of it. Our national net worth increases incrementally by that saved amount. Typical national savings rates are usually closer to 10%. So for each dollar of GDP generated, 10 cents of additional net wealth are accumulated is a more reasonable figure.

    Most people, of course, live hand to mouth and have a net worth that is much smaller than their annual income. The income comes in and is consumed, and their wealth doesn’t change much.

  19. Note that repairs maintenance and depreciation of housing are already deducted when computing net income. (PDF)

  20. Marko writes:

    Dan,

    For the wealth/gdp ratio to remain constant, wealth and gdp must grow at the same rate. If wealth is $400 and gdp $100 ( 4x ratio ) , and gdp goes up by 1% ( $1 ) , then wealth must rise by 1% ( $4 ) to maintain the ratio.

    So , $4 wealth per $1 gdp.

    See SR’s accounting statement in the referenced post , line 21 , for some real-world numbers.

    The U.S. ( or any developed ) economy has a fairly stable multiplier comparable to the p/e multiplier of a firm. If a firm has a p/e of 15 and earnings rise by a buck , the valuation rises by $15. Automatically. Out of thin air , so to speak.

    You provide a useful example of what sort of responses would arise to the question I posed , which illustrates why the H-S income concept should be part of our everyday economic discussions.

  21. Steve Roth writes:

    @Marko: “How many dollars of net worth is created in the U.S. economy for every dollar of GDP ( or GDI ) ?”

    I think you’re reading this wrong. It’s how many dollars of wealth/net worth exist for every dollar of GDP.

    The graph is often used to discuss the return on “capital” or wealth. Flipping it on its head, we can say that it also depicts the velocity of wealth — a perhaps-fruitful measure for further exploration.

  22. Steve Roth writes:

    @Marko: If you want to examine how much net worth is created per dollar of income/production, check out the explanatory precursor post linked from the post being discussed here:

    http://www.asymptosis.com/why-you-probably-dont-understand-the-national-accounts-in-pictures.html

    It graphically displays 1. Sources by amount and percent of total, and 2. Uses and Change in Net Worth by amount and percent of total — exactly what you see in the accounting statement, but in graphical form.

  23. Steve Roth writes:

    @SRW: I’m going to need to spend some time with your post, preferably when my brain is fresher. Thx.

  24. Steve Roth writes:

    @Dan Kervick:

    A question for you in the context of your thought experiment:

    The stock market goes up. All else equal.

    +Private-sector financial assets. No change to liabilities. So +NFA.

    Am I thinking about that right?

    If yes, it’s at least interesting that the stock market runup does not affect the Net Acquisition of Financial Assets measure. (Because that would require resort to the revaluation account.)

    The private sector can accumulate financial assets without acquiring them?

    Realizing: should it instead, properly, be Acquisition of Net Financial Assets? The existing term suggests acquisition, net of dis-acquisition, not acquisition of (Financial Assets – Liabilities).

    That usage would problematic, though, because the term “Net Financial Assets” does not appear in the FOFAs. I’s easily derived, of course, but the national accountants don’t seem to think it’s a measure worth reporting.

  25. Marko writes:

    SR,

    Yes , I think my question is phrased incorrectly. It should be : ” How much new net worth is created for every dollar of gdp GROWTH. ” This question corresponds with the response to Dan above.

    Clearly this is an “over the long haul” type of relationship. You wouldn’t expect a precise relationship q to q or even yr to yr , but that’s true for everything else , too. When we say that real gdp per capita has grown at x % over the last y decades , and point to the straight line on the log graph , we also know and accept that there’s a lot of variation that occurs within select periods.

    No , though , on the velocity of wealth speculation. I think that’s as backwards as a company with a p/e of 15 trying to raise earnings by a dollar by artificially boosting the valuation by $15. It’s cart-before-the-horse insanity , no matter what Roger Farmer’s intensely over-specified econometrics says.

  26. Marko writes:

    I should have done this right off :

    https://research.stlouisfed.org/fred2/graph/?g=21j3

    Annual change in net worth / annual change in gdp

    Dotted line marks 4 x multiple.

  27. Steve Roth writes:

    @Marco: https://research.stlouisfed.org/fred2/graph/?g=21j3

    That graph shows that annual change in net worth is far more variable than change in GDP. Generally 4x as much variation. But there’s much variation in that variation in change in levels. ;-)

  28. Detroit Dan writes:

    I’ll stick with MMT. MMT helped me to understand how the economy works, and all the criticisms are, as admitted here, stylistic. Sometimes simpler is better.

  29. Steve, no.

    Domestic stock market goes up, liabilities of the domestic corporations go up by that amount. Only to the extent those companies are held by domestic private agents does domestic private financial assets rise.

    That is, domestic private financial assets fall, on net, to the extent that the company is foreign and or government owned.

    Or, if you include private to include ROW and the government holds no stock, then private net financial assets are exactly unchanged.

  30. Marko writes:

    SR,

    Yes , there’s a lot more variability in asset values. What I find amazing is how rare are the years when change in NW/change in gdp falls to <2. In other words , net worth changes virtually always make a bigger contribution to H-S incomes than GDP changes.

    I just noticed I also messed up the country calcs for H-S incomes above. For the U.S. , for example , and using the 4x multiple , a typical year's U.S. aggregate H-S income would be (roughly) equal to GDP + 4 x deltaGDP , not simply 5xGDP as I noted above.

    U.S.capitalists make out like bandits at the expense of workers , and this is best revealed by combining the wealth and income discussions , rather than treating them separately. This is why a
    H-S income definition needs to become a standard , regularly reported dataset , preferably disaggregated by deciles , top percentages , etc.

  31. Neil Wilson writes:

    “Domestic stock market goes up, liabilities of the domestic corporations go up by that amount.”

    They don’t. It’s not true. The price of a share can go up way beyond what is in shareholder equity at that point. The idea that the two are the same is neo-classical bullshit.

    If the stock market goes up then the revaluation above the shareholder equity represents an unrealised gain, and therefore an unrealised loss to somebody else in the system. The problem is you don’t know which sector the unrealised loss will eventually be realised in.

    But you do know that the ‘holding gain’ is a zero sum operation. So you exclude it from proceedings until a transfer happens which makes the gain and loss concrete.

  32. Ramanan writes:

    Steve,

    There are various ways in which the word ‘net’ is used in national accounts. With different meanings. A guide of the flow of funds or UN guides on national accounts is your best friend but here’s a substitute.

    First, net is net of consumption of fixed capital (“depreciation”). So one has gross saving and net saving. The latter is saving net of consumption of fixed capital.

    Second, net is assets netted with liabilities. So NFA belongs to this type. Financial assets net of liabilities in this case. One can also have “Net Assets”.

    Third, netting with disposal, redemption etc. So I have acquire financial assets, sell them and buy more financial assets. So we can talk of gross acquisition and net acquisition. Remember it’s not the only thing. There’s also incurrence of liabilities. So there’s net incurrence of liabilities.

    One can mix second and third and create things such as net lending. This is net acquisition of financial assets minus net incurrence of liabilities. Sometimes people use “net acquisition of financial assets to mean net lending”.

    Fourth, (not used by national accountants), there’s net saving (saving net of investment). This is used in Wynne Godley’s papers and MMTers also use it.

  33. Ramanan writes:

    Dan,

    Your point is technically right. It’s possible the government budget is balanced, whatever the reason and the central bank can purchase nonfinancial assets and the private sector can have a positive stock of net financial assets.

    But everyone is telling a model of the world. There are always implicit assumptions. So when there’s a talk of net financial assets, it’s implicitly assumed that the central bank is not doing such a thing as in the first para. And even if it is, the central bank has been assumed to be a part of the government and the central bank has a deficit and hence also the government.

  34. Neil,

    This is not neoclassical theory. This is Godley and Lavoie. This is real life accounting in the published U.S. integrated macro accounts.

  35. Ramanan writes:

    David is right.

    The system of national accounts uses this convention. If stock prices go up, equity liabilities of corporations go up. Their net worth can become negative.

  36. JKH writes:

    “This is real life accounting in the published U.S. integrated macro accounts.”

    I think this is an area where it is wise to become extremely careful.

    There is of course a massive distinction between the way this is done in the SNA, the integrated accounts, and Godley and Lavoie versus the way it is done in standard financial accounting for a corporation for example.

    Suppose the world consisted of a single hard asset in the form of a factory and a single financial asset in the form of an equity issued by the owner of that factory.

    In standard financial accounting, the world holds an equity claim of some market value.

    It also holds a real asset of some value.

    The key is that you have to make a choice – to value the world according to the claim or to value the world according to the underlying hard asset.

    I think that is the accounting reality in fact.

    It is a choice. And because it is a choice, the value of hard asset claims and equity claims cannot be additive in a summing up of the world’s asset value.

    And I think that framing of choice is quite different than the framing that leaps directly to the assertion that world financial assets have a net value of zero.

    (I suspect both Ramanan and David Rosnick may be able to confirm this, but my recollection is that Godley and Lavoie in fact become very guarded and qualifying when describing the netting methodology they use in the case of equity claims, where the market value of equity is treated as both a liability and an asset in correspondence at the interface of issuance and ownership. I’m pretty sure they describe it very deliberately as an artificial device that is employed specifically and artificially for the purpose of achieving convenience in matrix additivity. In that sense, it is not “real world”. Financial accounting is the real world methodology in such a context, and in financial accounting there is definitely such a thing as net world financial assets including net world equity claims. That is what shows up as the direct and indirect equity claims of households on the rest of the world in the Fed flow of funds reports (although to be sure that indirect equity value component is itself a reflection of claims on claims issued further upstream in the business world.) The fact that net financial claims exist in the real world is a consequence of the fact that households do not issue equity claims, and that the telescoping of all world financial value stops at the household level. The “convenience device” of invoking an equity liability used by Godley and Lavoie spills over into SNA and the integrated accounts for essentially the same reason I suspect. I hope Ramanan and David don’t disagree with me on my recollection and interpretation here, or I will be forced to go diving into my Godley and Lavoie. And I much prefer laziness of recollection to the hard work of confirmation, or in the worst case disappointment. The latter outcome being sufficient reason to disappear as suddenly as I entered here.)

    The GL negative equity methodology is perfectly understandable for the purpose of macro aggregation ease. They explain why they want to use it quite clearly.

    But the “real world” accounting reality is not that the net equity claim value of the world is zero, in my opinion.

    I raise this here because it seems to me to be a pretty important qualification to keep in mind at the meta-level in understanding macro accounting options. There is sufficient confusion about how financial accounting works on its own at the micro level, that becoming hazy when it comes to fitting financial accounting into macro aggregation runs the risk of global blogosphere chaos on the subject – and we certainly wouldn’t want that. And whether equity is treated asymmetrically or symmetrically at the macro level seems to be a major fork in the road and building block of two different methodologies – one of which is arguably more real world than the other in my opinion. So let’s be clear that there is a choice in the road fork, and not become Yogi Berra in this regard.

    Finally, I use a little mental trick an an attempt to inject some intuition into the negative equity riff of Godley and Lavoie et al. If the market value of equity is less than its book value, then the issuing entity is deemed to have positive net worth in the GL world. The equity liability has become less than the book value – it has become a cheaper liability. If GL net worth is positive, then other things equal, the stock has also become cheap from a management perspective. In that case, it becomes a good time for management to do a share buyback, all other things equal. Positive GL net worth is a signal that that stock has become cheap, and that there is value to be gained in a share buyback, all other things equal.

  37. Ramanan writes:

    JKH,

    True SNA and G&L are different from business accounting. I don’t recall their arguments in detail except that they say that corporations feel pressured to pay dividends and hence it’s a liability like a debt security.

    My own argument is that if one doesn’t treat it like that, liabilities to foreigners may not be stated well.

  38. Dan Kervick writes:

    Ramanan, my account didn’t depend on any central bank purchasing of real assets. It is was based only on the idea that as a result of the conduct of its monetary policy, it is possible for the financial liability of the central bank to the private sector to exceed the financial liabilities of the private sector to the central bank. Whether a central bank operates with a negative equity position or a positive equity position is a policy choice.

  39. Ramanan writes:

    Dan,

    True. I see your point.

    It’s possible the government budget balance is balanced for a while and the central bank’s liabilities to exceed its assets. In that case, it becomes possible for the private sector to have a positive net stock of financial assets.

    Steve Roth may like your point.

    And with your example, I agree that it is dangerous to write statements
    “…. only possible if …”

    such as

    “the private sector can only have a positive NFA if the government runs a deficit”.

    Such things should be resisted.

  40. Ramanan writes:

    Dan,

    A more practical example is this:

    Suppose the government budget is balanced for whatever reason during a fiscal year.

    At the start of the fiscal year, bond prices start rising, maybe because of a rapid drop of interest rates by the central bank. One doesn’t even need the central bank to change its balance sheet much: announcement could be enough.

    Assume public debt to start was 100% of GDP. Assume the total value of government debt rises 10% during the year due to fall in short term interest rates leading to a rally in the bond market.

    In this case, the private sector’s net stock of financial assets has risen despite the government running a budget, mainly due to revaluations.

  41. Bob writes:

    Steve unrelated but there is a good MMT book/eBook here – understanding government finance:
    http://www.bondeconomics.com/p/books.html?m=0

  42. JKH writes:

    Ramanan,

    The part in the book I am thinking about is on page 30.

    They do note Joan Robinson’s dividend argument there.

    But they also refer to the necessity of having consistency for purposes of matrix aggregation.

    I think the Robinson argument is very flimsy. Lots of companies pay no dividends, and the payment of a dividend is always an option exercised one dividend at a time by the Board of Directors. A corporation simply does not have a liability to pay out a dividend or any other cash payment (e.g. share buyback) until the decision is made at the option of the Board, time after time. There is no ongoing legal liability the way there is for debt. Cutting a dividend is not a failure of liability – missing an interest or principal payment is.

    If the objective is creating a rationale for shoe horning a micro-founded reason for treating equity as a liability, I would instead use one that treats equity as a contingent liability under the wind-up of the firm. If there is any money left over after paying off creditors, it would go to the equity holders as a matter of liability. That is a more sound in terms of economics and legality than the dividend argument in my view, but it is still a forced treatment.

    The real reason is matrix additivity convenience. Nothing wrong with that, but it is pretty artificial.

    Again on page 28, their comment regarding the GL net worth calculation is sort of revealing:

    “… keeping in mind that the measured net worth of firms is of no particular significance. Indeed, in the book, no behavioral relationship draws on the definition.”

    The real world accounting definition and interpretation of net worth is of course very significant for behavioral relationships. It’s an interpretation that is fundamental to capitalism.

    I have great respect for how GL develop the book using this consistency approach, but I think this is why I prefer wading through the Fed flow of funds presentation in these matters.

    On the foreigner issue, I’d make the same observation. Showing the gross international investment position as a balance sheet of assets and liabilities, where liabilities may include equities held by foreigners, is a presentation convenience. In fact, foreigners holding any type of financial asset merely removes that asset from what may ultimately be available as claims held domestically. It is the removal of the asset from the domestic pot that is of importance, not that equity claims held by foreigners become liabilities of the domestic sphere.

  43. JKH,

    The IMAs are presented in the FoF.

    Also, foreign ownership is macroeconomically significant as there are income flows associated with foreign-held equity.

  44. Ramanan writes:

    JKH,

    Yes Joan Robinson is flimsy. A non-payment of interest is likely to be looked much worse than a non-payment of dividend. However I could argue that there is still pressure to pay dividends. It’s also true that a lot of firms do not pay dividends for a long time. Perhaps they convince their investors that it is the best thing to do and the investors don’t revolt too much. But there’s a big share of investors which loves dividends.

    There’s this “Dogs of the Dow” theory which believes that dividend is the best indicator. I don’t know how true that is (whether it works in practice) but as long as investors believe in the “Dogs of the Dow” theory, there is pressure by the stock market.

    I see your point about foreigners. I think you are saying that one should look at assets held by foreigners. There is however a consistency requirement there and it looks odd done in a way which is non-SNA/G&L. Suppose in that way of maintaining records, foreigners liquidate stocks and buy bonds instead. How is it that liabilities suddenly jump?

  45. Steve Roth writes:

    @Marko and @David:

    I was wrong about variablity. Ignore that. Brain was tired and I was heading out the door. Sorry.

  46. Nick Edmonds writes:

    JKH,

    “It is the removal of the asset from the domestic pot that is of importance, not that equity claims held by foreigners become liabilities of the domestic sphere.”

    An excellent point and one that is not often appreciated.

    The approach in G&L of recording all liabilities (including equity) at their asset value is indeed just a modelling convenience to facilitate the quadruple-entry system. There is, however, nothing to stop the modeller using a different valuation for the purpose of behavioural assumptions.

    It is perhaps worth mentioning that this whole private sector surplus issue is to some extent down to modelling convenience. The aggregation of households and firms was used by Godley and Cripps in CEPG UK model partly due to the difficulties of separately estimating business investment but also because various stylised facts about the UK economy at the time suggested this as a pragmatic approach. I don’t think it was ever held that this technique would always be appropriate and arguably developments since then, particularly the growth of household debt and of corporate cash piles, make it somewhat less useful.

  47. Steve Roth writes:

    @Marko: “H-S income definition needs to become a standard , regularly reported dataset , preferably disaggregated by deciles , top percentages , etc.”

    Totally agree. Unfortunately the only solid work on that I know of is the Armour, Burkhauser, and Larrimore paper I linked to in the post:

    http://www.nber.org/papers/w19110

    Which comes to conclusions that are wildly at odds with almost every other measure of distribution/concentration/equality.

    It’s quite possible to replicate their work (at least similar) using CPS and/or SCF, with IMA revals as indices, but hey: I don’t get paid to do this stuff, and it would be a hell of a lot of work. That’s what we pay national accountants for!

  48. Steve Roth writes:

    @Ramanan:

    Thanks for “net of WHAT, buster??” primer. ;-) Most I knew, but it’s often hard to find.

    You’re aware, right, that NFA also doesn’t appear in the FOFAs?

    If the national accountants didn’t consider it a measure worth reporting, I can sort of understand. It’s kind of arbitrarily posting all of an institutional unit’s liabilities against its financial assets.

    Curiously, I find that the Aussies don’t post NAFA, but they do post net acquisition of nonfinancial assets.

    http://www.abs.gov.au/Ausstats/abs@.nsf/0/F9DD5C2D78BA68AECA25697E0018FBCA?opendocument

    Must be a southern-hemisphere thing.

  49. Ramanan writes:

    SR,

    They do:

    page 22 here:

    http://www.ausstats.abs.gov.au/ausstats/subscriber.nsf/0/B1BA419C3D86ADE0CA2577FA0011CA6A/$File/52320_sep%202010.pdf

    Page 20 has net financial position of households, which is the difference between financial assets and liabilities.

  50. Ramanan writes:

    Sorry that’s Tables 22 (for NAFA) and Table 20 for NFA, not page 20 and 22.

  51. JKH writes:

    David R.,

    Right. I don’t look at it all that often and tend to recall the old version, pre-IMA incorporation. When I do look at it, I always start with B.101 and work out from there.

    Ramanan,

    Right. Visualizing the balance sheet presentation for the gross international investment position is very convenient indeed as a subset of the main show. I start with B.101 and then imagine the GIIP balance sheet as an overlay/adjunct – a hybrid of visualization techniques if you will. But in fact my point applies to equities (direct and indirect) held on the domestic household balance sheet whether issued by the domestic or the foreign sector. It is the asset identification that is important – not a convenient liability classification on the books of the issuer. As for equities issued by the domestic sector and held (directly or indirectly) on the foreign balance sheet, it becomes a neck twisting mind bender to imagine those items on the right hand side of the constructed gross international investment position being shotgunned out of the domestic pot ultimately into the rest of the world equivalent of the US Z1, whether via intermediaries or directly, again important primarily for the asset identification from that perspective.

    Nick E.

    Thanks be for your first point. That is authoritative confirmation for me that I’m not completely insane on this point of emphasis.

  52. Ramanan writes:

    JKH,

    It’s about how one looks at it … as you mentioned … the telescoping at the household level … There is an alternative way to look at national accounts, which is at the national level. Indeed some Austrians accuse national accountants to be both biased toward Keynesianism and Mercantilism. Which is true, except it’s not a bias. Keynesianism is true. Mercantalists were right on many things.

    So … imagine a stock market boom. A book value reporting will under-report the liabilities to foreigners. You could say that one can look at the asset side of the International Investment Position of the Rest of the World. But nobody produces such a thing. So anyone worried about debt to foreigners would look at the liabilities side of the International Investment Position and National Accounts is designed such that there is consistency between the two, as the national accounts itself has rest of the world positions.

    As in if national accountants were to produce statistics differently, where would they look if officials are worried about equities held by foreigners?

  53. JKH writes:

    Ramanan,

    I’m fine with that balance sheet representation, and I would never suggest equities be represented as assets at book value.

    I’m just saying that equities are not a liability in the real world, and the calculation of GL net worth is a perverse outcome of representing equities as a market value liability. They say so themselves. And even if you construe equities as a liability, it is another step further to construe them as a market value liability.

    My point is that the real world interpretation of an equity claim is not as a liability. So people viewing that balance sheet and thinking of nothing else may misconstrue the nature of the financial claim.

    In the case of equities, I would think of the right hand side of the GIIP as a subtraction of domestically issued equities from the domestic pot of what is otherwise available for domestic equity asset holdings. A subtraction rather than a liability, somewhat analogous to the depiction of an equity on the right hand side of an issuer’s balance sheet – where it is not classified as a liability either.

    But all that said, I’m OK with the GIIP representation. This is just a footnote to it.

  54. Steve Roth writes:

    Hey all:

    I don’t have much time at the moment to read and respond to great comments, so quickly:

    I also just re-read G-L 2.2.2, pp. 27-31, on the balance sheet of production firms (also to a great extent financial firms).

    Now please understand: I’ve read this at least three times in the past, and JKH schooled me on the subject at some length in the past: http://www.asymptosis.com/why-the-fed-hates-inflation-1-2-trillion-dollars-of-why.html#comment-25863

    But having now worked the math problem in rather excruciating detail by building that alternative account presentation, this is the first time I read it and said, “oh yeah, that’s obvious.” (As is usually true with math problems…)

    The one place that I’d embellish is in their discussion of bonds vs. equity, their distinguishing characters: “interest payments are a contractual obligation, where payment of dividends is not.” More to the point, I think, is that bonds’ and loans’ principal can (will) be forcibly redeemed by the lender — on demand or a fixed date — whereas a corp can never be forced to redeem equity. That more than the other makes equities’ treatment as liabilities problematic. SRW has written on this redeemability thing quite cogently, don’t have time to dig up the link.

    Amazing how G-L repeatedly pooh-pooh the immediate import of this treatment of firm net worth — it really doesn’t “make sense” (my words) and they say “no behavioural relationship draws on its definition” — even though it’s necessary for the stock-flow-consistent macro analysis.

    Viz: “Another way out, which national accountants seem to support, is to exclude the market value of issued shares from the liabilities of the firms. This is the approach taken by statisticians at the Federal Reserve.”

    In that approach, the households-own-firms HH net worth = Private net worth identity that I’ve been touting holds true, and market runups increase HH/private sector NW, as you (and G-L) would naturally expect. But there’s that missing balancing item.

    JKH’s best effort at describing it in the past was a “balance sheet mismatch.”

    But in any case, as implied by G-L’s “Another way out” usage and others, this remains a true conundrum — even (especially) if you understand clearly what the conundrum is.

  55. Steve Roth writes:

    Oh also, G-L:

    “In some sense, the value of these shares is something the firm owes to itself.” My kind of philosophy-of-accounting thinking…

    A thought experiment I’ve pondered:

    What if all firms bought back all their shares, so they were all independent entities sans shareholders, like nonprofits? Whither Citizens United?

  56. Ramanan writes:

    JKH,

    Inconsistency arises if you consider international matters. One is forced to treat equities as liabilities. And I’d argue that one is forced to use the market value.

    This is because foreigners can liquidate equities and do many things with this.

    Imagine if there’s only one investor from the UK holding C$1bn of equities in Canada. If one doesn’t treat them as liability at all, then inconsistency arises if the investor sells the equities and buys some Canadian T-bills. Liabilities jumping from 0 to C$1bn.

    And using the same argument leaves one with only the market value.

    But even with domestic matters, it’s not as if it’s not a liability. In addition to the dividend argument, there’s another I could put across: stock buybacks. It’s related in a sense. Management thinks it’s not good to pay dividends to people, so why not buyback shares. If there’s no pressure to pay dividends, there’s no reason one should buyback stocks.

  57. Marko writes:

    SR,

    “… Unfortunately the only solid work on that I know of is the Armour, Burkhauser, and Larrimore paper I linked to in the post….Which comes to conclusions that are wildly at odds with almost every other measure of distribution/concentration/equality…”

    Those guys have brass balls , I’ll say that. They’re career right-wing hacks , and damn proud of it!

    They rely heavily on survey data , missing the top 1% , where all the inequality action arises. They cherry-pick start and stop dates to get their desired results. (I’d like to see them update it through 2014. Even with all the hackery , their results would do a backflip followed by a faceplant.)

    Birthers – climate deniers – Burkhauser catchfarts. All peas-in-a-pod.

    Baker and others have done preliminary dissections. ( It makes no sense to do more than that – it’s not like you could shame them into becoming legitimate. ) :

    http://www.cepr.net/blogs/beat-the-press/thomas-edsall-on-richard-burkhauser-and-inequality

    http://www.tcf.org/blog/detail/graph-redefining-income-inequality-wont-make-it-go-away

    http://jaredbernsteinblog.com/when-the-results-look-weird-check-the-methods-carefully/

    Besides conforming to their immediate needs , I suspect their use of an H-S income variant was also a preemptive strike of sorts. A way of saying : “Careful what you wish for , progressives , you might not like the outcome. ” The reality is that the rich are petrified by the prospect that the combined wealth and incomes of individuals would become a regularly reported statistic , as opposed to our current practice of separate reporting , which allows for propagation of the fallacy that the income-rich and the wealth-rich are largely different populations.

  58. Steve Roth writes:

    @Marco:

    Everything you say.

    Only one correction:

    “combined wealth and incomes of individuals”

    Not really combined, certainly not additively. I would say instead “the full extent of income from ownership.” “Other Ownership Income” in my accounting.

    (I’ve changed my usage from “property income” to “ownership income” for reasons I’ll explain in a more appropriate forum.)

  59. Steve Roth writes:

    @JKH: “If there is any money left over after paying off creditors, it would go to the equity holders as a matter of liability.”

    That does point out (the?) one situation where a corp could be forced by a shareholder to “redeem” shares.
    Shows, I think, that those shares’ nature as a liability is more about their “principal” value than their periodic payouts.

    Of course understand why G-L include this (rhetorically and behaviourally inert) liability so the matrix works, but I’m really wowing at the construct that results:

    Firms Net Worth = Firms Net Worth (as we usually understand it) – Equity

    Have I got that right? Pretty amusing.

    @JKH:

    “But the “real world” accounting reality is not that the net equity claim value of the world is zero, in my opinion.”

    Right. The buck stops at households.

    “they describe it very deliberately as an artificial device”

    In my previous schooling JKH described them as “actuarial liabilities.” Not a bad description.

    @Neil: I would very much enjoy any links to your writings expanding on this zero world equity construct being “neoclassical bullshit.” I intuit likewise, and have some notions of why that is so, but your explanation would be much appreciated.

  60. Marko writes:

    SR,

    Yes , for some reason I’ve got the sloppies on this thread.

    I’d be content with an H-M measure that used either consumption ( i.e. income – savings ) or income , combined with change in net worth. To be scrupulously fair , the income measure should be comprehensive , post-tax , and including all supplements and in-kind income of all kinds , from all sources , eliminating the endless confusion about market vs post-tax/transfer incomes that has unnecessarily complicated the inequality debate. I’d even go so far as to say that some reasonable attempt be made to calculate individual’s social security present value , adding that to net worth and to changes in net worth as the present value evolves over time.

    Now that the FOF has the new B1 table showing national net wealth , measured strictly in nonfinancial terms , it seems to me that the goal should be to get to the point where a similar measure is made on the individual level such that all individuals added together with gov’t would equal the net national wealth figure , within reason. Gov’t bonds and such would be nonfinancial assets held by individuals , of course , but hopefully most of the accounting at the individual level could similarly be reduced to “real” measures. May be a pipe dream , though.

    Also , we need a comprehensive wealth registry , as per Piketty. I’m SURE that’s a pipe dream.

  61. Consider:

    Examples of equity instruments include… instruments that impose on the entity an obligation to deliver to another party a pro rata share of the net assets of the entity only on liquidation

    (PDF)

    So rather than focusing on an uncertain obligation to pay dividends, it may make more sense to think of the liability in terms of a genuine obligation to pay out upon liquidation.

  62. JKH writes:

    Steve R.,

    Standard net worth
    = book value of equity

    GL net worth
    = book value of equity – market value of equity
    = standard net worth – market value of equity

    GL net worth goes positive if the market value of the stock falls below book value.

    Which is counterintuitive to the standard interpretation of book value net worth, which is that lower net worth is a bad thing other things equal

    Similarly GL net worth goes negative if the market value of the stock rises above book value.

    Which is counterintuitive to the standard interpretation of book value net worth, which is that higher net worth is a good thing other things equal

    Standard book value net worth is silent on the market value of equity, which is more the holder’s concern for a given book value. Net worth book value increases through growth by retained earnings, so the market value of equity will normally increase along with that by putting an external value on that increasing book value and how the company deploys that capital.

    (To mix it up further, a household that buys an equity at market value will record that purchase price as the book value of its new asset. Then when it sells the equity, it will record the market value of the then sale price and pay tax on the capital gain over the book value.)

    Tobin’s Q
    = (market value of equity/book value of equity)

    So,

    GL net worth is positive if Tobin’s Q is 1

    This is consistently counterintuitive with the normal orientation for Tobin Q interpretation.

    As I said in a previous comment, I can force intuition into the GL construct by noting that a positive GL net worth (where Tobin’s Q is < 1) means that the stock is “cheap” from management’s perspective, and they may want to consider a share buyback. There is positive GL net worth evident because book value is now at a premium over market value, and the market is failing to recognize as least the full book value of the company. Think of it as management acting on disappointment in the market’s valuation of the stock they have previously issued.

    I’m bipolar on the overall GL construct. I think it’s magnificent from the perspective of global consistency in balance sheet construction. I think it’s terrible in terms of natural intuition, and unfortunate in the fact that it turns normal real world financial accounting of micro level net worth on its head, a disappointing inconsistency in the interface between the two.

    But my final impression of it is that it is very well done in the book and for good reason.

  63. JKH writes:

    Not sure what happened in the middle of that last comment.

    Should read:

    …………………………………………………

    Tobin’s Q is (market value of equity/book value of equity)

    So,

    GL net worth is positive if Tobin’s Q is 1

    This is consistently counterintuitive …

    ……………………………………………………..

  64. JKH writes:

    jeez

    Should read:

    …………………………………………………

    Tobin’s Q is (market value of equity/book value of equity)

    So,

    GL net worth is positive if Tobin’s Q is less than 1

    GL net worth is negative if Tobin’s Q is greater than 1

    This is consistently counterintuitive …

    ……………………………………………………

  65. JKH writes:

    Ramanan,

    Off the top of my head, if I were depicting the US GIIP, I might do it roughly along the following lines, where I’m showing only the right hand side of the balance sheet for illustration purposes:

    RHS

    Liabilities (e.g. bonds)
    Market value of equity claims issued
    Net equity value

    For the US GIIP, that last line would be a negative number, reflecting the US international balance sheet mismatch where the RHS value exceeds the LHS value (apart from that negative entry, which becomes the balancing net worth position for the GIIP alone).

    This would align the presentation somewhat more with micro financial accounting for RHS liabilities and equity.

  66. JKH writes:

    Ramanan,

    Net equity value above might also be called net worth as it pertains to the US GIIP balance sheet on its own

  67. Ramanan writes:

    JKH,

    Yeah, although the net worth would appear in the national balance sheet. Terminology wise, the IIP just restricts to foreign assets and liabilities. So actually the net worth is a large positive number because of the large nonfinancial assets the US has like office buildings, machines, houses etc.

  68. JKH writes:

    Ramanan,

    Of course I’m referring to the net worth of the GIIP balance sheet as a subset of total net worth.

    As you know, it currently makes a negative contribution to total net worth.

  69. Ramanan writes:

    JKH,

    Yeah.

  70. JKH writes:

    Ramanan,

    Here’s another way to make one of the points I want to make.

    Suppose the US GIIP consisted of nothing but equity claims on both sides of the balance sheet.

    Then you’d have roughly $ 24 trillion of equity claims acquired LHS and $ 30 trillion of equity claims issued RHS.

    There is no effective liability anywhere on that balance sheet. All that stuff could go to zero, without liability (except arguably in the case of windup of individual issuing entities, in which case some residual value if it existed might be paid out).

    (Within that GIIP profile, there is a net equity or net worth contribution of $ (6) trillion to US household net worth. With a counterfactual balanced GIIP position, US household net worth would be $ 92 trillion instead of $ 86 trillion, other things equal)

  71. Ramanan writes:

    JKH,

    I view it as a liability just like debt. That’s because a foreigner can sell equities to a resident economic unit and purchase a bond instead (in the simplest case from the same unit).

    Because foreigners always have this option, the nation as a whole is net indebted to foreigners in the example ($24 trillion LHS versus $30 trillion RHS).

    The other way to look at it is thinking of oneself as a policy maker. Over the next 5 years, so much of dividends will be paid affected the current account of balance of payments … so much of more accumulated debt and so on.

  72. Ramanan writes:

    JKH,

    And also instead of saying household net worth, I’d talk of the nation’s net worth. This is because a nation is bigger than households. The government has a lot of assets under the production boundary, for example. The US government I think has a lot of nonfinancial assets. It’s just that these are rarely officially calculated.

    So one can imagine a nation in which households are not rich overall but the government has a huge amount of non-financial assets compared to a case where households have the same sort of balance sheets but the government doesn’t have much assets. The two are different, although just looking at household balance sheets would lead one to believe that the two are the same.

  73. Steve Roth writes:

    @David: “liability in terms of a genuine obligation to pay out upon liquidation”

    Considering, this has import to “equities as liabilities” because in many cases, liquidation means “sale.” (If a business fails, net worth generally just goes negative, the shares become worthless, and there’s nothing to redeem.) And in many cases especially with smaller businesses (like the ones I’ve sold), the method is an asset sale, not a stock sale. The corporation has an obligation to pay ending (“liquidated”) net worth to the shareholders.

    (Mechanics of a full asset sale: Usually the original corp — often renamed because the name is one of the sold assets — becomes nothing more than a temporary container and conduit for the proceeds from the asset sale, fully netted before distribution, all final miscellaneous obligations paid. Those net proceeds, the final net worth, is distributed to equity share holders, then the corp is shut down.)

    But, again, the issue of forced redemption: in the real world of large business, shareholders have very limited power to force a sale hence redemption. It’s a long step from a loan or bond — an obligation of the corporate entity as a going concern — coming due on a fixed date.

    @JKH: I think you’ll be happy to hear that that comment tells me nothing I don’t already know — and fully understand! ;-)

    Just one petty cavil: precisely, Tobin’s Q is market value / replacement value. Replacement value, of course, is unobservable; what would be the cost of re-creating that complex of trained employees, procedures, brand recognition, etc. etc. etc. — the things that make a business a profitable, going concern?

    So we’re left with market value and book value, both of which are in a very real sense estimates of that replacement value.

    Wiki:

    “Although it is not the direct equivalent of Tobin’s q, it has become common practice in the finance literature to calculate the ratio by comparing the market value of a company’s equity and liabilities with its corresponding book values as the replacement values of a company’s assets is hard to estimate.”

    And just to be annoying: if market value is the best estimate of replacement value, Tobin’s Q is always 1!

    On the equity/liability treatment, here’s what I’m realizing seems odd to me:

    We know why G-L do the equity/liability thing. Perfectly sensible, because the artificial “net worth” of firms is immaterial and inert in their model.

    But why do the FOFAs/IMAs present it as they do, with equity tossed in among all the other liabilities? Equity is special and distinct, as explained so well in G-L and acknowledged on the RHS of standard corporate balance sheets: “Liabilities and Shareholder Equity“.

    So why not this for the last three lines:

    Net Worth
    Corporate Equity
    Net Worth Less Equity

    This labeling change and a position shift avoids the confusing problem where Net Worth seems to equal Net Worth – Equity.

    In your words re: changing (horrors!) the GIIP presentation, “This would align the presentation somewhat more with micro financial accounting.”

    Because in the “Real World,” really, nobody is going to suggest that the net worth of all American nonfinancial corps, year-end 1999, was negative $5.2 trillion. That is only true in (warning, oxymoron alert) “real life accounting.” ;-)

  74. But, again, the issue of forced redemption: in the real world of large business, shareholders have very limited power to force a sale hence redemption. It’s a long step from a loan or bond — an obligation of the corporate entity as a going concern — coming due on a fixed date.

    Hence the difference between debt and equity instruments, yes. But equity is still a liability to the firm in that it represents net worth to the owners and not the firm

    Because in the “Real World,” really, nobody is going to suggest that the net worth of all American nonfinancial corps, year-end 1999, was negative $5.2 trillion.

    Incorrect. Their year-end 2014 net worth of is negative $1.4 trillion after you pull $22.2 trillion in value off their balance sheet and onto that of the owners. That’s the whole point.

  75. JKH writes:

    David R.,

    Thinking of it in pure monetary terms, and allowing for the fact that dividends can be eliminated voluntarily by the firm at any time without further liability for dividend payments, there is no obligation for the firm to deliver the “medium of exchange” to equity holders at any point beyond that (other than in the wind up situation we both referred to).

    Whereas there is such an obligation with debt.

  76. JKH,

    Yes, that is my point– that it may be easier to think of corporate equities as equity instruments as the firm is obligated to deliver to the owners the entire net worth of the firm whenever the so owners decide. In this sense, the owners may choose to get an advance on this amount, extracting equity value via dividend distribution. This may have particular benefit; by extracting equity from the firm balance sheet it becomes protected to the extent there is limited liability to the firm.

    Now, final payouts sometimes do not always correspond to healthy share prices because in the case of actual liquidation the net worth (plus firm equity) is probably near zero. But firms are purchased all the time, producing nontrivial equity payouts to shareholders.

  77. Tom Warner writes:

    If I had a dime for every time a journalist mixed up net financial assets with savings and implied that the world’s savings always sum to zero, well, I’d have a lot of net financial assets. So this might help. Generally though there’s a wider problem among sectoral balance analysts of not thinking comprehensively enough about how these are components or factors of a constantly changing total spending/output that’s made up of each sectors’ real consumption, physical savings (property and durable goods) and net financial balances.

    Also this leaves one very important definition issue to be settled: are central banks private or public? All the sectoral balance data I’ve seen treat central banks as private sector. I think that’s misleading and I doubt it’s even intentional – it seems just to be following the practice of national financial accounts.

  78. JW Mason writes:

    I’m on Team Roth on this one.

  79. JW Mason writes:

    the firm is obligated to deliver to the owners the entire net worth of the firm whenever the so owners decide.

    Suuuuuure they are.

  80. JW,

    the firm is obligated to deliver to the owners the entire net worth of the firm whenever the so owners decide.

    Suuuuuure they are.

    Is that sarcasm I detect?

  81. JW Mason writes:

    Yes. As I said to you in email, nobody who has the slightest familiarity with the way actual corporations operate could think that shareholders can freely liquidate the company, let alone that they would do so in response to a high share price.

    Shareholders may (1) receive a dividend, when and if the board and management choose to pay one; (2) vote on members on the board, at times and on terms determined by the board; (3) sell their shares to someone else. That’s it. Shareholders are not and never have been the “owners” of the corporation, either legally or substantively. You could read Lynn Stout on this.

    You seem to be saying that a very high share price makes it likely that a firm will end as a going concern and have its assets distributed in a bankruptcy proceeding, just like very high debt would. But I can’t believe you actually think this.

    And this brings out the fundamental difference between corporate equity and what we normally mean by “liability.” A business (or other unit) is legally obliged to make the scheduled interest and principal payments on its debt. If it fails to do so, the creditor can claim the debtor’s assets in a bankruptcy proceeding, backed by the coercive power of the state. It’s this legal obligation to pay that defines a liability. But there is no legal obligation to deliver the market value of equity. Corporations are under no obligation to make payments to shareholders of any kind.

  82. JW Mason writes:

    Reading the thread properly now I see that I’m just repeating what Steve Roth has already said.

    There’s another puzzle along the same lines, which has not been discussed as much, in the treatment of debt. When there is a fall in prevailing interest rates, there is an increase in the value of bonds on the asset side, but no corresponding increase in the value of the corresponding liabilities. It might seem contradictory that there is no capital loss for the debtors to offset the capital gain for the borrowers. But I think it actually makes sense. The debtor is under the exact obligation to make future money payments as before. But the creditor now can generate additional money income from the debt contract, by selling it to a third party at a higher price than at the old interest rates.

    And of course things are more complicated because the effect of an interest rate change on the value of a debt contract depends on the horizons of the parties as well as the terms of the debt. If the lender has a short horizon and the borrower has a long one, it’s perfectly possible that a fall in interest rates could be correctly recorded as a capital gain by both parties.

    It is true that every dollar of future income by one unit is also a dollar of future payment by sone other unit. But only some of these future money flows are capitalized on balance sheets. There is absolutely no reason to think that the exact same flows must always be capitalized as assets and as liabilities, or that they must be discounted at the same rate on both sides.

    Actually David R.’s position on equities strikes me as just another form of Ricardian equivalence. There is no more reason to think that shares will eventually be “paid off” than that government debt will, or that it constitutes a liability for corporations any more than future tax payments do for the private sector.

  83. JW Mason writes:

    Or another example. If a borrower is in distress, creditors will, correctly, mark down the value of its debt as an asset. The expected future flow of income from a distressed borrower’s debt contracts is less than it was before. But there is no reduction of the debt on the liability side — the borrower is obligated to pay just as much as before. The possibility of insolvency is capitalized into the value of the asset value, but not into the value of the liability.

    Or again, if I own an apartment building, the value of the future rental payments is capitalized into the asset value of the building. But no one records those future rent payments as liabilities, even though someone will have to make them if they want a place to live.

    The idea that for every financial asset, there must be a liability of equal value just seems like dogmatism to me. It’s making a fetish of an abstract symmetry. Sort of like the mainstream’s insistence that every economic decision has to be represented as the solution to a constrained optimization problem. And actually I think the parallel goes deeper, because the asset=liability principle really only makes sense in a world of perfect information and identical units.

    OK enough from me on this.

  84. JKH writes:

    JW Mason,

    “But there is no legal obligation to deliver the market value of equity. Corporations are under no obligation to make payments to shareholders of any kind.”

    Precisely the point I tried to bring out in recent comment. Yours is a more complete statement of the fact of the matter.

    This is very important. Godley and Lavoie treat equity as both a liability and an asset purely for purposes of analytical expedience in the consolidation of the full economy for purposes of modelling. It is a very artificial construct, and they say so on page 28 of their book.

    MMT in effect treats equity as both a liability and an asset in order to get a “clean” definition of net financial assets from a consolidated private sector perspective. This is arguably more ideologically driven than modelling driven. It is correct as far as it goes, but it ignores the vast underbelly of the financial system.

    Both of these constructs serve their purpose, but are fraught with peril if one become too short sighted about that purpose.

    The fact is that all private sector wealth is telescoped directly or indirectly into the household sector as a subset of the private sector. And that household position itself includes a massive net financial asset position that has nothing to do with government debt. This is the great underbelly of how the financial system actually works beneath the convenient construction of a “consolidated private sector”. And understanding how it works depends on a reasonable grasp of how corporate equity is treated under normal financial accounting.

  85. Ramanan writes:

    “Actually David R.’s position on equities strikes me as just another form of Ricardian equivalence. There is no more reason to think that shares will eventually be “paid off” than that government debt will, or that it constitutes a liability for corporations any more than future tax payments do for the private sector.”

    How is it Ricardian equivalence. Just check Godley/Lavoie’s book. Where is any Ricardian equivalence assumed?

    Equivalently, I could assume your definitions and write a model which has Ricardian equivalence.

  86. Ramanan writes:

    JW,

    Also about your point on equity. Just think of Twitter and the kind of pressure Wall Street puts on it. Yes they are not obliged *by law* to pay any dividend. Cool. If they announce such a thing, they’ll never be able to raise any funds, debt or equity.

  87. Ramanan writes:

    But agree with JKH’s points mostly: in G&L’s book, although one convention is assumed, it’s not as if they are stuck up on that.

    (I personally prefer it for other reasons).

    MMT on the other hand uses it as a rhetorical device, driven ideologically. Moreover, in Wynne Godley’s analysis, he used the net stock of financial assets of the (domestic) private sector and didn’t merge it with the rest of the world at all.

    Here’s an example from one of his papers in the 70s.

    http://i.imgur.com/rdHope5.png

    Clearly separated out the private sector and the external sector.

  88. Yes, it is true that shareholders may never ever see a single cent in distributions no matter how highly valued the firm. I am not saying that the firm will eventually distribute an amount equivalent to the current outstanding equity value. Obviously, that figure could rise, or fall to zero.

    I am saying that stocks are called equities for a reason. I am saying we talk about the market value of corporations. I am saying that it is improper to say that firms have value and that outstanding value belongs on stockholder balance sheets without removing that same amount from the balance sheets of the firms. That is just double counting.

    And JW’s example pains me to no end. An apartment building is not a financial asset, but physical capital. So of course it has no corresponding liability. It should not, insofar as we count it as capital. As I have mentioned elsewhere, it may be politically devious to count housing this way– legitimizing rent as productive income– but the comparison to equity is totally invalid.

  89. Steve Roth writes:

    @JW Mason: “Shareholders are not and never have been the “owners” of the corporation, either legally or substantively.”

    I’d say, rather: Shareholders’ ownership rights are quite proscribed. “Ownership” is not a monolithic state. It implies anything from a very simple to very complex set of legal rights.

    Jumping a couple of conceptual steps, this leads me to think that a key job for economists is to develop a coherent taxonomy of different types of “assets,” defined perhaps along multiple axes. (Very sociological, this approach…)

    One important criterion for classification would be “Does it have an (explicit?) offsetting liability on another balance sheet?” ie is it a “financial asset”?

    Rules for forced redemption would be another. eg, at an extreme, can an “owner” say, “gimme my share of those real assets/goods”?

    “… the fundamental difference between corporate equity and what we normally mean by “liability.” A business (or other unit) is legally obliged…”

    I really want SRW to chime in on redeemability. He’s thought and written well about this.

    @JKH: “the great underbelly of how the financial system actually works beneath the convenient construction of a “consolidated private sector”.”

    Right. Different presentations make different economic realities obvious or invisible. My Comprehensive Household Income presentation, for instance, completely hides/obscures the surplus from production/value-added generated by Firms as Households’ wholly owned subsidiary. So MMT is not alone in this — inevitably not.

    @Ramanan and JW on Ricardian equivalence:

    Just to note that SRW hit this point in comment #14:

    “if you believe in “Ricardian equivalence”, then an increase in PRIVATE_SECTOR_SURPLUS caused by a government deficit is precisely offset by a reduction in HOUSEHOLD_NONFINANCIAL_ASSETS.”

    JW: “the value of the future rental payments is capitalized into the asset value of the building. But no one records those future rent payments as liabilities”

    That strikes me as a very telling point.

    I’d really like to hear people’s thoughts about the FOFAs and IMAs presentation for firms tossing equity in as just another liability, rather than breaking it out at the bottom as in corporate financial statements..

  90. Cullen Roche writes:

    We’re seeing the real problem with MMT. What the framework relies on is redefining well known terms for the sake of this “new paradigm”. All it does is create lots of confusion.

    Shareholders equity is shareholders equity. FULL STOP. It is assets minus liabilities. Not assets minus liabilities minus more liabilities = nothing. I don’t see why we’re now calling shareholder’s equity a “liability” and canceling it out as if it’s just some zero sum game. The Flow of Funds shows SE as its own component for a specific reason – the growth of shareholder’s equity is representative of the most important balance sheet item we have in the economy. When share values rise in conjunction with their total replacement cost this is a sign that the underlying private sector’s output has become more valuable and almost certainly increased the overall living standards of the society in which those assets exist.

    MMT reconstructs this view to diminish the importance of the private sector by saying that “government debt is the equity that supports the entire global credit structure”. This is misleading at best and deceptive at worst. It is a misrepresentation of where most of the great innovative and productive output comes from in the economy. And the entire “new paradigm” portrays the private sector as being overly dependent on the govt as if we’re flailing around just waiting for govt to “spend first”, supply NFA and engage in its role as the mythical money monopolist. MMT distorts the fact that productive corporations can thrive without govt, but govt cannot thrive without productive corporations.

    This is not to say that NFA and govt debt is not important. I am an ardent advocate of higher deficits (when appropriate and in a true Keynesian sense), but we have to be careful about veering too far in the wrong direction here and falling into the trap of abusing terminology for the sake of some “new paradigm”.

  91. Ramanan writes:

    SR,

    I don’t understand your reply to me on Ricardian Equivalence.

    Simply put, it is a behavioural assumption in a model. That model is wrong but economists use it as THE model.

    It says that if the government tries to expand domestic demand by say boosting its expenditure, consumers will respond by saving more, thinking that they’ll end up paying more taxes. It’s wrong empirically mostly because the private sector, especially consumers don’t behave that way.

    One can incorporate Ricardian Equivalence in a stock-flow consistent model. Just that it’s a waste of time.

  92. Shareholders equity is shareholders equity. FULL STOP. It is assets minus liabilities. Not assets minus liabilities minus more liabilities = nothing. I don’t see why we’re now calling shareholder’s equity a “liability” and canceling it out as if it’s just some zero sum game.

    The point of the equity liability is to prevent those “assets minus liabilities” from getting counted twice, not to prevent them from getting counted.

  93. Wow. You guys really get into this.

    I’m going to add a few points.

    1) There’s no such thing as a “true” answer to how one should set up accounts. It all depends on your purpose. That sounds all fuzzy and pomo, but anyone who has ever done financial statement analysis in a serious way has realized that, in order to make things comparable across firms, you have to redo accounting according to some rather arbitrary standards and conventions that you choose, not because anybody screwed up, but because even within the norms of e.g. US GAAP there are many ways to account for things, and the raw numbers won’t be “apples to apples” unless you dig and adjust. And US GAAP isn’t written on a stone tablet. It is very much a set of arbitrary, politically influenced conventions. NIPA conventions, flow-of-funds conventions, all of these things are conventions, not at all written on stone tablets. Different sets of conventions are useful for different sorts of analysis. If we treat any of them as normative, it is because we crave norms we can rely upon to speak authoritatively, not because any set of arbitrary choices is truly best.

    2) I think that debt and equity are very different. One of the core themes of this blog over the years is to prefer equity over debt arrangements, and to look for ways that address the problems debt solves with more equity-like collaborations. That said, debt and equity are similar along some dimensions. One way they are similar is that they both are efforts to account for claims on real assets, for who owns what.

    3) Financial assets like real assets might have different values to different parties. If I face a particular risk, my willingness to pay for a financial asset that hedges that risk may be different than yours. If we are doing accounts for individual firms or households, it strikes me as potentially reasonable and useful to use firm and household-specific values. But if we are aggregating, we need comparability among the things we are summing. If we claim an asset is worth some value, I don’t see great likelihood that you’ll come up with a very useful aggregate accounting framework if you value it differently for the issuer. This is an argument for an axiom: I claim that financial assets in aggregate accounts should sum to zero, because frameworks in which they don’t are rarely useful, not because it is the only true way. There are just too many degrees of freedom. Any way you do aggregate accounts there will be what seem from an intuitive perspective to be paradoxes and irregularities.

    4) If you don’t let financial assets sum to zero in value, you are claiming that total wealth is not the same as the total value of real assets (however broadly defined). You are attributing (positive or negative) value to finance per se, to the paper arrangements that we use to describe the value of and organize claims upon real value. That might be okay! It probably is true that finance itself is a valuable asset, that how we organize paper claims has a great deal to do with how much value we are able to generate from the resources around us (in whatever way we wish to think of value). But in practice, I have never encountered an accounting scheme that for which one could reasonably argue “net financial claims” captures the value added of finance. No answer to the question most tossed around here, whether the “liability” associated with firm equity should be priced the same as the “asset” held by its owner, would generate a reasonable estimate of value-added by finance. If we claim that equity is somehow a “weaker” liability than debt and so should be valued at less than market (or that it is not a liability at all and should be omitted from any netting), then we create value in our accounting system by taking any asset, wrapping it in a corporation, and issuing equity. We can daisy chain financial assets like this ad infinitum. You get to make up your own accounting system, but I think that’s a pretty dumb one.

    5) The practice of valuing firm liabilities at some cost-based or par value and letting the discrepancy between firm and market be a measure of value-added may be defensible at a firm level or even an aggregation-of-corporates level, but not in an accounting of an economy that includes households and businesses other than joint stock companies. Surely other entities create or destroy value? (Surely humans consume!) It makes no sense to make claims about value-added or destroyed in the aggregate economy based on a measure that only sees changes in the market value of corporations. As soon as you decide that you are going to include nonbusiness assets in your accounts, it’s hard to come up with useful accounting systems that let financial claims sum to anything but zero without making difficult to justify claims about why business real assets are different. Again, difficult does not mean impossible. Perhaps the financial arrangement of formal businesses is itself a valuable asset, and so the sum of business financial claims among issuers and holders should sum to the value of those arrangements. But I have never encountered a scheme that could credibly claim to capture such a thing. Almost always, when you let financial claims sum to other than zero and let real assets have independent value, you can come up with ways of reorganizing claims that create (destroy) value in ways that cannot be plausibly described as a value-add (value-destroy) of organization.

    For what it’s worth, that’s my 2 + -2 = 0 cents on the matter.

  94. JKH writes:

    “If you don’t let financial assets sum to zero in value, you are claiming that total wealth is not the same as the total value of real assets (however broadly defined).”

    Wrong, SRW.

    That contradicts the household net worth view of wealth – a legitimate, comprehensive presentation and arguably the most realistic.

  95. Steve Randy Waldman writes:

    JKH — Well, elaborate.

    To my simple mind, if we let real assets have what we estimate as their values (however we estimate that), and if we have financial claims that don’t sum to zero (I shouldn’t have used the word “assets”, I should have said “claims”), and if we sum the real and financial together, we end up at a value for total wealth different than the value of real assets.

    I don’t know what you mean by “household net worth view of wealth”. I’m perfectly happy with accounting systems that attribute all wealth ultimately to households. But the way you get there is…

    Total Wealth = Household Wealth = Household Real Assets + Household Financial Claims on Business (1)
    Household Financial Claims on Business = Business Real Assets (2)
    Total Wealth = Household Wealth = Household Real Assets + Business Real Assets (3)

    (I’m not dealing with govt or external sectors.)

    What does that imply of business financial claims? Well, we can leave the above as it is by definition, and just skip business claims entirely. But that seems unsatisfactory somehow. If we’re trying to aggregate the whole economy and we are going to include Household Financial Claims on Business (not just skip finance entirely), then we probably want to include business financial claims in the accounting somewhere. So we claim, axiomatically

    Total Wealth = Household Wealth (4)
    Total Wealth = Household Wealth + Business Wealth (5)

    By rather trivial algebra, that implies:

    Business Wealth = 0 (6)

    Which is good, since it is the assumption we began with and that every business balance sheet enshrines. The assets of every firm are ultimately claimed by some other entity. If we take the transitive closure of the ownership graph, the assets of every business are ultimately owned by households. (Again, putting aside come complexities of the govt and foreign sector.)

    So what does this imply about business financial claims?

    Business Wealth = Business Financial Claims On Households + Business Real Assets (7)

    But we’ve already established that Business Wealth is 0. So, we know

    Business Financial Claims On Households = -Business Real Assets (8)

    Substituting equation (2) we have

    Household Financial Claims on Business = -Business Financial Claims On Households

    Financial claims sum to 0.

  96. Steve Roth writes:

    @Ramanan: “I don’t understand your reply to me on Ricardian Equivalence.”

    Well I didn’t actually reply, just pointed out that SRW had touched on it. Retweet ≠ endorsement.

    Delighted that it or whatever spurred Steve to reply.

  97. […] Translating “net financial assets” Steve Waldman. Clearing up an MMT debate. […]

  98. JKH writes:

    Steve W.,

    I mean B101 in the Z1 Fed flow of funds report (plus B101e for a direct/indirect equity breakdown)

    Here is a simplified model that is consistent with that template:
    (Assume no debt for simplicity; there is no loss in generality because of this)

    Household Assets
    = House + Corporate Equity Security

    Household Net Worth
    = Value of Assets

    Assume the price of the corporate equity is the same as the value of its only (real) asset – say a factory

    (It is the equity security and not the factory that appears in the B101 analysis, and it is an arbitrary assumption to suggest these are not equal or that the value of the factory separate from the price of the security is any better a valuation than the price of the security separate from the value of the factory)

    That covers everything
    There is no required summing to zero of financial assets and liabilities in this model
    There is only the corporate equity security, which contributes a net positive value in the analysis

  99. SR & Ramanan — I don’t in general think much of Ricardian Equivalence. I just invoked it as an example of how you have to be careful about accounting relationships, how they say less than we are tempted to think.

    We had

    ΔHOUSEHOLD_NET_WORTH = PRIVATE_SECTOR_SURPLUS + ΔHOUSEHOLD_NONFINANCIAL_ASSETS + ΔBUSINESS_NONFINANCIAL_ASSETS

    It’s tempting, then, to say that deficit spending by government increases PRIVATE_SECTOR_SURPLUS and so increases ΔHOUSEHOLD_NET_WORTH, contributing to household wealth. But, there’s obviously something unsatisfactory about that. Government could issue $1T of debt to every household, but the world’s real assets, and so households’ real wealth, would not grow so much.

    But our equation is an accounting identity. It must always be true. How do we square the circle? Well, in reality, what would likely happen is a great inflation. In dollar terms household net worth would go up a great deal, but in real terms not so much.

    But suppose that government was going to use tax policy to ensure price stability. (Let’s suppose that central bank behavior is fixed for simplicity.) Then HOUSEHOLD_NET_WORTH could not go up so much. It would be constrained by real wealth, which is largely unchanged. How does the equation balance? Well, that aggressive tax policy the government plans is a pretty serious liability, which we can characterize as nonfinancial (since there is no particular contract enshrining each formal claim) or as financial (since the tax will be paid in money). It’s an arbitrary choice. If you characterize it as financial, then it nets out within PRIVATE_SECTOR_SURPLUS, and so not so much surplus is added in the first place. If you characterize it as nonfinanical, then the addition to PRIVATE_SECTOR_SURPLUS is going to be offset by a huge impairment of household or business nonfinancial assets, a large negative ΔNONFINANCIAL_ASSETS.

    “Ricardian equivalence” claims that people will behave as if each $ deficit spent by the govt is offset one to one by a reduction of net assets, whether you categorize that reduction as financial or nonfinancial. I don’t think that’s right, but it is a possible outcome. If people did behave that way, and if we account for tax obligations as “nonfinancial”, and if you assume agents are rational and informed, then the PRIMARY_SURPLUS term would not be independent of the other terms, and naive inferences from our accounting identity would be misleading.

    Ricardian equivalence probably does not reasonably describe people’s behavior, but it is also unlikely that PRIVATE_SECTOR_SURPLUS contributes to HOUSEHOLD_NET_WORTH in the straightforward, one-to-one way that a naive look at our accounting identity might suggest. Probably in complicated and nonmonotonic ways (that will also depend upon from where the PRIVATE_SECTOR_SURPLUS derives), PRIVATE_SECTOR_SURPLUS is not independent of ΔHOUSEHOLD_NONFINANCIAL_ASSETS and ΔBUSINESS_NONFINANCIAL_ASSETS.

  100. Steve Randy Waldman writes:

    JKH — We are describing precisely the same model. Look at my equations (1), (2), and (3) and compare them to

    Household Assets
    = House + Corporate Equity Security

    Household Net Worth
    = Value of Assets

    Assume the price of the corporate equity is the same as the value of its only (real) asset – say a factory

    They say precisely the same things.

    Financial positions don’t sum to zero only because we choose not to consider financial positions held by businesses. No one suggests that financial positions of households alone sum to zero!

    But the model you describe requires as a matter of arithmetic that financial positions sum to zero if we include business financial positions with respect to households as well. There can be nothing of the form, “households value equity of (all equity) Firm X at $50M, but Firm X’s position with respect to households is worth only -$30M, because equity claims impose a weaker burden on the firm than debt would have.” We can consider household positions alone, in which case financial positions will be large and positive. Or we can aggregate all financial positions, including those of businesses with respect to households, and they will sum precisely to zero.

  101. Ramanan writes:

    SRW,

    Didn’t catch your argument fully.

    There are several things. Neoclassicals argue that deficit takes away funds from the private sector. The accounting identity is invoked to show that the opposite is the case.

    Of course one can still argue that one shouldn’t argue with accounting identities but the same logic holds for those saying the government takes away funds via deficit spending.

    So one needs a behavioural model to back the story told by accounting identities.

    About your point on “real” vs “nominal” … Godley and Lavoie do create a system of accounts with inflation-adjustments and behavioural model around it … and the models show that fiscal policy is far from neutral as argued by neoclassical economists.

  102. Steve Randy Waldman writes:

    Ramanan — The “crowding out” argument is obviously, empirically, unreliable. Argument from naive interpretation of quantities in accounting identities as independent variables is also obviously unreliable. We are left always requiring a model.

    The point of the headline post here was to (informally) attach a model to the claims made by MMTers, which I think do have stylized empirical validity. I’m sure Godley and Lavoie’s model is much much better though!

  103. JKH writes:

    Steve W.,

    “But the model you describe requires as a matter of arithmetic that financial positions sum to zero if we include business financial positions with respect to households as well.”

    That makes absolutely no sense.

    Have a look at B101 and B101e.

    There are no business real assets in that model.

    That’s the whole point.

    The business sector is valued according to its financial capital structure.

    And that financial capital structure is part of the asset structure of households, directly and indirectly.

    So it is simply wrong as a matter of consistent methodology to include the business sector financial capital structure as a negative item in the valuation of the household sector’s interface with the business structure. It is included one time – as a positive item.

    Again, that’s the whole point of this.

    Valuation is done via the business financial capital structure rather than its real capital structure.

    And the result is just as legitimate a valuation of the consolidated private sector as the GL et al methodolgy.

    (E.g. it is the oil industry real asset values that are catching up to the oil industry equity and junk bond valuations as much as the other way around.)

  104. Ramanan writes:

    SR,

    Cool.

  105. Ramanan writes:

    SRW,

    Agree.

    And I also said in one of my comments about “NFA”. It’s not like it’s something like an interest rate to be set or something. The net financial assets of the private sector depends on many things such as their propensities to consume etc.

    In fact I came across this paper https://www.postkeynesian.net/downloads/Dafermos/YD200514.pdf

    who has Minskyian sort of behaviour and NFA should fluctuate a lot.

  106. There is only the corporate equity security, which contributes a net positive value in the analysis

    Right, because your analysis disappears the real asset which underlies the financial one. “B101 analysis” means the factory does not show up anywhere. Which is my problem with Steve Roth’s general analysis of saving.

  107. Ramanan writes:

    “if you believe in “Ricardian equivalence”, then an increase in PRIVATE_SECTOR_SURPLUS caused by a government deficit is precisely offset by a reduction in HOUSEHOLD_NONFINANCIAL_ASSETS.”

    I’ll just say that the accounting identities are all the way they are. But consumption function depends inversely on Government expenditure or something (if we assume Ricardian equivalence).

    So we might have a situation in which increase in government expenditure not able to do much in that model but public debt and deficit rising rapidly (if the government attempts to raise demand and fails because of Ricardian equivalence).

    The interpretation of identities and accounting etc won’t change. Just how stocks and flows move in time.

  108. Ramanan writes:

    I think the general line of argument here is … accounting identities are fine. Ricardian Equivalence is wrong but it points out that simply arguing from accounting identities won’t do.

    Is that your point SRW. Agree on that.

  109. Marko writes:

    You can prove to yourself that the household net worth valuation described by JKH is equivalent ( not precisely so , though ) to a valuation restricted to tangible assets by comparing the consolidated net worth of all sectors in the new B.1 “U.S. Net Wealth” measure to the equivalent measure you’d obtain by adding the standard FOF HH net worth to the combined state/local and national gov’t net worths.

    Here’s the B1 table :

    http://www.federalreserve.gov/Releases/z1/current/accessible/b1.htm

    Here’s the equivalent FRED graph using the combined HH and Gov’t values :

    https://research.stlouisfed.org/fred2/graph/?g=22Yh

    The values agree pretty well , and since the vast majority of the total is HH net worth , this implies that the two HH methods would correspond. Close enough for gov’t work , I’d say. The 2015 Q2 values of national wealth show the biggest difference , at $79,690 billion in the B.1 vs $81,609 billion using the FRED combo method.

  110. I think there is significant confusion as to what the net worth means. It seems like some folks think that a firm with zero net assets (after equity liability) is not worth anything. This is wrong. The net assets (before equity liability) are still valuable, and there are intangibles (e.g., power to exploit workers?) which ought to add value. The equity liability means everything of market value related to the firm is ultimately owned by the owners through the firm and not actually by the firm. No matter how much control the managers/officers/board of the firm may have, they are still playing in large part with other people’s capital.

    That is, zero net assets (after equity) says the firm ultimately owns nothing itself– not that ownership of the firm ultimately is worth nothing.

  111. Ramanan writes:

    Hey Marko,

    Didn’t know about this new table B.1

    Just checked the intro and it says it’s a new table.

    I’ll remove Line 13 in my preferred way. That’s $30 trillion. The reason it happens is because the flow of funds doesn’t count equities in liabilities.

    With one qualification though.

    It counts equities held by foreigners as liabilities of the United States. That’s why it has the net IIP in line 24.

  112. Marko writes:

    Ramanan ,

    Here’s the Fed Note descibing the new series :

    “Derivation of U.S. net wealth table: A new table on the derivation of U.S. net wealth (table B.1) has been added to the summary section of the “Financial Accounts.” The calculation of U.S. net wealth includes the value of nonfinancial assets (real estate, equipment, intellectual property products, consumer durables, and inventories) held by households and nonprofit organizations and noncorporate businesses. For the federal government and state and local governments sectors, only structures, equipment, and intellectual property products are included; values for land and nonproduced nonfinancial assets are not available. The measure of U.S. net wealth also includes the market value of domestic nonfinancial and financial corporations, and is adjusted to reflect net U.S. financial claims on the rest of the world. This definition of U.S. net wealth differs from the sum of the net worth of sectors shown in the Integrated Macroeconomic Accounts (IMA). A forthcoming FEDS Note will provide additional information.”

    from :

    http://www.federalreserve.gov/apps/fof/fofhighlight.aspx

    ( line 7 )

  113. Marko writes:

    Also , here’s the Z1 B.1 table in pdf form , which is a little better than the table linked above as it includes the series identifiers for each datum and the footnotes are appropriately linked :

    http://www.federalreserve.gov/apps/FOF/Guide/P_6_coded.pdf

  114. Ramanan writes:

    Thanks Marco.

    I would have thought the government had more non-financial assets than that but $13 trillion is not bad.

  115. Peter writes:

    @SRW

    I may be way off base here (trying to get into the mind of various people in this disagreement… always a dangerous proposition) but…

    I think the issue is a disconnect in your equation:

    Household Financial Claims on Business = Business Real Assets (2)

    If all Business Real Assets are fully capitalized to incorporate all future cash flows, then this is what you would see. Rarely in the “real world” of balance sheets do you see this, though.

    I think this is what JW Mason is getting at with the distressed debt example, where you don’t see a reduction in a company liability even if you do see a reduction in a creditor asset.

    Either you have to subscribe to the view that:
    (A) All financial assets/liabilities net — which would mean some weird balance sheet accounting not often encountered in the wild, e.g. businesses marking up/down assets and liabilities based on business conditions, or
    (B) you accept that you’ve got financials claims “left over” — yes, this would lead to your scenario of creating value by simply dropping an asset into a corporation and having the market value of the equity appreciate w/o marking up the asset as well (per your point 4 above)

    Thoughts? Helpful? Crazy?

  116. Peter writes:

    Also, I just realized that I’ve had this open too long on my browser and my comment (115) seems hopelessly late to be of value.

    For what it’s worth, it’s a response to comment (95) with a reference to comment (93).

  117. Peter

    I think you’ve fallen into that same trap I pointed out earlier.

    The “weird” accounting is not found “in the wild” because it is not important for those to be macro-consistent. It is not like households report balance sheets which might conflict with those corporate balance sheets.

    In the wild: “Hey, this says something about what the firm might be worth to shareholders!”
    IMA: “This is what is left once the owners claim market value from the firm”

    Macro-consistency means we do not want to record market value of the firm on the household sheets and the “wild” value on the corporate sheets because these are more or less different measures of the same thing. We do not want to double-count the value of the firm. So we could just leave domestic equities off the owners’ balance sheets and use “wild” net worth, but you’d be left with an entire mess trying to figure out households’ indirect holdings. Much better to simply record the appropriate market values on the owners’ sheets and leave any remaining value on the corporate.

    This does not mean that firms do or ought to make decisions based on IMA net worth. It certainly does not mean shareholders and potential shareholders do or ought to price based on IMA net worth. It simply says that if we want the wealth implicit in the outstanding market value to fall on household balance sheets, then macro-consistency requires it to not appear the firm balance sheets.

  118. Peter,

    Ditto that, I guess.

  119. Maybe one last attempt for the time being.

    IMA corporate net worth is essentially, “firm net worth, n.e.c.”

  120. Steve Roth writes:

    @Marco:

    Thank you for noticing the new B.1. What a great addition.

    Amazing: it’s been more than eight decades since Simon Kuznets started slaving away on this, and we finally received this table on September 18.

    I can’t resist pointing out that this table is inescapably, inevitably based on Haig-Simons accounting.

    US net wealth does not equal household net worth for obvious reasons: HH doesn’t include gov, and the HH sector has different “external” liabilities than US to ROW.

    Would love to work the math of adding up private sector net worth (a measure that would be “true” and = HH net worth by construction), but: while it’s easy to add up the assets here (lines 2, 7, 12, and 13), can’t just add up liabilities from each sector’s BSes because most of those liabilities are among/between those four sectors, not “external” sectors — gov and row. Thoughts?

  121. Marko writes:

    SR ,

    I posted a note about the B.1 on the evening the last Z.1 came out , on a site called “Angry Bear” :

    http://angrybearblog.com/2015/09/open-thread-sept-18-2015.html#comment-2676609

    It’s a good site. You should check it out. ;)

    Like you I’d like to see the comparable private sector breakdown , but I think the differences from the current Z1 measures wouldn’t be great , as I noted above. The offshore “hidden wealth” as described by Zucman is more in need of a proper accounting at the moment , I think.

  122. Steve Randy Waldman writes:

    JKH — We’re really not disagreeing about anything. You just think we are.

    Table B101 of the flow of funds shows a household & nonprofit sector balance sheet. We all agree that the financial assets of the household sector don’t sum to zero. Summing to zero only happens when you aggregate the positions of both holders and issuers of financial assets. B101 doesn’t do that.

    The flow of funds treats the financial position of owners and issuers symmetrically. Both are measured at market value, which means by definition, they must sum to zero (assuming you attach a positive sign to holders of assets and a negative sign to issuers).

    It’s not so easy to verify this, but we can see that things approximately work out. Look at the September 18 report, table B101, lines 25 and 26, corporate equities and mutual fund shares. Sum the latest values, and you get 21758. Now go to table B103 nonfinancial corporate net worth (market value). It’s 21449.

    This is a surprisingly good match. Mutual funds can hold things other than corporate equities. Households hold shares of financials (which these tables look through). Equities might be held in other categories, pension funds, life insurance reserves, etc. This is actually smaller a discrepancy than I’d expect from such a crude eyeballing. But it is not a coincidence. These values track one another very closely throughout the time period shown.

    Now look at B101 line 29, equity in noncorporate business. Compare it with the net worth line of B104, nonfinancial noncorporate business. Again, you’ll see they track closely.

    The Flow of Funds, like any reasonable full economy aggregation, adopts the convention that the value of a financial asset to its holder (households here) is the mirror image of its right-hand-side-of-the-balance-sheet value of the issuer. Again, it’s the only sane way to do things, and the Flow of Funds is sane.

    Note that if you are NOT aggregating the full economy, that is, if you are not accounting for both sides of financial positions, it’s perfectly sane to do things differently. Individual businesses sometime value assets at cost, sometimes at market value, and accounting conventions include all different kinds of rules about that. For some purposes, liabilities are marked at principal value, for other purposes they fluctuate with market values. That’s all fine. When you are analyzing a business, there are lots of reasons to make lots of choices to answer different questions.

    The same would be true of an isolated accounting of a sector. We could aggregate business sector balance sheets and adopt the convention that assets are valued at least of cost or market (as they most commonly are in conventional statements). Or we could try to value assets at market, as the flow of funds tries to do. Aggregate net worth based on least of cost or market would be much lower than the market value of equity. (It would be close to the sum of book values). As has come up in this thread and over at Steve Roth’s, the difference between cost-based value and market value may be an interesting sort of value add measure. But it would not be an appropriate thing to sum against household positions in a computation of net worth. (The value at cost ought to be part of household net worth. Valuing the net household financial position at the difference between cost-based and market value either excludes the cost basis, which is still owned by households, or else double counts the value add if business real assets are also included.)

    There is nothing wrong with, a whole lot right with really, sector-specific accountings in which aggregate sectoral financial positions are nonzero. But its very hard to come up with a reasonable accounting system of the aggregate economy, in which all sectors, including financial positions of both holders and issuers of assets, are aggregated to some value other than zero. I won’t say it’s impossible. You could for example define all claims against business as 0 and then let total value of financial claims against business stand in for business real assets. You might justify that choice on the theory that financial claims are easier to value, so this procedure gives better results. But you can’t use that trick to value nonfinancial assets of households, so you really haven’t solved any problem, and setting liabilities of household-on-household financial claims to zero would yield the absurd result that lending money to your cousin creates household net worth.

    So, yes, a for a household sector aggregate balance sheet, financial positions are large and positive, and that’s great. But I have yet to see a total aggregation in which it doesn’t make sense to let financial claims net to zero. (More accurately, I’ve never encountered a proposal for such a thing where you wouldn’t run into indefensible ways of altering important values by reorganizing claims.) The flow of funds is yet another example of an exercise in which financial claims across all sectors would net to zero, absent statistical discrepency. It is not any kind of counterexample.

  123. Steve Roth writes:

    @Marko: “I’d like to see the comparable private sector breakdown , but I think the differences from the current Z1 measures wouldn’t be great”

    So you’re saying HH net worth would roughly equal private-sector net worth? I think so too. Oddball that I am, I’d just like to work through the math problem and see it happen. Helps me internalize the accounting so it’s really intuitive. But I don’t know how to consolidate private-sector liabilities — how figure out, for instance, how much HHs owe to gov and ROW.

    Yes re: Zucman and offshore. A hundred times Yes.

  124. Steve Roth writes:

    @Ramanan: “I’ll remove Line 13 in my preferred way. That’s $30 trillion.”

    I don’t understand. The value of firms isn’t reflected anywhere else on B1. Are you saying 1. American firms are worth nothing, form no portion of American wealth, or 2. They should be valued at nothing whatever their value because it allows for closed-loop accounting? Or…??

    “The reason it happens is because the flow of funds doesn’t count equities in liabilities.”

    Corporations don’t include shareholder equity in liabilities either. That’s why they label the right side of their balance sheets “Liabilities and Shareholder Equity.” Is that an accounting error on their part?

  125. Steve Randy Waldman writes:

    Peter —

    There is no such thing as the true worth of a firm, or of a household, or the household sector, or the world. We impose conventions upon the world to produce accounting estimates. What we can hope for is a kind of consistency (or nonabsurdity), and that the values that we end up are meaningful to address some kind of question that we’re interested in.

    We can compute some kind of aggregate net worth based on conservative GAAP corporate balance sheets. We can do the same using “fully capitalized” estimates, which is what market value balance sheets attempt to capture. But whichever value we pick, we have to stick to it. If we intentionally use conservative values (per GAAP) for business assets, then we are making a claim that, for the purpose of our exercise, that’s how we want to measure business assets. That usually becomes the value of business real assets by definition. We don’t have to do that. We can do anything we want. We could claim to have an independent estimate of real business asset value, a book value, and track the difference. But whatever we do, the value of the claims issued by firms and the value of claims to their holders will need to be consistent. Suppose we oddly decide to value firm real assets as (book value) + (delta to estimate). How shall we evaluate the value of claims against those firms held by households then? We could set them at book value, in which case we’d have to add some kind of shadow or nonfinacial claim to account for the delta. Or we could inflate the equity claims to account for the delta on household balance sheets. Then equity wouldn’t net properly, right? Households would have total value while firms “liability” would be only at book. Households would show a net positive position of financial assets! Well, yes, but it wouldn’t be very meaningful, it would actually be the value of the (delta to estimate) and you’d have to include the value at cost, and only the value at cost, as a nonfinancial asset to reconcile what you are calling the total value of the business with the holdings of households. If you are going to have household “look through” finance to hold the fraction of assets at cost, why wouldn’t you have them look through to hold total business assets, however you want to estimate that? Again, it’s just very hard to come up with any reason or scenario why, in a fully aggregated all sector accounting, it’s useful to value opposite sides of a financial arrangement differently.

    That is true despite the fact that the parties to any financial arrangement typically value them differently! If a firm issues stock and I buy it, the firm’s managers has decided my liquidity is worth more to the firm than their stock, and I have decided the stock is worth more to me than liquidity! The same is true of impaired debt: if a firm has an illiquid liability whose market value is low, sure the burden to the issuer is more than the value to the holder. But summing those two different values is useless. The value of the impaired debt is no less than zero to the household that ultimately holds it. If I add the market value and subtract the firm’s accounting value at par, I create a value less than zero, a subtraction from household wealth, which is meaningless.

    A better way of thinking about it is that, when you aggregate all claimants, all claimants are equity holders. The capital structure is going to disappear, except in the shadow it potentially leaves on real value by virtue of tax benefits, near-insolvency/illiquidty costs, etc. So the capital structure affects the real value of the firm (contra Modigliani & Miller), but then the real value of the firm is what is owned by households in aggregate.

    A minor note on JW’s point re impaired debt: Firms do actually treat debt at market value for some purposes. Again, accounting is an exercise in discretion, not a thing handed down from God. For the purpose of computing their capital positions, financial firms during the crisis sometimes valued their own liabilities at market, so that their positions “improved” as their debt became more impaired. Was that reasonable? Well, it was regulator-acceptable, and that’s what really mattered to them. It’s easy to mock, but even in terms of economic reality, a case can be made for the practice depending upon the details of the debt. If the debt is marketable, firms can actually realize the gain implicit in the impairment of their liabilities by repurchasing the debt cheaply. (Of course you might ask, if they are liquid enough to do that, why is there debt so impaired? But it has definitely been done!)

  126. Steve Randy Waldman writes:

    Ramanan — We are pretty completely in agreement I think. The accounting identity is what it is. Ricardian equivalence is just a (counterfactual) example of how imposing an assumption of independence between variables (an assumption not remotely inherent in the accounting identity) might lead one astray.

  127. JKH writes:

    Steve W.,

    I’m aware of the nature of those connections you point out, and I think we’re talking past each other.

    Perhaps it is likely that what is obvious to me is obvious to you, and vice versa. But perhaps it is not so obvious that I’m making a legitimate point of distinction, so let’s make the example even simpler.

    Suppose the B101 household net worth representation consists of a corporate equity as an asset and net worth equal to the value of the corporate equity.

    This model of household sector net worth cannot logically be construed as the “household sector alone” as you refer to it. That is the crux of the point I’m making.

    That’s because there is an inter-sectoral interface with the business sector.

    The full value of the business sector is transmitted to the household in the form of the equity claim.

    So the household net worth model “incorporates” the entire value of the business sector as well.

    In this calculation, there is no reliance whatsoever on depicting the corporate equity as a negative financial asset, as you suggest.

    Which is what I read you to say when I responded with “wrong”.

    And the reason there is no reliance is that this model does not incorporate the real asset value of the business sector directly – i.e. the factory in my first iteration of this. As is the case with the other table connections you point to.

    What you are saying is that you can’t aggregate without attaching both positive and negative values to financial representations with a zero sum. That is true for a model that concerns itself with real business assets in a direct way, but it is not true in the case of this household net worht model.

    Other tables to which you refer show a connection of this B101 model to models that do incorporate real business assets, but the connections between these two different types of models are not a feature of the B101 representation of household net worth.

    This is in the nature of the model construction in the case of B101, which aggregates the private sector without looking at real business assets or their values directly. If you’re saying that other representations exist in which the financial asset values depicted in B101 are depicted in effect as negative values in juxtaposition with real asset values, then I would say that is obvious as well. But that is an outside connection to B101 rather than a feature of B101. B101 depends directly on financial capital structure and financial markets – not on the real business assets that connect to B101 financial values in other representations.

  128. Ramanan writes:

    Steve R,

    That’s one of the discussions above. See JKH’s comments.

    In the SNA and in Godley/Lavoie models, equity securities is in the liabilities and the market value of equities is used.

    It’s true also in IMA. See Marko’s comment:

    “This definition of U.S. net wealth differs from the sum of the net worth of sectors shown in the Integrated Macroeconomic Accounts (IMA). A forthcoming FEDS Note will provide additional information.”

  129. Paul Boisvert writes:

    @ J.W. Mason (back at comments 82 and 83)

    Josh, I think you’re comparing apples to oranges: you compare the real value of the asset to the nominal value of the debt/liability. But if interest rates (a proxy for inflation) drop for bonds, the real (future) value of the bond-issuers’ debt goes up. In real terms, assets still equal liabilities. (For financial assets, that is; the separate critique above of your “landlord” example as being about physical capital, not financial assets, seems correct to me.)

    And if a borrower’s “distress” leads to a real discounting of the value of the asset on the lender’s balance sheet, so too does the “real” value of the debt go down–the lender will now be compelled (rationally) to settle the debt at a discount proportional to the probability of default. The borrower’s (real) (expected value of the) amount she will happily pay off the debt for has gone down. Again, real liabilities match real assets.

    You mention that in reality, there may be different expected time horizons for financial instruments on the part of borrowers and lenders, so some arbitrage of “gains”, perhaps even considered “mutual” by both parties, may occur. Sure, but that’s true of all transactions–Joe won’t sell Sally a car unless each thinks it’s a good deal for him/herself, given their expectations of their needs (for money or transportation) over whatever time-horizon they privately envision (and which they rarely share with each other.)

    This doesn’t alter the fact that when a transaction actually takes place, “clearing” the market for that particular asset/liability, the expectations will by definition have collapsed into a “real”, “instantaneous” balance–including all relevant time-horizons, but now discounted by whatever real circumstances of vitality or distress are manifest. So, when the pedal hits the metal, and the deal is signed, real assets will equal real liabilities–which means that rational financial actors will always be “currently” valuing assets and liabilities by imagining a thought experiment in which they are actually sitting behind the altar, er, I mean desk, signing an actual deal. The envisioned real values of assets and liabilities in that thought experiment are what will (generally, absent incurable mental illness) match at a clearing price, as I understand MMT. This is by no means to imply, of course, that capitalism is not a form of (curable, one hopes) mental illness… :)

  130. JKH writes:

    Steve W.,

    Having reached comment 130 within a blizzard of MMT/Roth/Waldman inter-connected sub-issues, I just wanted to make an entirely separate observation that is unique to your post.

    “Net financial assets are special, because they serve insurance functions that assets produced by the domestic private sector simply cannot provide.”

    This is actually not quite right.

    Net financial assets are unambiguously not special in this way. And as you know, this is not core message broadcast by MMT about NFA.

    Assume for simplicity an economy that is either closed or balanced with respect to the international books.

    Suppose the government runs a cumulative balanced budget. That means NFA = 0 in the MMT sense.

    Now suppose it issues debt and parks the money in its array of commercial bank deposit accounts.

    NFA is still 0.

    But the insurance function otherwise provided by NFA is present without NFA.

    The insurance function in question is basically orthogonal to the core NFA function, the latter which from an MMT perspective is required to top up income and saving from income in the private sector.

    The core NFA function is associated with its creation by deficits. That is the MMT message as it relates to the issue of aggregate demand and desired savings. But insurance can occur with or without deficits.

    A real world example of this is the case of TARP. The superficial government accounting for deficits was probably wrong for this, but in fact TARP provided 0 effective NFA for the private sector. It was just an asset swap. Yet it did provide the insurance function, because TARP was funded at the margin with government debt.

    Deficits and NFA are sufficient for the insurance function. But they are not necessary.

    Which is why NFA is not special in the sense of insurance.

    This is not entirely an academic analytical point. At one point, Greenspan was concerned about this particular problem in the context of the emerging Clinton surpluses.

  131. Ramanan writes:

    Good example JKH.

  132. Steve Roth writes:

    @Ramanan et al:

    RE: “This definition of U.S. net wealth differs from the sum of the net worth of sectors shown in the Integrated Macroeconomic Accounts (IMA). A forthcoming FEDS Note will provide additional information.”

    Folks might be interested to know that I queried my BEA contacts about the revaluation methods last October (them because I had contacts there). Asked if there was a writeup on it. (I couldn’t find one.)

    They offered to forward it to the Fed, and offered this response:

    “conceptually, the revaluation account figures reflect the holdings gains/losses for the different asset categories and are calculated using changes in market rates/current prices from period to period. For fixed assets such as structures, software, and equipment, much of the price information used to inform the models comes from the consumer price index and producer price index surveys.”

    And from the Fed:

    “Currently, we do not have a simple paper that outlines the revaluation calculations. These are calculated in a multitude of ways, including using price indexes to estimate gains, backing them out from changes in positions and flows, etc.”

    I look forward to the FEDS note.

    I find it very interesting that they choose to use the term “wealth” in B1, a term that to my knowledge has never entered the accounts lexicon before.

  133. Ramanan writes:

    SR,

    That’s not true. Check any national accounts guide and you’ll see the word wealth. Even the Fed Z.1 has the word in other places.

    About your point on that IMA quote (FEDS note): That’s a different point altogether. It has little to do with revaluations. It’s about whether to put equities in liabilities or not.

  134. Steve Roth writes:

    @JKH: “Deficits and NFA are sufficient for the insurance function. But they are not necessary.”

    God you took the words right out of my mouth (thoughts).

    “Suppose the government runs a cumulative balanced budget. That means NFA = 0 in the MMT sense.

    Now suppose it issues debt and parks the money in its array of commercial bank deposit accounts.”

    For gov to issue debt, there must be private-sector bank deposits to swap for newly-issued bonds. But because of the cumulative balanced budget, NFA = 0, there should not be any.

    But there are, because private lending has created has created them ab nihilo (notwithstanding that it has also created matching liabilities).

    It’s government’s creation of assets — whether it’s 1. deposits created directly by deficit spending sans borrowing (+NFAs), or 2. bonds created and issued in exchange for other assets (no +NFAs) — that provides that insurance function. Full faith and credit and all that, adding to or replacing more tenuous private-sector-backed assets.

    (In practice, treasury never does #1 absent #2, cause they’re not allowed to except via Fed end-runs/”monetizing the debt.”)

    I would only add that market revaluation has also created balance-sheet assets (with no associated liabilities), ab nihilo. Unambiguous +NFA (even for equities, if you don’t post equity to corps as liabilities; but certainly for RE and other assets).

    And that’s where I question the “MMT message as it relates to the issue of aggregate demand and desired savings”.

  135. Peter writes:

    @SRW (& David)

    Totally agree with basically everything you say in comment (125). It definitely IS absurd to add across two different accounting schemes and it felt (at least to me!) like that was part of misunderstandings in the ~100 comments before mine, with both sides talking past each other on that point.

    If you go into the national accounts (or really anywhere) to do work on proving something using accounting “identities” it is pretty darn important to not mix and match market and book (and understand which type of accounting sources are using / be relatively specific about what assumptions you’re carrying into your analysis).

    Also, a minor note on your minor note — as someone who isn’t an economist but actually trades debt for a living, the weirdness that occasionally happens in GAAP accounting was actually the only reason I felt like I might have something to add here! Credit Valuation Adjustments are definitely strange (but mandated) and, for what it’s worth, buying back your own debt below par is surprisingly common.

  136. Steve Roth writes:

    Just to say, it seems like it would be easy to add some “look-throughs” to the HH balance sheet. Desirable? I’ll leave that to others.

    For instance:

    Corporate Equities
    Book value
    +Market premium
    =Market value

    Or in the revaluation account, change in:

    Corporate equities
    Due to increase in book value
    Due to inflation
    Pure real market revaluation

    These are like addenda, or windows out of the account, making visible the wider world beyond — they don’t “go” anywhere else in the HH account. But they do provide a look-through to underlying economic mechanisms. Easy to imagine a user interface with buttons letting you expand or collapse these look-throughs.

  137. JW Mason writes:

    SRW:

    if we are aggregating, we need comparability among the things we are summing. If we claim an asset is worth some value, I don’t see great likelihood that you’ll come up with a very useful aggregate accounting framework if you value it differently for the issuer. This is an argument for an axiom: I claim that financial assets in aggregate accounts should sum to zero, because frameworks in which they don’t are rarely useful

    This sounds reasonable but I don’t think it is. Because we are not (only) trying to come up with an ideal set of accounts to use in constructing a formal model of the economy, we are trying to organize the economic data we actually observe “in the wild.” And the only source of that data is the accounts kept by private units. You might believe (I still don’t really see why) that no asset should appear on anyone’s balance sheet without an equal-value liability on someone else’s balance sheet. But the only way we actually measure anything, is by looking at the entire that actually are on private balance sheets, and they don’t follow this rule.

    As we all know, the financial accounts prepared by the fed for the US do NOT follow your axiom that total financial asset must equal total liabilities. Are they useless? Are they strictly less useful than any other accounts that do follow your axiom? Obviously not. So you do don’t really believe your axiom. At best, it is not a hard rule but one desirable property of a system of accounts that has to be weighed against others.

    f you don’t let financial assets sum to zero in value, you are claiming that total wealth is not the same as the total value of real assets (however broadly defined).

    But there’s no sharp line between “real” and “financial” assets. An asset is just a property title that gives the holder certain legal rights and which is expected to yield a certain money income. We could say that the “real” assets are the tangible assets recorded on a business’s books, at their historical cost or at their replacement or market value. Or we could say that the “real” asset is the business itself, valued by the market value of its equity. If we really think that corporations are simply property of their shareholders, the latter approach (as in Piketty) seems more sensible than the IMA approach preferred by Godley and Lavoie and various folks in this thread. The IMA approach says that when the value of shares rise, that should be treated as an increase in the corporation’s future payments to its shareholders, but not in the increase in its future earnings. I can’t see any justification for that inconsistency. Much better to look only at the net worth of the household sector. If you think share prices are informative about future payments flows, then add them to both corporate earnings and corporate payouts, and say they reflect unmeasured real assets. If you think they are not informative about future payments flows, then ignore them and just add corporate assets less debt to household net worth. But the IMA approach of treating them as informative about corporate payments to shareholders but not informative about corporate earnings, makes no sense.

    Shorter: Measures of net worth we actually observe do not in fact equal our measures of real wealth. So we can impute unobserved financial assets/liabilities, or we can assume that the discrepancy means we have incorrectly measured real wealth. (Which is not so unthinkable, is it?) Or we can accept that preserving this particular symmetry is less important that preserving the correspondence of our aggregate accounts with the private accounts that actually shape economic behavior.

  138. As we all know, the financial accounts prepared by the fed for the US do NOT follow your axiom that total financial asset must equal total liabilities.

    We all know this? Funny, considering the IMAs are part of that release by the Fed.

    The IMA approach says that when the value of shares rise, that should be treated as an increase in the corporation’s future payments to its shareholders, but not in the increase in its future earnings.

    Wrong. It the increase in earnings shows up on the owners’ balance sheets.

  139. I think, JW, you are confusing stocks and flows.

  140. JKH writes:

    Steve R. # 134

    Very good

    Had you been looking for a moment to go out on top, that might have been it

    But its too late now

    :)

  141. Ramanan writes:

    Just wrote this about the new table B.1 in Z.1

    http://www.concertedaction.com/2015/10/07/united-states-net-wealth/

  142. JKH writes:

    Nice post Ramanan

    Although I think you may be confusing sector analyst Steve Randy Waldman with the late great guitarist Stevie Ray Vaughn

    The net result is OK though

    :)

  143. Ramanan writes:

    :-)

  144. JKH writes:

    Somebody else may want to try and reconcile the data on this, but I think there’s a very interesting relationship between household net worth and the B.1 national net wealth calculation that is quite pertinent to the subject of this post – which is the interpretation of MMT NFA.

    I think Ramanan has covered at least some of this in his post, but here are my first thoughts:

    Without checking the data (or second checking my own thought process for that matter), I suspect and expect the reason that household net worth exceeds national wealth is roughly because of the MMT private sector NFA factor.

    As previously discussed, NFA more or less telescopes down from the private sector to the household sector, in the form of direct and indirect holdings of the government debt that mirrors the cumulative generation of government deficits. This is the marginal financial asset that in MMT land has filled some of the gap in private sector (and in turn household sector) saving desires.

    On the other hand, the national net wealth calculation ignores government debt in the aggregation. In effect, it treats government debt as a liability or negative financial asset – along the lines of the SRW et al paradigm. That negative contribution cancels out against the corresponding positive contribution from the asset holding side.

    But this is only the case for the government interface. National net wealth still includes private sector financial assets as a net contribution, along the lines of the household net worth calculation.

    The interesting thing is that the private sector NFA perspective is quite at odds with the national wealth calculation. This is because the national wealth calculation implies the presence of a constituent government balance sheet that features real assets on the LHS and government debt on the RHS.

    The profile of this government position that is embedded in the national net wealth presentation is actually that of a government sector net worth profile. And that net worth profile is in contradistinction to the debt only profile as seen through the MMT NFA lens. The NFA asset that is contributed ON A NET BASIS to the private sector view is very different to the NET WORTH that is contributed on a net basis to the national net wealth profile. It is a different data set in total.

    Another way of saying this is that the MMT NFA view is agnostic with respect to the existence or non-existence of real assets held by the government. That real asset side does not matter to the MMT message. But it does matter to the national net wealth calculation.

    Go figure this difference in the context of MMT messaging on aggregate demand and saving desires and the role of the government in all of this.

    Those are thoughts off the top. I’ll leave it there because this is the first time I’ve seen B.1 and I haven’t explored the data yet.

  145. Let me try again to clear something up.

    I certainly hope we can agree that If the net worth (before equity liability) of a firm is $1 trillion, and the market value of the firm is $1.1 trillion, then we definitely do not want to say that the firm is “worth” $2.1 trillion.

    IMAs treat the market value of nonfinancial assets as the ultimate aggregate net worth, so it makes sense for the IMAs to distribute that $1 trillion among the various balance sheets. If we put $1.1 trillion of that $1 trillion on that of owners, then that leaves the firm with -$100 million.

    Another may argue that the ultimate net worth of the firm is its market value– or $1.1 trillion. Once we put the $1.1 trillion on the balance sheets of owners, that leaves the firm with nothing. This would make the IMAs look funny, as the firm would be missing from its balance sheet $100 million in some sort of intangible and only indirectly marketable net worth. On the other hand, this solves JW’s concern, because an increase in the stock price increases the intangible net worth and the liability for the firm while also increasing the net worth of the owners.

    There is nothing wrong per se with either, though I definitely favor the IMA approach. In either case, however, there is an equity liability of $1.1 trillion.

    Now, I suppose one might say this makes things simple by eliminating the need for corporate balance sheets in the macro accounting (firms would still need to report in the ordinary way, obviously.) So we could simply define stock equity an asset with no liability. But I think this is hugely misleading– not only confusing the actual accounting– but also practically. If the equity value of a company goes up, this might just be revaluation or it might reflect real investment and/or net acquisition of financial assets.

    Incidentally, that is the real problem with Steve Roth’s proposal. It is not easy at all to work out the look-through because different sectors have different equity holdings. Not everyone’s $100 of revaluation will break down the same way. Figuring out how to distribute such to households versus foreign owners would be nearly impossible. It is much more important to recognize that these things do lie behind the veil of equity ownership.

  146. Marko writes:

    Ramanan ,

    Regarding your blog post on B.1 , I believe there’s double-counting of owners’ equity in noncorporate business using the summing method you describe. That same ~$ 10 trillion is already included in the household assets included in the hh net worth figure. See line 29 , here :

    http://www.federalreserve.gov/releases/z1/Current/z1r-5.pdf

    I’m thinking that’s one of the things the forthcoming explanatory Fed Note will cite for reasons that the sectoral net worths do not sum to national wealth as per the B.1.

    BTW , WTF are they waiting for in giving us a proper explanation of this thing , anyway ? How many smart people in this country are spending hours sitting at their computer in their underpants trying to make sense of this crap ?

    And what’s the big deal with migrating the new B.1 data to the FRED database ? Is that really so hard to do ? My guess is we’ll finally get it in ~2017.

    Sometimes I think they do this stuff just to pull my chain.

  147. Marko writes:

    One more thing – Why can’t we get a nice , neat summary of national wealth like the UK provides ?

    http://www.ons.gov.uk/ons/rel/cap-stock/the-national-balance-sheet/2014-estimates/index.html

    http://www.ons.gov.uk/ons/dcp171778_386098.pdf

  148. Steve Roth writes:

    Okay, the UK presentation vs the B1 pretty much illustrates the situation. Thanks Marko.

    UK:

    Sector Net worth
    Households and NPISH 8,455.10
    Non-financial corporations -1,186.30
    Financial corporations 533.5
    General government -157.8
    Of which: Central government -663
    Of which: Local government 505.2
    Total 7,644.50

    If you want to just add up the sectors like this, you have to deduct corp equity from corp net worth/book value. Cause the equity’s already there in households. But the (negative) corp net worth number is totally artificial, has no relationship to what it would cost to buy or re-create them, or to their future cash flows.

    But the B1 uses a more sophisticated approach, tallying only nonfinancial assets for non-equity-issuing, noncorp private sectors, then just showing full market-value equity for the equity-issuing corps — the market’s best estimate of what they’re “worth” You really can’t deduct corp equity here from…corp equity…cause you’d get zero, and unlike the UK sum-it-up sheet, that wealth is not counted anywhere else.

    You can’t just add up the sectors cause one is owned by another. Double counting. I’m liking the B1 solution better. And full household NW is helpfully provided as an addendum item at the bottom. Cool.

  149. Marko writes:

    SR ,

    Yes , that’s been my interpretation for quite a while , and I’m not willing to revisit the topic until I get an official Fed explanation of how the the B.1 relates to sectoral NWs.

    I’ve been operating under the impression that private sector net worth is collapsed into the hh sector FOF net worth figure , full stop , and thus the best estimate of national NW ( until the B.1 came along ) could be achieved by summing hh and gov’t net worth , and for gov’t net worth we have the IMA data , for better or worse. Frankly , it was never clear to me whether ROW net worth should be deducted , so in the interest of my sanity , I simply ignored it. With the way we’ve been shipping wealth offshore , I know that strategy has its limits.

  150. Steve Randy Waldman writes:

    JKH — I agree with the substance about your point about insurance not captured in NFA. But I want to put it a bit differently.

    You are absolutely right that, for example, insured bank deposits serve the same insurance function as Treasury bonds, but they are not captured by the MMT-standard sectoral balances analysis. To generalize, the contingent liabilities of the state, as well as its overt deficit finance provides this sort of insurance. That is why, for example, Gary Gorton advocates state guarantees of the shadow banks that broke. He wants to preserve and increase the ability of private-sector financial intermediaries to offer this sort of insurance, but he understands that absent a state backstop, it cannot adequately do so.

    But before writing off sectoral balance accounting on this basis, it’s worth pointing out that this problem is absolutely ubiquitous under all standard accounting schemes. Under US GAAP, if I recall, has very broad rules about when contingent liabilities can go unbooked, especially if the probability of realization is uncertain. That’s a practical necessity. Every mom and pop restaurant has a contingent liability several times the net worth of the business, should they serve some food that sickens or kills a customer. The probability of that is small, the expected value of that position if realized is large, its expected value might be material, but it cannot be estimated and so is omitted from accounts. (Note the liability-on-realization should be understood as larger than net worth, even though the firm is limited liability and will payout no more than that, because the amount of an actual award conditions the risk of new investment in the firm.)

    TARP, again, is best understood as the realization of a deeply uncertain contingent liability of the state. Part of the systemic risk insurance that private sector entities hold has to do with the uncertain but real possibility that the state will intercede to clip certain bad outcomes even when it has issued no formal promise to do so. That asset of private sector entities is matched by a contingent liability on the part of the state (where the politics of deciding to payout is part of the contingency).

    Godly/MMT sectoral balance accounting fails to capture this. But that doesn’t render it worthless, any more than the omission of conditional liabilities at a restaurant render those accounts worthless. A set of accounts is always an imperfect model. Godley/MMT sectoral balances is no exception, but it does capture and convey real information.

    A more comprehensive measure of NFA, one that consolidated guaranteed bank deposits and other state guarantees that would be offered with near-certainty in a bad state of the world and whose amounts are estimable, would be a worthwhile exercise. Formally insured bank money is special, as opposed to other liquid privately issued liabilities, because of the certainty that the state would intervene if necessary to make holders of those liabilities whole, even at the expense of other members of the polity. Whether the same is true of other bank liabilities, and of precisely which others, is a more difficult and contentious question.

    The standard NFA analysis is not complete or perfect. But it still conveys a great deal of useful information in my view.

  151. Steve Randy Waldman writes:

    JKH — With respect to our prior discussion, I think we’ve come to common ground.

    We don’t disagree, I think, that if we are divising an accounting system where financial assets are going to offset financial liabilities, they should sum to zero.

    But your point is well taken, that sometimes you can construct very useful aggregate financial accounting systems without netting financial positions, just letting financial assets stand in for assets. The fact that financial assets have market values may make estimation more practical than systems that try to look through financial asset values.

    As David Rosnick points out, the downside of this sort if procedure is that it leaves the valuation of whatever is accounted for as only a financial asset as a black box, where usually (though not always) most of what we think of as the value of the thing being accounted for derives from some sort of real asset (broadly defined to include “organizational capital” and whatnot). But sometimes that’s okay.

    Your point is also well-taken that, under some circumstances, a sector-specific balance sheet is sufficient to capture the “net worth” of the aggregate economy, and so might be called a fully aggregate set of accounts. I don’t think those conditions completely hold with respect the household sector in the US: the state and foreigners hold claims against US assets, not just US households. At some level, one can argue, that US households “own” the assets of the state, but those are usually off-balance sheet positions. One can of course put them on-balance sheet, as the new B.1 table to which Ramanan introduces us does, and that is an interesting new accounting.

    Note that the new B.1 table nets out financial claims of US entities against US governments, and looks through them to government real assets. So this is a set of accounts that would entirely omit NFA. And it is a fine set of accounts. The government paper portion of NFA is not a contributer to US net worth. It is represents a set of distributional arrangements within the US that provides insurance to some parties. Different accounting schemes reveal different things, and many of them are useful.

  152. Steve Randy Waldman writes:

    JW —

    This sounds reasonable but I don’t think it is. Because we are not (only) trying to come up with an ideal set of accounts to use in constructing a formal model of the economy, we are trying to organize the economic data we actually observe “in the wild.” And the only source of that data is the accounts kept by private units. You might believe (I still don’t really see why) that no asset should appear on anyone’s balance sheet without an equal-value liability on someone else’s balance sheet. But the only way we actually measure anything, is by looking at the entire that actually are on private balance sheets, and they don’t follow this rule.

    So, you are misunderstanding. There is no one true accounting system. Accounting systems are instrumental. (Like models, they are always wrong but sometimes useful. They are often more straightforwardly and tangibly useful then economic models though.)

    I certainly don’t believe that “no asset should appear on anyone’s balance sheet without an equal-value liability on someone else’s balance sheet”. Real firms, “in the wild” as you put it, value their assets and liabilities in ways that address their own concerns. Some of those concerns are regulatory, which introduces a certain measure of consistency to accounting statements between firms, but not all that much. Even with in firms, most nontrivial firms have multiple sets of accounts for multiple different purposes. A firm produces a depreciation estimate for its “normative” income statement and balance sheet, but a different one for tax accounting purposes. Accounting, “in the wild”, serves many purposes, but it is not about providing some “one, true” picture of a firm. (I have more on that here.)

    I don’t think accounting statements are consistent within a single firm. I certainly don’t expect them to be consistent across firms. Obviously, “in the wild”, firms book financial assets at values different from the liability booked by the issuer.

    But an aggregate accounting is not an aggregation of the reported financial statements of individual firms. That would be an entirely incoherent procedure, as arbitrary accounting choices made by different firms would translate into changes in aggregate net worth. An aggregate accounting exercise (just like a comparative accounting exercise performed by financial analysts) has to adjust valuations to be consistent with one another. Usually, in an aggregate accounting exercise, market values are taken for financial assets, and those are treated consistently as liabilities or as equity at market value by issuers. Market value is not the only reasonable standard. But it is hard to come up with an aggregate accounting scheme in which not valuing financial assets at the same value for holders and issuers would be useful. “In the wild”, accounts are quite properly heterogenous. In an aggregation, they are recast on a consistent basis.

    As we all know, the financial accounts prepared by the fed for the US do NOT follow your axiom that total financial asset must equal total liabilities. Are they useless? Are they strictly less useful than any other accounts that do follow your axiom? Obviously not. So you do don’t really believe your axiom. At best, it is not a hard rule but one desirable property of a system of accounts that has to be weighed against others.

    Do we know that? I think I just went through an exercise with JKH where we agreed that in fact in the flow of accounts financial assets are valued at market for both holders and issuers and they are in fact consistent. I can ensure you that I really believe what I believe. What we can agree on is that there are circumstances where it might not be useful to net financial assets against issuer liabilities, but just leave them as “black box” financial assets, as the new B.1 table does for domestic corporations but not for noncorporate domestic businesses. Note how pragmatic that choice was. Clearly the intention of the B.1 table is to look through to the values of real assets, as it does with noncorporates, but since “valuing” real assets is a practically and philosophically fraught exercise, where market value data does exist (the corporate sector), the compilers of the B.1 table chose to use it. That strikes me as perfectly reasonable. You’ll note that both the new table B.1, which offers a holder-side view of for example corporate equity, and tables like B.103, which offers an issuer-side view, consistently use market values of equity. They show close but not identical values only because they are “adjusted to reflect net U.S. financial claims on the rest of the world” (latest Z.1, p i.v.).

    But there’s no sharp line between “real” and “financial” assets. An asset is just a property title that gives the holder certain legal rights and which is expected to yield a certain money income. We could say that the “real” assets are the tangible assets recorded on a business’s books, at their historical cost or at their replacement or market value. Or we could say that the “real” asset is the business itself, valued by the market value of its equity.

    This is a good point in the abstract, but again, accounting imposes a reality, it does not try to define the one true reality. With respect to any accounting system, it is always important to judge whether the names we attach to our accounting relationships map to what we think we are talking about in the real world. You are right to point out that we don’t know, exactly, what we are talking about in the real world when we distinguish between real and financial assets. So we ask at least in a pornography-inspired “know it when we see it” kind of way.

    I’m going to start with an imperfect but pretty good definition of what we mean by a financial asset. A “financial asset”, I’ll tentatively claim, represents a claim by one legally recognized entity against a distinct legally recognized entity. So the fact that a home is represented on paper and in legal proceedings by a “deed” doesn’t make it financial. That deed is a paper representation of an asset, but it is not a claim against a distinct legal person or entity.

    Again, we can choose different definitions of financial, but that’s the one I’m hoping to capture when I say that the sectoral balances equation is an accounting of financial claims.

    Now if you are very careful, you will note that it is in practice a not-quite-adequate accounting of financial claims by that definition, because NET_FOREIGN_FINANCIAL_POSITION usually includes FDI, which usually represents a claim of a parent company against a subsidiary (which is fine), but also includes natural persons holdings of foreign property. You can try to remedy this by definition, by declaring that a natural person in country X is a different legal entity than the same natural person in country Y. Or not. In aggregate this kind of investment is (I think, am I wrong?) small, and however persnickety we choose to be, the accordance between what the sectoral balance accounts represent as “financial” and one real-world understanding of what it means to be a financial asset is close enough to qualify as pornography to any Supreme Court justice.

    Shorter: Measures of net worth we actually observe do not in fact equal our measures of real wealth. So we can impute unobserved financial assets/liabilities, or we can assume that the discrepancy means we have incorrectly measured real wealth. (Which is not so unthinkable, is it?) Or we can accept that preserving this particular symmetry is less important that preserving the correspondence of our aggregate accounts with the private accounts that actually shape economic behavior.

    There is no one true measure of net worth, individual, corporate, or aggregate. There is no one true measure of net wealth. We invent accounting schemes or estimating procedures that we know are imperfect for certain purposes. Most of our estimates disagree with one another, but that doesn’t render them useless for their purposes.

    I know of no way to construct a useful accounting system that 1) nets together the value of financial positions of holders and issuers; 2) purports to incorporate nonzero values of that sum in some measure of aggregate net worth (as opposed to value added); and 3) seems useful. That doesn’t mean that such a thing is impossible, but I certainly haven’t seen one, and consider it unlikely from first principles. It strikes me as unlikely because it opens the scheme to a kind of “regulatory arbitrage”, whereunder reorganization or reshuffling of the paper claims affect our net worth valuation in ways that we wouldn’t consider useful mappings of changes in the “real net worth” we are trying to model.

    Again, all of this is arbitrary, and we have no true measure of net worth, and if we want to we can aggregate individual unadjusted valuations of everything and call that aggregate net worth. I just don’t think it would be very useful.

  153. JKH writes:

    SRW,

    We are largely in agreement I think.

    One point I might emphasize. I have a natural preference for the long-standing household net worth presentation due to decades of habitual on-the-job active use on a pre-SNA basis. And I also think that the financial asset values captured directly are very interesting. I don’t have a problem with the SNA type format and agree that it is useful, although as a complement rather than a substitute in my case.

    A distinction I might draw now is that it is the association of equity claims with the idea of liability that I find potentially dangerous – not the treatment of equity within a formal macro accounting scheme that subtracts the value of equity in netting. Those are two different albeit intersecting relationships. The association of equity with the idea of liability is dangerous because it contradicts a foundational element of financial accounting and by extension an understanding of capitalism at a deep level of its mathematical depiction. And I think the lazy association of these two things because of an otherwise benign optional macro accounting technique is somewhat unfortunate.

    “Note that the new B.1 table nets out financial claims of US entities against US governments, and looks through them to government real assets. So this is a set of accounts that would entirely omit NFA. And it is a fine set of accounts. The government paper portion of NFA is not a contributor to US net worth.”

    Yes. This was the point I attempted to make in a perhaps unwieldy way in my # 144. It looks to me like a very interesting hybrid of the household net worth presentation and the SNA type “no-net-financial-assets” presentation.

  154. Ramanan writes:

    Marko,

    I don’t see any double counting. The table B.1 mentions household net worth but doesn’t use it anywhere. They have household non-financial assets. For noncorporate businesses, non-financial assets of these are added on top on household non-financial assets. For corporate business, values of equities are added on top of household non-financial assets.

  155. Marko writes:

    I was surprised to discover tonite that the Fed has updated the Z1’s back to 1945 to include the new B.1 report.

    I decided to plot the sum of sectors method of arriving at national wealth and compare it to the B.1 values as well as my previous method of only summing hh plus gov’t sectors. The total sum of sectors runs much higher , due I’m sure to double-counting of noncorporate business equity. Leaving that out of the sum , I thought maybe any residuals would be inter-sector quantities that would cancel out , but a significant residual remains , causing a high-side bias most of the time. The hh plus gov’t method seems to be a good proxy , all the way back to 1945 , running at a level just above the B.1 value for the few numbers I checked. See :

    https://research.stlouisfed.org/fred2/graph/?g=24hm

    The blue curve is hh plus gov’t , the red curve is all sectors less noncorp business equity and the black lines are manually set at the B.1 values for 2011-2014 , 1985-86 , and 1945-46.( It’s best to adjust the time slider to see the earlier part of the curve )

    I didn’t subtract ROW net worth ( though it’s there in the line 2 series if anyone wants to play with it ) because this page indicates that the alternative national wealth calculation is arrived at by only summing the domestic sectors :

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL892090025&t=S.2.A&bc=S.2.A:FL112090205&suf=A

    If you apply a factor of .97-.98 to the series 1 formula ( blue curve ) , the curve aligns with the B.1 values I’ve plotted , so I’m pleased that the national wealth proxy I’ve relied on matches the new “official” figure with only a couple percent or so of deviation. I’ll be glad when the new data shows up in FRED so I can plot the entire curve. I’m way too lazy to check all the numbers manually.

  156. JKH writes:

    Ramanan,

    Right.

    No double counting.

  157. Marko writes:

    Ramanan ,

    You’re right that it’s not mentioned explicitly on the B.1 page , but if you look at the footnotes ( here: http://www.federalreserve.gov/apps/FOF/Guide/P_6_coded.pdf ) , it says :

    ” (5) Household net worth is calculated as the difference between total assets and liabilities of the household and nonprofit organizations sector. See table B.101. ”

    That’s the table with the “line 29″ figure of hh assets that I referenced above.

  158. Marko writes:

    I’m sorry. I see what you guys mean now. The hh net worth that footnote 5 refers to is just the memo item.

    OK , now I really need to see how they’re going explain the differences , because the sum of sectors line (red ) in my graph above is going to be $10 trillion higher in the later years.

  159. Marko writes:

    I still don’t see , however , how if you follow this procedure for the wealth calculation that it’s not double counted :

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL892090025&t=S.2.A&bc=S.2.A:FL112090205&suf=A

    i.e. the first line already includes the second line :

    + FL152090005.A Households and nonprofit organizations; net worth
    + FL112090205.A Nonfinancial noncorporate business; proprietors’ equity in noncorporate business (net worth)

    as detailed , here :

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL154090005&t=S.2.A&bc=S.2.A:FL112090205,FL892090025,FL152090005,FL152000005&suf=A

  160. Ramanan writes:

    Marko (#159),

    I think the seriesanalyzer you link is for the IMA way of doing it which it does by just adding net worths of all resident sectors.

  161. Marko writes:

    Ramanan,

    Yes , but that’s exactly the way you proposed doing it in your blog post , with the exception that you added ROW. From your blog :

    “…..So let’s use those numbers.

    Flow of funds’ net wealth for 2014 = $77.89 tn (Table B.1, line 1).

    Now, go to Table S.2.a. These tables use SNA definitions. Add lines 76-81 and subtract line 82.

    This gives us a value of $81.65 trillion….”

    Here’s lines 76-81 from S.2.a ( 2014 data ):

    Net worth
    76 FL152090005 Households and nonprofit institutions serving households 83425.7
    77 FL112090205 Nonfinancial noncorporate business 10021.3
    78 FL102090085 Nonfinancial corporate business -1411.3
    79 FL792090095 Financial business -827.2
    80 FL312090095 Federal government -11745.6
    81 FL212090095 State and local government 7873.5
    82 FL262090095 Rest of the world 5687.8

    It’s double counting of line 77 , because it’s already included in line 76 , traceable in the series analyzer as I’ve demonstrated.

  162. Ramanan writes:

    Marko,

    Good catch. Think I see your point.

    Around $10 trillion as per your comment #146.

  163. Ramanan writes:

    So basically for the IMA way of calculating the net wealth, it uses IMA definitions but then for noncorporate businesses uses the FoF definition of net worth.

    Should correct this in my post.

  164. Ramanan writes:

    So B.1 is fine. S.2.a line 77 is wrong.

  165. Roberto writes:

    Steve,

    You say: “And I worry much more than I think most MMT economists do about the unjust distribution of risk-bearing that might accompany a large stock of net financial assets very unequally distributed.”

    I think, if I understand you correctly, that you are being a little unfair to the MMT guys. Most of them propose policies that address how to add the financial assets to the economy.

    For instance, in http://bilbo.economicoutlook.net/blog/?p=11911 Bill Mitchell address this problem in the context to where the QE money has gone.

    Also, the Job Guarantee program is integral to most MMT proposals and gives a pretty straightforward answer: the new financial assets are used to employ the unemployed, creating a new automatic stabilizer.

    Of course it’s possible to agree with the descriptive aspect of MMT and not the prescriptive.

  166. JKH writes:

    My very strong suspicion is that S.2.a line 77 is OK

    The proof will lay fallow pending reflective confirmation or not

  167. Ramanan writes:

    JKH,

    It’s correct as per the FoF definition but not right as per the SNA definition.

    Also I have updated/edited my post. http://www.concertedaction.com/2015/10/07/united-states-net-wealth/

    Surprisingly, the IMA number and FoF number turn out to be quite similar! That’s strange because equity prices can be anything and this difference depends on stock markets.

  168. JKH writes:

    Ramanan,

    S.2.a line 77 looks fine to me

    It is derived in detail in table S.4.a

    And the IMA calculation is a positive number quite similar to what would be the calculation for net worth from the older flow of funds template, because there is no financial asset interface between the net worth of a non-financial non-corporate firm and the owners of that firm. There is no negative netting of equity financial claims because there are no such financial claims issued as there are in the corporate case. The ownership is direct, without financial intermediation.

    You can see the difference by comparing that with table S.5.a, which shows the corresponding format in the case of nonfinancial corporate business. The pertinent classification there is line 140, “corporate equity”. That reflects the issuance of a financial asset, which is why it is netted out as a negative in the calculation of IMA net worth.

    Conversely, there is no such financial asset issued in the case of non-corporates, which is why there is no negative netting.

    Indeed, this is quite analogous to the treatment of the government sector in IMA, in which case there is also obviously no equity financial claim issued and netted out on a negative basis.

    That all makes sense to me in terms of the overall IMA framework.

    And so the line item 77 looks fine to me.

    As to where that leads in terms of further aggregation and reconciliation challenges, I don’t know because I haven’t gone through it yet.

    But I fail to see why line 77 is problematic, because the profile looks perfectly fine based on the IMA “rules” as I understand them to be differentiated from the flow of funds template “rules”.

  169. Ramanan writes:

    JKH,

    The error is everywhere. Not just line 77 of S.2.a.

    I think the IMA errs in thinking that since no equity is traded publicly, it has to ignore it altogether. Hence S.4.a is also wrong about net worths.

    The best way is to think of the consistency rule that for a closed economy, the net worth should add up to the sum of nonfinancial assets.

    Imagine all businesses are non-corporate and households’ value of houses is next to nothing. And firms’ financial assets is next to nothing. And firms have some bank loans.

    In that economy the sum of the net worths of all sectors should add to non-corporate businesses’ non-financial assets.

    All financial assets can be seen to cancel each other except ownership in non-corporate businesses. There should be a cancelling item so that net worths add to businesses’ non-financial assets, right. How does this cancellation happen?

  170. JKH writes:

    And I see line 77 appears directly as the household asset line 121 in S.3.a

    Provided this position is treated as a net positive value subset of household assets and therefore of household net worth (with no financial asset netting) rather than additive to household net worth (which would require financial asset netting), line 77 makes sense as a carve-out item

  171. JKH writes:

    Ramanan,

    I want to make sure I understand what you’re saying at a high level of comparison.

    Are you saying that the IMA is supposed to reflect the SNA aggregation technique, and that the SNA aggregation technique in respect of the NF NC net worth line nets it out the same way as it does for corporate equity claims, and that therefore the IMA report is wrong in this regard?

  172. JKH writes:

    And again …

    On S.2.9 there is no direct indication that all the lines following line 76 are intended to be additive to line 76; i.e. line 77 can be interpreted as a positive carve out from line 76

    Which is consistent with how line 121 is presented in S.2.9

  173. JKH writes:

    sorry … meant line 76 and 77 in S.2.a, and line 121 in S.3.a, both in # 172 above

  174. Ramanan writes:

    JKH,

    Line 121 in S.3.a looks fine to me.

    It’s things below line 106 in S.4.a that looks wrong to me.

    I am not sure about your point about carving out. I see the ownership as an additional asset for households and hence affecting household net worth. In the sense that had this asset been zero, net worth would have been lower.

    About comment #171, not just because of SNA but for the sake of its own self-consistency, the IMA has to treat NC in the same was as corporates.

    Else it’s double counting. When I add household net worths and NC net worth and all net worths to arrive at the national net worth, there’s double counting because there’s no corresponding liability which cancels out ownership in NC.

    Or else think of a single company being moved out of its classification from non-corporate to corporate or the other way round via an IPO or a full buyback. The net worth of the nation shouldn’t change because of this alone.

  175. Ramanan writes:

    JKH,

    It’s possible what you are saying is right. But Marko has pointed out this:

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL892090025&t=S.2.A&bc=S.2.A:FL112090205&suf=A

    The second line FL112090205 is added. This is the net worth of non-corporate businesses as calculated in S.3.a and is equal to around $10 trillion.

  176. JKH writes:

    Ramanan,

    “When I add household net worths and NC net worth and all net worths to arrive at the national net worth, there’s double counting because there’s no corresponding liability which cancels out ownership in NC”

    Consider how IMA treats NF NC, as I have described it

    It’s treats it the same as a house in effect

    There is no separate balance sheet net worth position for a house alone (in this case there is no balance sheet, but you could certainly construct one with an equity net worth on the right hand side) that needs to be added to the existing household net worth balance sheet – because the net effect is already embedded as a household asset

    It’s the same principle according to the way in which IMA treats the NF NC balance sheet

    There is no balance sheet net worth for NF NC that needs to be added to household net worth – because the net effect is already embedded as a household asset

    So there is no double counting because there is no addition of balance sheets required in the first place

    That’s what I mean by carve out

  177. Ramanan writes:

    JKH,

    Consider an example. Let’s suppose that the Fed’s statistics is near perfect. Except that it ignored Ramanan’s firm. Assume I am a US citizen. I have a registered office which has a bungalow valued at $1,000,000 and has a computer priced $3,000 which I use to make an ad-blocking App for iPhone/iOS9. My firm has a bank account and whenever someone purchases an App, I transfer it to my personal account and spend it all. My personal bank account is next to nothing. My consumption is estimated well by the Fed. So no need for changes there. Only thing surveyed incorrectly is my bungalow and my computer.

    As per Z.1, assets of my firm is $1,003,000 and the net worth of my firm is $1,003,000.

    And because of its incorrect survery my net worth was calculated incorrectly by the Fed and it is $1,003,000 instead of zero. Because assets = $1,003,000 (ownership) and liabilities = 0.

    But when national wealth is calculated by the IMA, it will add my net worth ie, $1,003,000 in households and also the net worth of my firm $1,003,000 in non-corporate businesses.

    So FL892090025 is calculated (the link in my previous comment) it will report an extra $2,006,000.

    It should have been $1,003,000.

  178. Marko writes:

    JKH ,

    This…..

    “There is no balance sheet net worth for NF NC that needs to be added to household net worth – because the net effect is already embedded as a household asset

    So there is no double counting because there is no addition of balance sheets required in the first place. ”

    ….is not in dispute. The problem we see is with the IMA calculation of national wealth , as outlined in the Series analyzer for FL892090025.A (All sectors; U.S. wealth (Integrated Macroeconomic Accounts) , where the addition of sectoral values does result in double counting.

    It’s telling , perhaps , that the summed value that FL892090025.A identifies does not actually show up anywhere. If it did , however , using the method outlined , it would be ~ $10 trillion over the B.1 value for national wealth.

  179. JKH writes:

    Marko,

    “It’s telling , perhaps , that the summed value that FL892090025.A identifies does not actually show up anywhere. If it did …”

    Spending a little more time perusing it all, that’s exactly the question I was about to ask upon returning here – because I expected it didn’t show up anywhere.

    I now think that the code we see there for FL892090025.A is actually wrong.

    I.e. I don’t mean wrong conceptually – I mean that the code that appears there has been transcribed incorrectly from what is actually used in compiling the reports – and that the actual underlying code omits the double counting to which you refer.

    I’m thinking that for two reasons:

    a) Maybe you guys have crunched the numbers and are getting something closer than me, but when I do the actual addition manually, the result seems too far away from what I’m eye balling and calculating overall otherwise

    b) More noteworthy, if you look at the logical development and drill down of the Integrated account series in sequence – with the total, the household sector, and then the NF NC net worth as a subset of the household sector – it all makes sense to me and looks orderly and logical and aggregated correctly and consistently though the piece.

    So I think it’s the code that is wrong literally on the surface, and that the true underlying code omits the addition of NF NC net worth.

  180. Marko writes:

    Also telling , I think , is that the NF-NC business , proprietors’ equity series is the only one , other than hh , that does not carry the designation “(Integrated Macroeconomic Accounts)”.

    If the private sector net worth is effectively consolidated in the hh sector in the FOF analysis – which is the general understanding as I’ve always perceived it – then the NF-NC business net worth should show up as as a near-zero residual in the IMA. I think it might have just slipped thru the cracks.

  181. Marko writes:

    ” So I think it’s the code that is wrong literally on the surface, and that the true underlying code omits the addition of NF NC net worth. ”

    JKH ,

    Yep. That’s probably it. And since they never actually published any incorrect data using the bad code , it’s a “no harm , no foul” situation from their standpoint. Meanwhile , we’ve all frittered away valuable time we could have spent drinking , or something.

  182. JKH writes:

    yes

    that time opportunity cost is a tragedy

    but correctible

    :)

  183. Steve Roth writes:

    So sorry to have inflicted those costs on everyone. I, at least, have benefited greatly from the results. Thanks to all.

    Bottoms up!

  184. JKH writes:

    Not just yet Steve R.

    I have discovered something else, and will have more later today.

  185. Steve Roth writes:

    Oh just one more item.

    SRW: “The government paper portion of NFA is not a contributer to US net worth.”

    Right, because they’re fed gov liabilities, deducted from national net worth. Fed gov net worth is -11.8 trillion, largely due to treasury-bond liabilities of…$14.4 trillion. (ca $3 trillion of which are held by the Fed, which is another issue.)

    Just to say: based on 200+ years of U.S. history and 300+ years of U.K. history, they are liabilities that will never be paid off. Kind of special that way.

    http://www.asymptosis.com/the-meme-that-refuses-to-die-government-debt-must-be-paid-back.html

    Even right now, the markets clearly have no expectation that those liabilities will ever be paid off.

    http://www.asymptosis.com/the-market-doesnt-think-the-fed-will-ever-sell-those-bonds-back.html

    You all get this: we need a different term than “government debt.” SRW has talked about government money as equity (roughly paraphrased), and Chris Cook has suggested the term “National Equity.”

    http://www.asymptosis.com/yes-the-government-must-pay-its-bills-in-the-long-run-every-few-centuries-questions-for-krugman.html

    Is this equities-as-liabilities thing sounding familiar?

    As Cook says:

    “Only by liquidating US Incorporated could this National Equity ever be redeemed.”

    As we keep discovering, “equity” is special. No forced redemption.

    This points out yet again how accounting terms and definitions — which are hugely privileged, because everyone assumes them to be unambiguously “true” — can have profound rhetorical and political implications.

  186. Ramanan writes:

    Cheers Steve R,

    I just noticed this:

    SRW: “The government paper portion of NFA is not a contributer to US net worth.”

    While true just via accounting, as it cancels with domestic holders of the debt, fiscal policy drives the economy and production and leads to a rise in the US net worth.

    Steve, about your points, I don’t think there’s anything more neutral than national accounts, flow of funds etc. There are of course differences as we have discussed here (for example between flow of funds and IMA) but these are minor ones.

    About your point of not paying off … that’s not correct. Public debt – although never retired completely – cannot keep rising relative to GDP. It’s about sustainability of the debt. For example, “liability dollarization” may happen to government debt.

    Neither is “national equity” useful. As we have seen here, net wealth is different from public debt and quite large compared to it. It’s just some Moslerism (“to the penny”) in different words.

  187. JKH writes:

    I’ve always preferred the original Fed Flow of Funds table B.101 on household net worth as my macro accounting methodology of choice – on the basis that one should be able to transit from B.101 to the integrated macro account (IMA) reports with fairly simple adjustments, according to the “negative equity methodology” of the latter.

    So now:

    The full Fed Z1 report with original formats + IMA in pdf form:

    http://www.federalreserve.gov/releases/Z1/Current/z1.pdf

    Go to page 6 of this report
    B.1 “Derivation of U.S. net wealth”
    It employs IMA type calculations

    Go directly to the memo items and their Q2 2015 values:
    Line 31 U.S. net wealth $ 79.690 trillion
    Line 33 Household net worth $ 85.712 trillion
    Difference $ 6.022 trillion

    Line 31 derives from the IMA “negative equity methodology”
    Line 33 derives from the old B.101, found in detail on page 134

    So the objective is to reconcile the way in which IMA national net wealth is $ 6.022 trillion less than B.101 household net worth.

    Go to page 145
    This is the tail end of the IMA table:
    “S.2.a Selected Aggregates for Total Economy and Sectors”
    Look at the six line items at the bottom
    Note that these are all IMA calculations

    76
    Households and non-profit institutions serving households net worth (IMA basis)
    $ 83.425 trillion

    77
    Nonfinancial noncorporate business net worth
    $ 10.021 trillion

    78
    Nonfinancial corporate business net worth
    $ -1.411 trillion

    79
    Financial business net worth
    $ -.827 trillion

    80
    Federal government net worth
    $ -11.745 trillion

    81
    State and local government net worth
    $ 7.873 trillion

    82
    Rest of the world net worth
    $ 5.687 trillion

    Now add the following:
    lines 78 + 79 + 80 + 81
    = $ 6.110 trillion

    Very close to the difference to be explained of $ 6.022 trillion.
    I don’t think it’s a coincidence.

    Explanation:

    The main factors to be explained in moving from B.101 household net worth to IMA national net wealth are the IMA net equity adjustment (78 + 79) and the government net worth adjustment (80 + 81).

    The equity adjustment essentially replaces the B.101 value of household equity asset financial claims with the underlying corporate real asset values. And the government net worth adjustment does essentially the same thing, replacing claims in the form of government debt with the underlying government real asset holdings.

    Importantly there are several line items in the list above that are not part of the required adjustment:

    I’ve ignored the top line 76, which is the IMA version of household net worth. This value is somewhat different than B.101 (by about $ 3 trillion), and reconciling the difference with B.101 would require further analysis that I’m not interested in at this point. I’m interested only in the transition from the original B.101 to IMA national net wealth.

    Line 77 showing the value of non-financial non-corporate net worth must be excluded from the IMA national aggregation calculation. (The IMA calculation of this line is essentially the same as the B.101 derivation.) Net worth is calculated according to conventional financial accounting, without being netted away with a negative equity factor as is done in the case of corporate equity. This conventional financial accounting measure of net worth is fully included as an asset in B.101, and therefore captured in the starting point of B.101 net worth, so we can ignore it for purposes of any further adjustment. I now believe the reason for this conventional net worth treatment may be that there is no public market for this equity and that there is therefore no such thing as public data reflecting fluctuating trading values. To attach a financial equity value in an attempt to reflect valuation changes based on financial markets would be guesswork and artificial and pointless. So the net worth is calculated on the basis of the value of the underlying assets and rolled into household net worth on the asset side without further ado. (My earlier suggestion that this is not a financial asset was incorrect. It is classified as a financial asset in both methods of reporting. But it is a different kind of financial asset – one whose financial value cannot be observed by continuous public market evaluation, as is the case with corporate equities.)

    Line 82 is not part of the IMA aggregation calculation. This is the net asset position of the rest of the world as a positive claim on the US. It is excluded from aggregation because its netting effect is already reflected in the residual net worth of those assets that do remain available to US claimants. So this item can be ignored – it does not need to be subtracted in aggregation. To do so would be double counting the netting away of the foreign claim from a household net worth that already takes this into account.

    The important point about lines 76 to 82 is that this is just a list of sector net worth calculations that are not necessarily additive. This sort of listing of incongruent line items seems to occur frequently in macro accounting tables, making full reconciliation more challenging.

    Regarding Marko’s comment # 261:

    As I said earlier, I believe the inclusion of the non-financial non-corporate sector is an error in the transcribed code for FL892090025.A. It should be excluded because as explained above there is no netting of the book equity value of that sector, the way there is for the corporate sector. The $ 10 trillion net worth of the result is already captured as a subset of the $ 83 trillion aggregate net worth (IMA version) of the household sector (it appears as a household asset). Whereas the $ (2) net worth of the corporate sector (calculated by netting out equity claims) is not a subset of household net worth and instead is additive to it in the calculation of the total economy’s net worth, as per the usual IMA additivity using negative values for equities. And the construction of the accounts through the various tables makes it clear that the non-financial non-corporate sector is treated according to the conventional financial accounting version of equity – not according to the treatment of equity as a negative item or liability in the additivity of net worth across the economy. Finally, as explained above, the code does not subtract ROW $ 5.6 trillion net worth – it merely omits it (correctly).

    I think this transformation from B.101 to national net wealth is intuitively sensible, and the numbers seem pretty good.

  188. Marko writes:

    JKH ,

    Thanks for that useful summary. I’m going to work thru the numbers myself , but not until I’ve had a break for a couple of days. I’ve had more than my fill of FOF (FOFOF) lately.

    It’s satisfying to know that we’ll now have a reasonably consistent ( and officially-determined , rather than tentatively proxied ) measure for “national wealth” that goes back to 1945. I suspect that researchers will attempt to extend the series even further back via various approximations.
    Frankly , it seems like something that should have been done since Day One , but better late than never , I suppose.

  189. Ramanan writes:

    JKH,

    Yeah yesterday I edited my post to not count the rest of the world (which I erred earlier):

    http://www.concertedaction.com/2015/10/07/united-states-net-wealth/

    Regarding the point about non-corporate sector. One way to say is as you say it … ignore it in adding. The other way is to treat the equivalent (relevant) assets of households as the liabilities of this sector, so that the net worth of this sector is close to zero.

    About your point that since the rest of the world is ignored so that it’s not as if all lines have to be added in 76-82 in S.2.a, my point will be that national net worth is sum of resident sectors. So that automatically excludes 82. That’s not a justification for not adding line 77.

    About your point about coincidence: it’s really a coincidence. The net worth calculated in B.1 uses the market value of equities but there’s no corresponding liability anywhere. So it depends on stock market indices. One could have imagined much higher stock prices, because markets can do anything. And the difference between IMA net worth and FoF net worth can be large.

  190. JKH writes:

    Ramanan,

    It’s not a coincidence.

    If the market value of equities was different and/or the underlying value of real assets was different, then obviously the differential would be different. But the equivalence relationship I used would still hold.

    And in connection with the rest of the world, as I explained its line 82 that’s not added – not line 77.

    Line 77 is not added for an entirely different reason, which I also explained.

    The code referred to by Marko correctly omits line 82 and incorrectly includes line 77.

  191. Ramanan writes:

    JKH,

    Agree on the equivalence relationship part.

    There is however, a complication which neither the FoF or the way the IMA is currently doing it in the report which I am worried of.

    The code actually seems right to me.

    The Table B.1 which actually is not IMA (and hence the talk of differences between B.1 and S.2.a) already treats equities held by foreigners as liabilities.

    More complication however arises if you consider Direct Investments.

    Let’s take S.2.a and consider Direct Investment of 100% from say the UK in one firm in the US. It’s a wholly owned subsidiary.

    By the logic of distinguishing using whether equities are traded in stock markets or not and whether to treat net worth differently on this basis might lead to some inconsistencies. How will you account for this firm?

    There should be something which shows that US residents have a liability toward UK residents for this case which is accounted in S.2.a.

    (An example can be built by thinking of me as a UK citizen purchase a building for next to nothing and it super-appreciates in value in a year whereas other things are about the same. Wouldn’t the US’ net worth reduce and if so, in the Table S.2.a, where would it appear?).

  192. Ramanan writes:

    Okay building in this case may be a wrong example of what I am trying to say. Think of the subsidiary investing in the stock market.

  193. Ramanan writes:

    JKH,

    Or maybe my examples are non-sense. Ignore. I am trying to come up with an example to show that BoP/IIP forces one to define things in the SNA way… Hence the code seems right to me … just that the report doesn’t.

    I’ll try to think of an example and also put some numbers.

  194. JKH writes:

    Ramanan,

    I’ve attempted to shoe horn the best explanation I can come up with as to why the non-financial non-corporate $ 10 trillion net worth is treated the way it is. I can’t get inside the minds of the macro accountants who make the decisions, but the rationale I suggest seems reasonable for now.

    I have no problem with the general idea that the foreign sector should be represented and treated as a net subtraction (i.e. the current actual net profile) from what would otherwise be available for domestic balance sheet asset holdings. In the case of the US profile, there is clearly a subtraction from the pot of assets created by domestic issuers and sellers that would otherwise be available for domestic balance sheet holdings, but that have been claimed by foreigners in compensation for the cumulative current account deficit along with revaluation adjustments, etc. etc.

    As for the criterion that I suggested might apply to the treatment of the non-financial non-corporate sector, I have no problem with the idea that this same treatment would not necessarily apply to the international position. That is not necessarily inconsistent. The domestic financial accounting net worth treatment is applied to the domestic absorption of those net asset positions into the domestic balance sheet of the household sector. But the criteria that determine the treatment of positions on the international balance sheet logically override domestic criteria. This is because the international treatment is logically prior to the domestic treatment – the international facts of where the assets are held and their corresponding treatment eliminates assets from the domestic sphere of holdings, and makes them unavailable for consideration of how they should be treated in the interface between the non-financial non-corporate sector and the household balance sheet for example. So the logical order is international balance treatment followed by domestic treatment, and those two treatments don’t necessarily have to be exactly the same.

    P.S.

    I think the code is wrong, but proceed by all means.

  195. Ramanan writes:

    JKH,

    Till recently, there was no IMA in Z.1. Only a year or two back it got added.

    The Z.1 philosophy is to not treat equities as liabilities at all.

    However for B.1, it is forced to view equities held by foreigners as liabilities. It’s not just about the domestic pot. Suppose on Monday, for some reason, firms’ stocks whose equities are held domestically don’t see much movement but firms whose stocks are mainly held by foreigners see a huge rally. In that case, U.S Net Wealth will change as per B.1. (Thinking of it being recorded daily). But the domestic pot is still the same. So it’s not just about the domestic pot.

    So B.1 actually is not firm about its original philosophy.

    Now to the IMA part. Still thinking of a correct example but I believe that the only consistent way is in a way in which the non-corporate sector has around zero net worth (which of course doesn’t mean they are worthless in an ordinary sense – just that the company owes to the owner) and the code being right.

  196. Marko writes:

    Someone at the Fed read my post #146 , above , and started quaking in their boots.

    Et voila ! :

    http://www.federalreserve.gov/econresdata/notes/feds-notes/2015/us-net-wealth-in-the-financial-accounts-of-the-united-states-20151008.html

    Do I have clout , or what ?

    ;)

  197. Marko writes:

    Just fyi , since I know some of you will be interested , another Feds Note that posted yesterday :

    “Federal Debt in the Financial Accounts of the United States”

    http://www.federalreserve.gov/econresdata/notes/feds-notes/2015/federal-debt-in-the-financial-accounts-of-the-united-states-20151008.html

  198. JKH writes:

    Ramanan,

    I understand your first two points and haven’t said anything or intended to say anything to the contrary.

    Table B.1 is clear on the fact that the net claim of the rest of the world on the US is netted out in order to arrive at national net wealth. So it is the residual domestic pot of equities available for domestic holdings that is relevant to the US net wealth calculation. In your example in which only the value of equities held by foreigners increases due to stock market valuation, the US net wealth will be unchanged, because the increase in the net claim is offset by the increase in the domestically issued value.

    No change – the code is wrong in my view.

  199. JKH writes:

    Ramanan,

    “About your point about coincidence: it’s really a coincidence. The net worth calculated in B.1 uses the market value of equities but there’s no corresponding liability anywhere. So it depends on stock market indices. One could have imagined much higher stock prices, because markets can do anything. And the difference between IMA net worth and FoF net worth can be large.”

    Let me think about that. Maybe I forgot that B.1 is a hybrid.

    Lots to keep track of.

  200. JKH writes:

    The following from Marko’s first link:

    “Unlike in the Financial Accounts, however, in the IMAs, the market value of a sector’s corporate equity outstanding is included as a liability. Thus, the sector net worth numbers shown on table S.2.a Selected Aggregates for Total Economy and Sectors in the IMAs exclude the market value of equity, and therefore differ significantly from those reported for the sectors on table B.1.”

    Specifies corporate equity.

    The non-financial non-corporate sector does not issue corporate equity claims, so there is no liability adjustment to net worth.

  201. Marko writes:

    All in all , I’d rate the new Feds Note as lame , and I’m in a good mood.

  202. The net worth calculated in B.1 uses the market value of equities but there’s no corresponding liability anywhere.

    Right, because it is just a stand-in for the nonfinancial assets of domestic corporations. We do know the nonfinancial assets of domestic corporations, so you might wonder why we would use a proxy for a figure we already have. The problem is we do not know what percentage of domestic corporate nonfinancial assets are owned domestically. Neither do we know how much foreign corporate nonfinancial assets are owned domestically.

    So rather than make further guesses, we simply use with market value of the equities. This works for us both ways, so we can use the total market value of domestic corporations and add the net cross-border equity position to get an estimate of the nonfinancial position of the U.S. vis-a-vis global corporations.

    Point being, there is no corresponding liability presented in B.1 because it (meaning lines 13+26-29) is just an estimate of the corporate nonfinancial assets owned domestically.

  203. The non-financial non-corporate sector does not issue corporate equity claims, so there is no liability adjustment to net worth.

    And here comes the rub. In B.1, nonfinancial noncorporate equity does not appear on the owners’ balance sheets, either. The nonfinancial assets simply appear on line 7. Disturbingly, this equity does appear on the owners’ balance sheets. S.4.a line 117 shows their net worth (shown on S.2.a line 77). But, say, household equity in all noncorporate business shows up on S.3.a line 122.

    Year-end 2014 in billions:

    S.2.a line 77: $10021.3 in nonfinancial assets of nonfinancial noncorporate business
    +B.1 line 12: $17.5 in nonfinancial assets of financial noncorporate business
    =$10038.8 in nonfinancial assets of all noncorporate business
    =S.3.a line 122: $10038.8 in household ownership of all noncorporate equity

    Coincidence? Or double-counting?

    Looks to me as though adding up the IMA household, nonfinancial noncorporate, and government balance sheets and subtract the nonfinancial assets of nonfinancial noncorporate business gets U.S. net worth, short the nonfinancial assets of noncorporate financial business. (Do not count corporate sheets because we want to shift to market value when we reconcile, and IMA corporate balances only reflect that difference between market value and asset value.)

    So at year-end 2014:

    S.2.a lines 76+77+80+81
    + B.1 lines 12-7
    = $77896.4 billion

    This is within $5 billion of B.1 line 1. I would guess the difference has to do with the net foreign ownership of noncorporate business. Here is a comparison of U.S. net wealth and this IMA reconstruction.

    It is ridiculously close most of the time.

  204. Ramanan writes:

    JKH, (#198),

    Don’t understand why US net wealth is unchanged in the example I gave. US liabilities to foreigners has increased without any change in US assets. So the net wealth falls.

  205. Marko writes:

    David ,

    Nice job. Thanks for that.

    So , the blue curve in your graph reproduces the B.1 values. That’s handy to have , at least until the B.1 data moves over to FRED.

  206. Ramanan writes:

    David,

    Surprising they are so close.

    What’s the second link?

  207. JKH writes:

    Ramanan,

    Suppose US net wealth consists entirely of the value of a single US stock S1.

    And the international investment position consists entirely of a single US stock S2 owned by foreigners.

    I understood in your example that the value of S2 goes up and the value of S1 stays the same.

    So US liabilities to foreigners increase but the US net wealth stays the same.

  208. Marko writes:

    Ramanan ,

    David’s second graph is just the difference between the curves , I believe.

    Here’s David’s graph with my long-time shortcut method for national wealth , in yellow. It does fairly well , also , though it runs a little high in the late-80s to early-90s , then around 2000 , but otherwise holds pretty close to the blue B.1 curve.

    https://research.stlouisfed.org/fred2/graph/?g=26gz

  209. JKH writes:

    David,

    The underlying assets of US non-financial non-corporates of $ 12 trillion correctly appears on S.1, not net worth of $ 10 trillion.

    Financial accounting liabilities of approximately $ 2 trillion are eliminated on consolidation within overall S.1. – liabilities issued equal assets held of $ 2 trillion. That’s not inconsistent with a sector presentation that shows $ 10 trillion net worth.

    S.2.a line 77 is non-financial non-corporate net worth of $ 10 trillion – not assets which are $ 12 trillion.

  210. Ramanan writes:

    JKH, (#207),

    In both B.1 and S.2.a, the US net wealth changes in that example.

    In B.1, this affects line 24 and reduces it, thereby reducing US Net Wealth as per its definition.

    In S.2.a, this affects 78 and reduces it and reduces US Net Wealth.

  211. Marko writes:

    The era of qe-infinity has been good not just to domestic asset-holders , but to foreign holders as well :

    http://blog.bea.gov/2015/09/29/value-of-u-s-liabilities-decreased-more-than-u-s-assets-in-second-quarter-2015/

  212. JKH writes:

    Ramanan,

    In B.1 it affects line 13 in the offsetting direction.

    In my example, the domestic equity S2 has gone up in value and increased that line as well.

    So no change in net wealth.

    Etc.

  213. Ramanan writes:

    JKH,

    I was implicitly assuming line 13 is for resident held equities.

    Let me revisit the point after doing some analysis.

    Meanwhile, US Net Wealth as per S.2.a changes in the example right? Reduces?

  214. JKH writes:

    Ramanan,

    Line 13 is the value of domestic equities issued, wherever held.

  215. Ramanan writes:

    Yes, realize.

    My argument is along the line that it won’t survive because of self-contradictions. My previous example didn’t work for B.1, so I am thinking of some other example.

    But it looks strange to me that S.2.a calculates differently and in the example reduces US Net Worth. Not that I find anything wrong: but contrary to the FoF philosophy.

    As I mentioned, I prefer the SNA method. S.2.a is close, code is closer, except for some changes which I think is artificially removed in the published tables.

  216. JKH writes:

    This is a retooling of my # 187, where I overlooked the hybrid calculation nature of B.1

    B.1 line 1 US net wealth $ 79.7 trillion
    S.2.a line 80 Federal Government Net Worth $ (11.7) trillion
    S.2.a line 81 State Local Government Net worth $ 7.9 trillion

    $ 79.7 trillion – ($ (11.7) trillion -$ 7.9 trillion)= $ 83.5 trillion

    And:

    S.3.a line 140 Household Net Worth = $ 83.4 trillion

    I’ve replaced B.101 household net worth with the IMA version S.3.a line 140.

    And on that basis the government net worth adjustment (in brackets above) is the only significant change in the transition between the IMA version of household net worth $ 83.4 and net national wealth of $ 79.7 trillion in B.1.

    Going back to the “code” line items:

    Lines 77, 78, 79, and 82 are all irrelevant to the reconciliation, for various reasons:

    76 is the top line of household net worth of $ 83.4 trillion
    77 as I described earlier is already included in 76
    78 and 79 are the corporate equity adjustments that are irrelevant to both the hybrid $ 79.7 trillion calculation and the household net worth calculation, so they are not a factor
    80 and 81 are the government adjustments are key to the reconciliation
    82 as I described earlier is the international position that is already netted out to arrive at the top line $ 83.4 trillion net worth and therefore irrelevant to the reconciliation

  217. Ramanan writes:

    That’s easy to see JKH.

    National Net Worth = HH Net Worth + Firms’ Net Worth + Government Net Worth

    Firms’ Net Worth is low and negligible compared to others in the equation. So ignore.

    So,

    National Net Worth = HH Net Worth + Government Net Worth

    That’s in SNA language. There are some definitional issues but the IMA numbers are SNA numbers only (except for saying things in a round about way).

    But it’s just because of coincidence that B.1 number is the same as IMA number. (Although for that reason, I find it even more surprising that historically they are close).

  218. JKH writes:

    Ramanan,

    Should be easy to see.

    Which is why adding up all that code stuff to get to the top line makes no sense.

  219. Ramanan writes:

    JKH,

    That code definition is the SNA definition. I don’t find anything wrong with that.

    It’s also possible that the IMA is moving toward SNA’s definitions (as it has been doing for other things) and hence you see the right code but it’s not yet reflecting in the Z.1 quarterly release.

  220. JKH writes:

    Ramanan,

    “Meanwhile, US Net Wealth as per S.2.a changes in the example right? Reduces?”

    no

    the increase in the asset value S2 occurs outside of 76 net worth and outside of B.1 net wealth

    so does not affect 76 net worth or B.1 net wealth

    78 (or 79) becomes more negative

    but 78 and 79 do not affect either 76 net worth (because the stock value as offset is held outside of the domestic economy) or B.1 net wealth

    82 becomes more negative

    but that is offset by the increase in asset value S2

    and that occurs outside of 76 net worth and outside of B.1 net wealth

    so neither 76 net worth or B.1 net wealth are affected

  221. JKH writes:

    sorry

    maybe that’s not right

    I’ll check that later

  222. Ramanan writes:

    JKH,

    B.1 is different from S.2.a.

    What’s the S.2.a US Net Wealth/worth definition?

  223. Ramanan writes:

    JKH,

    http://bea.gov//scb/pdf/2013/04%20April/0413_macro-accts.pdf

    “The organizing framework of the IMAs comes from the System of National Accounts (SNA) 2008.”

    (First page, right column).

    But although most things seem fine, it seems to mess up things around net worth.

  224. JKH writes:

    Ramanan,

    On the run, at risk of error, but regarding your example:

    B.101 net worth no change because no negative equity netting

    B.1 net wealth no change because no negative equity netting

    IMA thingy if non-corporate equity no change because no negative equity netting

    IMA thingy if corporate equity a reduction in total net worth, because of negative equity netting with no domestic offset

    I find juggling the various versions to be a pain, and I think the format of the entire Z1 report is now a mess – very poorly organized top down in that the “integrated accounts” are poorly integrated with the other accounts

  225. Ramanan writes:

    JKH,

    I’ll be out too, so we can discuss this some other time.

  226. Ramanan #210,

    No. Even in this contrived example. Think of line 13 as split between your two types of firms. Call them 13d and 13f. You are saying that 13d is unchanged, but 13f rises. But this is mostly balanced by a rise in line 29. Only if line 13f is wholly foreign-owned is there is no change in line 1. Otherwise, line 1 will also rise.

  227. Ramanan #215,

    Yes, I believe so if the price change is purely due to intangibles. Effectively, the net worth of 78f-type corporations would fall with no corresponding increase in 76.

    The reason why this happens is the IMA does not count (implicitly) intangibles the way B.1 does. B.1 implies that the increase in market value of 13f-type corporations reflects an actual increase in global wealth– some “asset” not considered in the IMA. So if these intangibles or whathaveyou have increased, then the domestic share of ownership raises domestic wealth correspondingly. But in the IMAs a pure increase in the price of 78f-type corporations does not imply an increase in global wealth, so insofar as it increases the wealth of the foreign sector it must lower the wealth of domestic ones.

  228. Steve Roth writes:

    @JKH: “And on that basis the government net worth adjustment (in brackets above) is the only significant change in the transition between the IMA version of household net worth $ 83.4 and net national wealth of $ 79.7 trillion in B.1.”

    VERY nicely done. Demonstrates that HH net worth is a darned good representation of “the wealth of nations” — the very thing I’ve been trying to wrap my brain around in strict accounting terms since we discussed these matters almost four years ago (!!):

    http://www.asymptosis.com/an-mmt-thought-experiment-the-arithmetic-and-political-mechanics-of-net-financial-assets.html

    Note the first two comments there: SRW and JKH. (Ramanan [and Marko] had yet to join us.)

    Which leaves us with a smaller task of much less import, to close this loop among B.1, S.3.a, and B.101: providing an equally simple, clear, and one hopes intuitive explanation for the discrepancy between HH net worth in B.101 and S.3.a — dispensing, as here, with measures that are “irrelevant to the reconciliation.” (Though I fear this reconciliation may be less amenable to simple explanation.)

    Again, thanks. For working the math problem for us. Makes things very clear. In particular, it makes sense in terms of people’s normal (vague) usages — and clarifies those usages.

  229. Steve Roth writes:

    Though I still wonder why “Net Wealth”.

    1. Is there such a thing as Gross Wealth? As always: net of what? (I’m thinkin': ROW.)

    2. It implies a conceptual distinction between wealth and net worth. I, at least, would be hard-pressed to explain that distinction simply and clearly.

  230. Marko #208,

    Yes. The major difference is the gap between equity and nonfinancial assets of nonfinancial noncorporate. That is, (because equity is defined there as net worth) the major remaining difference is the financial net worth of nonfinancial noncorporate.

  231. Steve Roth writes:

    IMO Econ 101 should begin with the B.1, the accounting practices of that “derivation,” and the theory and philosophy of “value” and “utility” (and “ownership”) necessary to truly understand those practices.

    As in the discussions here, necessary economic theorizing arises inevitably from there.

    Microeconomics requires and is based on macrofoundations. Like, really macro.

  232. Marko writes:

    Steve ,

    Re : Economic theorizing – I’ll present here my rough thinking on the “theory of everything” , incorporating as a base concept the Haig-Simons income theory , strongly favored by you and me ( and maybe ONLY by you and me ).

    I think the economy is set up to generate income and wealth ( or net worth ) , and those who do the setting up are largely indifferent in their preferences between a dollar obtained as income vs a dollar obtained as wealth. So , using H-S rather than GDP as the definition of our macroeconomy , we’d have :

    H-S Econ = Consumption + change in net worth , or :
    = ( GNI-Gross Saving ) + change in B.1 national wealth

    If we start in the mid-1940s as FRED does for most data , we can forget about initial values as these are rounding errors in this exercise , so change in NW is simply current NW. Our H-S Econ over the time since then is thus ( thru 2014 , in $trillions , nominal ) :

    = 14.8 + 77.9 = 92.7

    Where does the money come from ? What is the money ? I believe the money that matches our H-S Econ comes from monetary base , credit-market debt ( i.e. domestic nonfinancial ) , and cumulative gross savings , so :

    M = MB + C-M debt + GSAVE

    = 3.9 + 43.3 + 65.9 = 113.1

    So , from a monetarist viewpoint , the velocity of money ( or H-S Econ / M , in this case ) is very nearly 1. If the missing H-S Econ in the forms of offshore wealth and consumption leakage were added back in , it might be even closer to 1.

    Unfortunately , FRED doesn’t have a way to get cumulated GSAVE , so I can only show a graph using the annual flow values. If you move the time slider back and forth , you can see that the annual money flows move in a plausible range relative to annual H-S Econ flows. The outlier areas are the bubble periods which then revert to more normal ranges :

    https://research.stlouisfed.org/fred2/graph/?g=26r1

    This is how the machine works , IMO , or something close to this. And this is why the Fed and the banksters will make sure that for most of us , the only item on the menu will be credit , and it won’t be organic. For our masters ,on the other hand , growing their wealth at the expense of our incomes is organic indeed.

  233. Ramanan writes:

    Marko,

    “M = MB + C-M debt + GSAVE”

    Don’t think that’s right.

    The best way to answer it is using G&L’s models.

  234. Ramanan writes:

    Steve Roth,

    About your question on gross verus net ….

    From the FEDS Notes:

    http://www.federalreserve.gov/econresdata/notes/feds-notes/2015/us-net-wealth-in-the-financial-accounts-of-the-united-states-20151008.html

    “In estimating U.S. net wealth, we use direct measures of the value of households’, nonprofits’, noncorproate businesses’, and governments’ nonfinancial wealth. For corporate businesses, we use the market value of their outstanding equity shares to better capture the value of intangible assets, such as intellectual property. We then net out financial obligations between U.S. resident households, businesses, and government agencies and the rest of the world, because the concept of U.S. net wealth should exclude nonfinancial assets that are financed abroad rather than domestically, and include the value of nonfinancial wealth held by U.S. entities abroad. Taking all this together, we define net U.S wealth as the value of tangible assets controlled by households and nonprofits, noncorporate business, and government sectors of the U.S. economy, plus the market value of domestic nonfinancial and financial corporations, net of U.S. financial obligations to the rest of the world.”

  235. Ramanan writes:

    I think in the FEDS note, it looks to me as if equities are used to estimate the value of non-financial assets … as some kind of proxy.

    It uses the SNA concept of net worth and everywhere there’s non-financial assets except for corporations as it thinks this is a better measure to estimate the value.

    There is however one issue I can think of. A corporation can have a lot of debt and the value of its outstanding stocks needn’t be a measure. In the extreme case, think of a corporation with mainly nonfinancial assets and liabilities which is mostly debt and just some equity. The total outstanding will be a poor measure of non-financial assets of this company.

  236. Marko writes:

    Ramanan,

    I was expecting a suggestion like that , in fact , I expected a more immediate and brutal pushback. I guess I need to work on my provocation techniques. ;)

    I’m not really that interested in getting deep into the accounting weeds on this myself. One common narrative that I’d like to see disabused is the notion that the national wealth , such as it is , is fleeting and can disappear at a moment’s notice because it’s mainly a function of fickle markets. There’s an element of truth to that , of course , but the great bulk of the wealth is as stable as the overall economy , in fact it’s inextricably tied to the productive economy , just as market values of huge blue chip corporations are tied to their business activities. Only the U.S. economy is multiples more stable and secure than even the bluest blue chip , with a historically defined and limited downside.

    Here’s David’s B.1 net wealth reproduction , bracketed by gdp x 3.8 and gdp x 4.6 in dotted red :

    https://research.stlouisfed.org/fred2/graph/?g=26Cw

    If one accepts that , given 70 years of history , national wealth is not going to fall below 3.8xgdp by very much or for very long , and you take 4.6 x gdp as equal to 2014 net wealth , then you can say that 3.8/4.6 x 100% = 82.6% of current net wealth is stable , fundamental value , not likely to evaporate because of haunted animal spirits. That’s $64.3 trillion of the $ 77.9 trillion 2014 national net wealth. It may be a coincidence that that’s so close to the $66 trillion in cumulated gross savings over time , but it would fit a story where it’s the underlying productive economy that’s generating the great bulk of the wealth. I can remember a time when this would not have even been a topic for debate. I think all the talk of QE , helicopters , and negative rates has fried a load of neurons in this country.

    The “fleeting wealth” meme is simply part of the propaganda designed to deflect attention from the extreme wealth inequality in this country. The accounting fiction of outsized capital consumption/depreciation is another similar coverup , not to mention the tax dodge aspect.

    It’ll take a massive accounting rework to accurately portray the patterns of U.S. capital accumulation as experienced by a properly disaggregated population. Long ago Richard Ruggles pointed to the weakness of the current system in this regard , and suggested ways to fix it. Here’s a fairly recent paper that uses Ruggles’ method to look at households’ capital and savings , etc :

    http://www.researchgate.net/publication/265041820_Households'_Capital_Account_Investment_Gross_and_Net_Savings_in_the_NIPA

    We have accounts designed by the 1% , for the benefit of the 1%. That needs to change.

  237. JKH writes:

    Ramanan,

    Your point # 235 is incredibly fundamental to the exercise of understanding everything that has been discussed in these various blog posts and comments.

    Suppose the world consisted of a single US real asset held by a single US corporation with a financial capital structure entirely held by single US household.

    Suppose that financial capital consisted of 9 parts debt and 1 part equity.

    Then US wealth could be calculated according to either the value of the real asset (SNA/IMA?) or the value of the financial capital structure (B.101). (Because I’ve assume away government debt, B.1 would calculate as per B.101).

    The point being that it is not sufficient in either the B.1 or B.101 calculation to include only the value of corporate equity. The value of all net financial claims held by the private sector must be included – debt and equity.

    This point is not an issue of any controversy for B.101, because the calculation of what is included is obvious on the surface – every asset held by households is taken into account one way or another.

    But surely this must be specified for B.1 somewhere?

  238. JKH writes:

    damn; should read above:

    “The point being that it is not sufficient in either the B.1 or B.101 calculation to include only the value of corporate equity. The value of all net corporate financial claims held by the household sector must be included – debt and equity. (Excluding government debt in the case of B.1).”

  239. Marko writes:

    I should have kept it simple in the graph above. This is better :

    https://research.stlouisfed.org/fred2/graph/?g=26dX

    Call the range 3.3-4.4 x gdp , changing the calculation to 75% , or 58.4 of the 77.9 trillion.

  240. Marko writes:

    I’m officially brain-dead : One more time :

    https://research.stlouisfed.org/fred2/graph/?g=26Fn

  241. Ramanan writes:

    JKH,

    Yeah seems B.1 is messed up, right?

  242. Ramanan writes:

    Marko,

    “We have accounts designed by the 1% , for the benefit of the 1%. That needs to change.”

    That is simply not true.

    Here’s a quote by Morris Copeland who did the most work in this:


    The subject of money, credit and moneyflows is a highly technical one, but it is also one that has a wide popular appeal. For centuries it has attracted quacks as well as serious students, and there has too often been difficulty in distinguishing a widely held popular belief from a completely formulated and tested scientific hypothesis.

    I have said that the subject of money and moneyflows lends itself to a social accounting approach. Let me go one step farther. I am convinced that only with such an approach will economists be able to rid this subject of the quackery and misconceptions that have hitherto been prevalent in it.

  243. JKH writes:

    Ramanan,

    To be honest, this is the first time I’ve looked at the IMA section of Z1 with any focus on detail.

    And it’s obviously the first time I’ve seen B.1.

    Given that, my initial impression is that the overarching logical organization of Z1 as a whole is now a complete disaster. I mean overarching.

    Whether or not all the micro reconciliations line up, I don’t know.

    But there are so many loose ends that seem to be evident to a number of intelligent people trying to make sense of the entire thing that something is really wrong with how all this is set up – from the standpoint at least of the overall organization of Z1. And that gets manifested in the new B.1 table.

    I have no doubt that Godley and Laovoie’s overall treatment in terms of the use of financial asset netting etc. is pristine – they’re simply too good for it to be otherwise – although this is not my preferred format for fundamental analysis purposes. And I imagine the SNA and IMA are pretty good in that way as well.

    But this S.1 hybrid thingy looks like a chickens come home to roost dog’s breakfast to me.

    And all of this is why I much prefer the original B.101 as a starting point – where in a sense I’ve been fighting an uphill intuitive battle with the SNA’ers and IMA’ers for several years now.

    (Steve Roth and yourself for example, maybe? Certainly no intellectual disrespect there; it’s a matter of personal preference among some pretty intelligent people.)

    The reason B.101 feels so natural to me is that I find it fundamentally satisfying to consider the lens of the full financial markets and monetary system in looking at the concept of wealth and how it gets transmitted from both the hard real asset and financial asset perspective. And also that the household is the unit in the economy where the transmission of wealth ultimately stops – at the end of the day that’s really because households are the unit does not emit equity financial claims.

    (Note that according to IMA, governments in effect do in effect emit Ricardian negative equity claims in the form of negative net worth. Interestingly, this is anathema to the MMT net financial asset concept.)

    And all of this as you know is tied in a semi-secret way to the S = I + (S – I) boolean perspective on the net financial asset position of the household sector, which is quite different than the MMT net financial asset position of the integrated private sector. And in fact that theme underpins in different words the message of both source posts for this overall discussion – Roth and Waldman.

  244. Ramanan writes:

    JKH,

    Wrote an update post http://www.concertedaction.com/2015/10/11/united-states-net-wealth-part-2/

    Right, MMTers will not like this. They even come close to saying that national wealth is public debt.

    There is however a reason I like a national net worth: because it includes claims vis-a-vis foreigners. Keynes suspected that Mercantilism had an element of truth in it. Mercantilists probably didn’t articulate their points well and wouldn’t accept that their strategy does others harm. It’s beggar-thy-neighbour. Indeed Adam Smith’s book titled “The Wealth of Nations” noticed this but promoted free trade instead. The phrase “beggar-thy-neighbour” comes from his book.

    Whatever the story, there’s always a need for “wealth of a nation”. So an interest to capture it, measure it, study its dynamics etc.

  245. JKH writes:

    Right. Your post is what I described in my example.

    That Fed note is absolutely absurd:

    “Financial wealth is excluded in the estimation of U.S. net wealth because financial wealth represents agreements between parties regarding future payments, such that a financial asset of one party is always matched with an offsetting financial liability for another party.”

    They make that statement, while at the same time including the entire value of the publicly traded US stock market? That’s a massive “exception” to what they purport to be doing.

    Just a ridiculous opening summary of what they are doing. Dog’s breakfast even in the description.

    B.101 takes into account the effect of NIIP. NIIP can easily be looked up as a separate report.

  246. Ramanan writes:

    JKH,

    Yeah the note is funny. Using the same quote, they shouldn’t have included line 24 “Net U.S. financial claims on the rest of the world”.

  247. Ramanan writes:

    JKH,

    There are countries with a “resource curse” in which the nation has a lot of resources but people are poor. In such cases it’s likely that the government (and few firms?) has a lot of assets. That’s another case for statistics on national wealth.

  248. Marko writes:

    Ramanan ,

    My issue is less with Copeland and the FOFs and more with NIPA. For example , if you think financial services ( the playground of the 1% ) are appropriately represented in the NIPAs as to their “contribution” to national output , I think it’s safe to say you’d be distinctly in the minority.

  249. Marko writes:

    I’m curious about an entry on F.2 of the Z1 – on lines 36 and 37 , it looks like there might have been a sizable transfer from gov’t to corporate of ~$120 billion in “nonproduced nonfinancial assets” in Q2 2015. Mineral rights or something like that , maybe ?

  250. Steve Roth writes:

    @Marko: “Where does the money come from ? What is the money ? I believe the money that matches our H-S Econ comes from monetary base , credit-market debt ( i.e. domestic nonfinancial ) , and cumulative gross savings”

    As your questions imply, I think you need to be very clear on what you mean by “money.” I’ve given an exact definition for what I think is a useful meaning of the word as a technical term of art. It’s problematic relative to millennia of vernacular usage, because, for instance, a dollar bill is not “money.” But it works as part of a useful and mutually coherent set of technical definitions (“asset,” “wealth,” etc.) — something that is completely missing in economics. Like physics without agreed-upon definitions of energy or mass.

    I’ve suggested that market revaluations, cap gains, are the primary source of “money.” When the markets go up, there’s more money out there, right? Ab nihilo.

    @Marko: “a dollar obtained as income vs a dollar obtained as wealth”

    I don’t think that’s a coherently formulated concept. A dollar “obtained” by an individual contributes to their wealth. A dollar created contributes to aggregate wealth.

    @Ramanan: “About your question on gross verus net ….”

    Yeah sure. B.1 attempts to estimate the value of our existing real goods, then subtracts our net financial position relative to ROW — in JKH’s words, telescoping ROW’s “financial structure” against an estimate of our real structure.

    But that doesn’t address the central question: why “wealth”? Net worth seems a perfectly adequate term to me. What does the term “wealth” mean or imply?

    @Ramanan: “I think in the FEDS note, it looks to me as if equities are used to estimate the value of non-financial assets … as some kind of proxy.”

    Exactly what I’ve been saying. HH Net Worth (incorporating firm NW via market value of firm equity) is the markets’ best consolidated estimate of what all our stuff is worth, whether or not it’s represented explicitly on balance sheets. A well-educated populace, for instance. That hits entities’ balance sheets through market evaluation/estimation of the “value” of those entities’ assets, though there’s no way to track any transaction flow for that. If aliens gave every American programming skills or higher IQ, all our stuff would instantly be worth more. (Marko: there would be more “money.”)

    @Ramanan: “A corporation can have a lot of debt and the value of its outstanding stocks needn’t be a measure. … The total outstanding will be a poor measure of non-financial assets of this company.”

    The theory is that the markets take that debt into account, and still render the best (only?) possible estimate of that corporation’s value (net worth, replacement value…). To repeat previous: hence, Tobin’s Q is always 1.

    Agree completely with JKH #237: “Your point # 235 is incredibly fundamental to the exercise of understanding everything that has been discussed in these various blog posts and comments.”

    @JKH: “I much prefer the original B.101 as a starting point… looking at the concept of wealth

    Agree that it’s an excellent nexus to start with. Especially when you have clear in your mind the understanding of “real structure” vis-a-vis “financial structure” of accounts that you’ve discussed here. Key to the philosophical/theoretical understanding of accounts/accounting that Econ 101 students should take in with their mother’s milk.

    But I’ll reiterate that we/they/you need to be very clear on what’s meant by “wealth” if we want to avoid talking (and thinking) past each other.

    @JKH: “that theme underpins in different words the message of both source posts for this overall discussion – Roth and Waldman”

    Just to say that even though we’ve worrying at this for nigh-on four years, I at least don’t think we’re spinning our wheels. I can imagine (only a fond imagining, perhaps) a reformulated MMT emerging from this thinking that is more coherent, more usefully (and intuitively) explanatory of how economies work, and more rhetorically and politically powerful.

    @Marko: “F.2 of the Z1 – on lines 36 and 37 , it looks like there might have been a sizable transfer from gov’t to corporate of ~$120 billion in “nonproduced nonfinancial assets” in Q2 2015.””

    NPNFAs are a very sticky wicket. Note that government land, for instance, is not included in the estimates of gov NW.

    We really needed to start with a balance sheet of all the world’s stuff, before humans created any value or laid any claims to that value.

    LHS Assets: Stuff
    RHS, no liabilities, only “shareholder” equity.

    Then we tally the value of annual captured sunlight… ;-)

  251. Steve Roth writes:

    Just add: the “World” balance sheet I describe at the end rather embodies the view of things that is at least professed by many early indigenous peoples, who found the concept of ownership rather mysterious.

  252. Ramanan writes:

    Steve R,

    “The theory is that the markets take that debt into account, and still render the best (only?) possible estimate of that corporation’s value (net worth, replacement value…). To repeat previous: hence, Tobin’s Q is always 1.”

    What I meant is something like this http://www.concertedaction.com/2015/10/11/united-states-net-wealth-part-2/

  253. Steve Roth writes:

    @Ramanan:

    Great post, love especially your emphasis on the conceptual in the final sentence. What do these measures and terms mean?

    But in your para 3 you seem to use wealth and net worth synonymously:

    “…the meaning of the measure of the “U.S. Net Wealth.” The definition is similar to the System of National Accounts 2008 (2008 SNA). The net worth of a nation is…”

    My difficulty is not with the construction of these measures. I understand them perfectly and re-explaining it doesn’t help me.

    So my question remains unanswered and unaddressed: why “wealth”?

  254. Steve Roth writes:

    Basically the Fed sed: we can give estimates of “real”-asset values for everything except firms. For firms, we need to throw up our hands and just take the markets’ best guess.

    A perfectly sensible decision, IMO.

    What’s not stated: that market equity estimate embodies not just the tally-able assets that we can see in firms’ book values, but the value that is attributable to the firms’ environment — an educated, healthy populace, effective government, etc. Because that environment determines, to a great extent, the volume and value of firms’ future output. If that environment improves, so does firms’ “value.”

  255. Steve Roth writes:

    Optimally (arguably), all real-asset values should be estimated by the market. But since noncorp firms aren’t traded, that’s impossible.

    It’s worth noting, though, that the other big asset category, real estate, is estimated using market indices. That’s as it should be, IMO.

    As a result, most of the changing value resulting from “the environment” is revealed, in these tallies, via the revaluation of corp equities and real estate. That’s how owners of equities and real estate pick up the “rent’ from that environment, adding it to their balance sheets.

    Owners of noncorporate firms also pick up rent, of course, but since those firms’ value is tallied book value, the rent component of their “income” is largely invisible — it all appears, rather, via the changing market estimates for, cap gains from, equities and real estate.

    Again, see Jorgenson, Hulten, Hall, etc. on the “zero-rent model” architecture of the national accounts.

    https://books.google.com/books?id=rqM2GtTl9NcC&pg=PA211&lpg=PA211&dq=%22zero+rent%22+accounting&source=bl&ots=k9zadIrCkk&sig=mda5uNlVGPzqhnoVjfhYAW5eCxA&hl=en&sa=X&ved=0CDAQ6AEwA2oVChMI1sKitN-_yAIVSfZjCh1wmwNd

    https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=5&ved=0CDIQFjAEahUKEwjWwqK037_IAhVJ9mMKHXCbA10&url=http%3A%2F%2Fwww.nber.org%2Fchapters%2Fc0137.pdf&usg=AFQjCNEz6cAXu6zMKllg66Hcev7IIsOZpw&sig2=Rp3QkGf7KjJyXrB5ws64Ug

    https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=6&ved=0CDwQFjAFahUKEwjWwqK037_IAhVJ9mMKHXCbA10&url=http%3A%2F%2Fwww.nber.org%2FCRIW%2FCRIWs04%2Fhulten.pdf&usg=AFQjCNELYNbaxJPln0DT5GtdF0p52RRIHA&sig2=8gb95boaKScsRMiYCHzoFQ

  256. JKH writes:

    Once again, B.1 seems fundamentally incoherent to me, along the lines of my # 237 comment example. Ramanan made essentially the same point near the conclusion of his second post.

    It seems incoherent not only in concept but according to empirical data presented in Z1 as a whole.

    There are at least 3 parts to the incoherence:

    a) At the end of 2014, the value of non-financial corporate equity was stated at $ 23 trillion (B.1. and S.5.a). The value of non-financial real assets was stated at $ 20 trillion (S.5.a). Not bad, but nonsensical in concept. The reason being that for any given value of corporate equity, a firm can either raise debt to buy back stock, or raise equity to pay off debt. This has absolutely no effect on real assets, yet corporate equity changes in both cases. Why should we believe that corporate equity is anything but a fluke estimator of real asset value? It makes no sense.

    Moreover, here are the extraordinary numbers for the particular case of Apple:

    In its recent Q3 results, Apple reported real assets on its balance sheet of about $ 25 billion. The book value of its equity was about $ 125 billion. Earlier this year, the market cap of Apple reached about $ 750 billion. That’s a 30 times corporate equity to real asset value ratio. And yet that’s the number that would have been plugged into B.1 on the same date.

    http://www.apple.com/pr/library/2015/07/21Apple-Reports-Record-Third-Quarter-Results.html

    b) At the end of 2014, the value of financial corporate equity was stated at roughly $ 7 trillion (B.1 and S.6.a). The value of financial corporate real assets was stated at something less than $ 2 trillion (S.6.a). This illustrates more clearly how disconnected the two measures can be in reality. Financial services is obviously light in real assets, but heavy in the generation of equity value because it is so services oriented.

    c) B.1 simply aggregates the corporate equity values of a) and b). This seems bizarre, since the financial services asset portfolio in aggregate certainly includes non-financial corporate equity holdings (within total financial asset holdings of $ 85 trillion!). How does this not engage in obtuse double counting?

  257. Steve Roth writes:

    @JKH: “a firm can either raise debt to buy back stock, or raise equity to pay off debt. This has absolutely no effect on real assets, yet corporate equity changes in both cases. Why should we believe that corporate equity is anything but a fluke estimator of real asset value? It makes no sense.”

    Because the market presumably takes all of that into account, holds up its thumb and squints, and decides what the firm (hence the equity in the firm) is worth. Flawed as that might be, I truly can’t imagine any better estimate. There is far too much unaccounted for in book value, for instance, to even get close. Think, for instance, of that favorite buyout accounting catchall asset, “good will.”

    The book-to-market discrepancy you point out is arguably far more due to the hopeless efforts of “book value” to impart net worth, than to the efforts of the market at estimating that value.

    In particular, book value will and can never impart a going concern’s future ability to capture economic rents — which I’m gonna suggest is one very valuable asset.

  258. Basically the Fed sed: we can give estimates of “real”-asset values for everything except firms. For firms, we need to throw up our hands and just take the markets’ best guess.

    That’s very obviously wrong, given that the Fed presents such estimates for firms. The reason for using market value for corporations in B.1 rather than nonfinancial assets is that we don’t have a good estimate for how much of domestic corporate nonfinancial assets are ultimately owned domestically versus abroad. Likewise, we have no good idea how much foreign corporate nonfinancial assets are ultimately owned domestically versus abroad.

    It is far easier to track general equity holdings.

  259. Steve Roth writes:

    @JKH:

    Adopting the excellent terms you bruited in #103:

    The B.1 attempts to depict the domestic “real capital structure,” only resorting to “financial capital structure” for ROW, as a necessary accounting convenience. (It assumes that the market value of equity is the best possible estimate of firms’ “real” value.)

    By contrast look at the Sources and Uses and Assets and Liabilities FOF matrixes, which depict only the financial capital structure. The real capital structure is completely invisible there — viz the complete absence of real estate, for just one very large example.

    None of us would suggest that the financial abstraction in the FOFs matrix “makes no sense”…

    (Though I would say that its failure to balance with net worth and ∆net worth makes the presentation confusingly problematic.)

  260. Ramanan writes:

    Good analysis JKH.

    You are saying that in addition to the liability side analysis (the debt/equity example), there’s an asset side issue also with B.1. Such as with the example of Apple and financial companies.

    I think the table B.1 and the FEDS note only makes sense in the approximation that firms only have non-financial assets in assets and equities in liabilities. I think they have been reading textbook models wrongly. Some textbook model may have such a thing but that’s just a model and the Table B.1 seems to literally take it as the real world.

  261. Steve Roth writes:

    @David: “That’s very obviously wrong, given that the Fed presents such estimates for firms. The reason for using market value for corporations in B.1 rather than nonfinancial assets is that we don’t have a good estimate for how much of domestic corporate nonfinancial assets are ultimately owned domestically versus abroad. Likewise, we have no good idea how much foreign corporate nonfinancial assets are ultimately owned domestically versus abroad.”

    ?? We have a perfectly good estimate of GIIP — though not for individual companies. Error of composition?

  262. JKH writes:

    Steve R.,

    Sorry, but with respect, I’m not sure you understand how the capital transactions I referred to work by necessity.

    When a firm issues debt and does a share buy-back, it actually shrinks not only the number of shares, but both the book value and the market value of equity.

    The reason it does a buyback is to increase earnings per share – because the cost of debt capital is less than the cost of equity capital.

    But the impact of the transaction is that the amount of equity capital will shrink as a result of the share buyback – in both book and market value terms.

    This is basic financial accounting.

    I can give you a further numerical example if you like.

    The issues you raise are extremely tangential to the reality of what must happen to the numbers in question – the latter directly concerning the conceptual veracity of the B1 estimates/substitutions at a very basic level.

  263. ?? We have a perfectly good estimate of GIIP — though not for individual companies. Error of composition?

    I think we have good numbers on financial transactions. That’s why we compute B.1 as total nonfinancial plus net financial (net being both net of liabilities and net vis-a-vis the foreign sector.)

  264. Steve Roth writes:

    @JKH: “When a firm issues debt and does a share buy-back, it actually shrinks not only the number of shares, but both the book value and the market value of equity.”

    And the market bids up the shares becuz same total value less shares. U know that. ???

  265. Steve Roth writes:

    By total value I mean NPV capitalized value of future earnings. Which are unchanged by the buyback.

  266. By total value I mean NPV capitalized value of future earnings. Which are unchanged by the buyback.

    I think it is fair to speculate that the additional debt will lower that NPV.

  267. Steve Roth writes:

    @David: “I think we have good numbers on financial transactions.”

    Right, as I said above: they resort to financial capital structure for ROW because deconstructing ROW’s real capital structure is unrealistic, or just not worth it given its fairly small impact on the net wealth estimate.

  268. JKH writes:

    Steve,

    “And the market bids up the shares becuz same total value less shares. U know that. ???”

    That’s not correct Steve.

    David,

    “I think it is fair to speculate that the additional debt will lower that NPV.

    That’s correct. But it’s not speculation. It’s elementary corporate finance.

    I’ll be back later with a simple example.

  269. Steve Roth writes:

    Apple:

    Real assets: $25 billion
    Book value: $125 billion
    Market value: $750 billion

    How much is Apple “worth”?

    Two of these estimates are wildly wrong. Which ones?

    Now Apple buys back $5 billion in stock — either cash out of pocket or borrowing to finance it. In either case:

    New Book value: $120 billion, down 8%.

    Does anyone want to suggest that Apple’s future earnings prospects have declined by 8%, and that market value should also decline 8%, to $690 billion?

  270. Steve,

    I guess I still do not understand what your point is at #254. Your first statement there still seems indefensible on its face. Perhaps you might clarify?

  271. That’s correct. But it’s not speculation. It’s elementary corporate finance.

    Since we cannot actually measure the NPV, it’s not that elementary. But basic accounting makes it seem pretty likely– reducing the net wealth (before equity liability) of the firm is probably going to reduce the net flows to the (owners of the) firm.

  272. Now Apple buys back $5 billion in stock — either cash out of pocket or borrowing to finance it.

    Your question sidesteps the issue. Before the buyback, Apple had $5 billion in retained earnings that they have effectively distributed now rather than the future. Thus, the future earnings prospects indeed are lower by $5 billion. I am pretty certain that prices almost always fall after distributions– hence the need for dividend adjustments.

  273. Sorry, I meant future distributions, obviously.

  274. Steve Roth writes:

    @David:

    First, did one piece of arithmetic wrong: it’s 4% not 8%. But the story’s the same.

    Capitalized future earnings prospects completely overwhelm one-time distributions. The miracle of compounding interest and all that. As exemplified in the Apple’s 6:1 market-to-book ratio.

    “I am pretty certain that prices almost always fall after distributions”

    You might want to do some research into that.

    Again, are you saying that a $5-billion stock buyback will cut Apple’s earnings prospects and market value by 4%, $30 billion? That seems to be what you’re suggesting. Yea or nay?

  275. Adding, on top of the immediate change in capacity for future distributions, a buyback may signal lower growth and lower capacity for future earnings (for not having been reinvested.) This may further depress the share price– especially if the price was high in anticipation of front-loaded distributions (and hence a high NPV prior to the buyback.)

  276. Steve Roth writes:

    @David:

    You just exited the arithmetic, entered speculation land (in two senses of the word “speculation”).

    Suppose a company maintains (roughly) zero cash, uses all earnings over time to buy back all its shares, except one.

    How much is that last share worth?

  277. Again, are you saying that a $5-billion stock buyback will cut Apple’s earnings prospects and market value by 4%, $30 billion? That seems to be what you’re suggesting. Yea or nay?

    I am saying it surely cuts its prospects for future distributions. As I wrote,

    Your question sidesteps the issue.. the future [distribution (see #273)] prospects indeed are lower by $5 billion.

  278. Steve,

    One share only makes for a very illiquid market. You want me to speculate on how the sole owner values the firm? I would guess at least the liquidation value of the firm as that is presumably an option. Could be worth a tremendous amount or nothing. Maybe less than that if the owner does not want to deal with the hassle of owning.

  279. JKH writes:

    Here’s a simple example (as simple as I can think of), serving to illustrate the point I made earlier in # 262.

    Assume a firm with book value assets of $ 200.

    Financed by book value $ 100 debt and book value $ 100 equity.

    Assume to make things simple the market value of debt and equity are also each $ 100.

    To simply, no taxes.

    (For later reference, assume the interest rate on existing debt is 10 per cent and the return on equity is 20 per cent. So the starting aggregate interest cost is $ 100 and the starting aggregate earnings are $ 200. We don’t really need these numbers to explain the intuition on what happens, but just in case.)

    Suppose the firm buys back $ 80 in stock and finances this with more debt.

    So the book value capital structure is now $ 180 in debt and $ 20 in residual equity.

    Miller-Modigliani is the starting point for what happens.

    And in this context, the starting point is that the firm’s “enterprise value” was $ 200.

    This is the firm’s total market value, according to the market value of its financial capital of both debt and equity.

    Nothing has changed after the buyback with respect to the firm’s operations or how it generates the total cash flow required to service its debt interest and earn its return on equity. The buyback is just financial engineering.

    Under idealized assumptions, MM states that nothing happens to the enterprise value merely because of financial engineering of the capital structure. Intuitively, the interest rate cost of issuing debt will increase as the firm buys back more stock. This is because the lower equity capital buffer that remains after the stock buyback exposes the debt holders to a greater risk that their interest will not be paid – due to potential downside earnings and cash flow volatility. So in order to sell debt of $ 80, the market will demand a higher interest rate. That 10 per cent rate will increase on the incremental debt that is sold.

    So MM suggests that under ideal conditions, the enterprise value in the example will remain $ 200, and so the new market value of equity will only be $ 20. This does not mean that earnings per share won’t increase. Expected earnings per equity share remaining will still increase, because the average interest cost on the $ 80 in debt is still less than the previous cost on the $ 80 of equity it displaced. (Perhaps for example the interest cost increases from 10 per cent to 15 per cent for the last dollar of debt – junk bond status.) The leverage still works in terms of increasing the expected return on equity and expected earnings per share. But despite increased expected earnings per share, the stock price will not increase proportionately. That’s because those expected earnings are now more at risk due to the increased leverage – so the market will not accord the same price/earnings multiple as before.

    The firm can attain its objective of increasing earnings per share under MM, but in this example the market value of equity will not increase beyond its residual book value (now $ 20), because enterprise value in total doesn’t change, and $ 80 of new debt has displaced $ 80 of old book value equity. The stock price remains the same under MM in this example, because the higher risk to earnings results in a lower valuation of each dollar of earnings than before. So the new market values in the example are $ 180 for debt and $ 20 for equity, with proportionately fewer equity shares outstanding earning a higher return on equity.

    (The cost of debt will reach the cost of equity at the limit as the buyback goes to the entire stock float. That increase in interest cost is necessary in order to attract the amount of debt sufficient to fund the buyback. The increased interest cost of the additional debt of $ 80 will be a marginal drag on what would otherwise be the resulting increase in earnings per share, but there will nevertheless be some EPS increase because the interest cost on the $ 80 will still be less than the original cost of equity on the $ 20 that remains. In my base case example, I assumed the marginal interest cost on the debt increased from 10 per cent to 15 per cent – still less than 20 per cent previous return on equity.)

    MM is idealized under certain assumptions, but it is the lift off point for the real world logic. Whether or not the stock price itself goes up in the real world depends on imperfections that exist beyond the MM simplified assumptions. But the stock price in the example would have to increase by 500 per cent after the share buyback in order to get back to the original equity market cap of $ 100, which isn’t going to happen. In a world of market imperfections, the buyback and the leverage effect may possibly increase the stock price, but obviously not to that degree. My main point in the earlier comment was that the market value of equity will decline after a stock buyback – which in this idealized example is represented by a decline in total equity value from $ 100 to $ 20. In the real world of imperfections, leveraging up that equity might increase the resulting number from $ 20, but certainly not back to $ $ 100.

    I’ve tried to keep this very simple as an illustration. For example, stock buybacks where the existing market value of equity is greater than its book value are more “expensive”, other things equal. The per share book value of the stock that remains will actually be lower than before, due to the cost of buying back the stock at a premium. And that reduced book value per share will also be a depressing drag on its resulting market value even from what was just calculated, other things equal.

    In summary again, the interest rate cost on the debt will increase as more debt is floated. The expected return on equity will increase because the cost of debt is still less – that’s the leverage effect. But the stock price won’t go up proportionately because the stock is so much riskier. And therefore the full enterprise equity value loss due to the stock buyback will not be recouped by an increase in equity share price. In the simple example, the stock price would have had to have increased by 500 per cent in order for the equity market value not to decline as a result of the buyback. And that would have increased the enterprise value from $ 200 to $ 280, which doesn’t happen under MM and certainly won’t happen to such a degree in the real world.

    I think some of the confusion in some of the comments above is that enterprise value is being mistaken for the market value of equity. In particular, even if the enterprise value remains unchanged after a share buyback (unknowingly implied in some comments above), the market value of equity will decline a fortiori because some equity has been displaced by debt within that unchanged enterprise value.

  280. JKH writes:

    SR,

    “Now Apple buys back $5 billion in stock — either cash out of pocket or borrowing to finance it. In either case: New Book value: $120 billion, down 8%. Does anyone want to suggest that Apple’s future earnings prospects have declined by 8%, and that market value should also decline 8%, to $690 billion?”

    If Apple buys back its stock through debt financing, its aggregate earnings will decline due to the additional interest cost of the debt. Earnings will not decline by 8 per cent, because the cost of debt is less than the cost of equity. But earnings will decline by some amount, because the cost of debt isn’t zero. Shares outstanding will decline due to the buyback and the market value per share may increase if the benefit of expected return per share from lower cost leverage outweighs the cost due to the marginal increase in risk due to more leverage. But the aggregate market value of equity will decline, other things equal, because the interest cost of debt still isn’t zero.

    I went through all that in my generic example. Apple is not immune from the logic of that example, merely because of its rather extreme balance sheet and market value configuration. The numbers are obviously very different to start, but the directional causalities due to change in capital structure don’t change.

    The point of the Apple data was to illustrate an extreme case of a fundamental error in using the market value of equity as a proxy for the value of underlying real assets. The point was not that Apple serve as a proxy for some standard quantitative reaction due to a buyback. Apple is obviously an extreme case in many dimensions of its balance sheet.

  281. JKH writes:

    correction near the beginning of # 279 above:

    (For later reference, assume the interest rate on existing debt is 10 per cent and the return on equity is 20 per cent. So the starting aggregate interest cost is $ 10 and the starting aggregate earnings are $ 20. We don’t really need these numbers to explain the intuition on what happens, but just in case.)

  282. Steve Roth writes:

    @JKH: M-M just sez: with two identical firms, the leveraged one has the same value as the unleveraged one. (In a somewhat idealized world where, in particular, investors borrow at the same rate as firms.) Capital structure is immaterial to value.

    This begs the question: what’s the firm’s value?

    B.1 assumes that the market’s estimate is the best estimate. My question: what estimate is superior to that?

    I think you’re objecting because this estimate breaks closed-loop accounting, which is totally true. Because the Revaluation Account has no sources. New value/assets simply appear(s) on balance sheets as a result of the market’s improved expectations. The market decides (discovers?) that the stuff we produced in previous periods is worth more than we thought it was.

  283. Ramanan writes:

    Steve,

    JKH’s point with his examples are that lines 13-15 in Table B.1 has huge conceptual problems.

    And JKH has also defined things such as enterprise value and market value of equity. I am not sure why you then ask “what’s the firm’s value”.

  284. Steve Roth writes:

    @Ramanan:

    “lines 13-15 in Table B.1 has huge conceptual problems”

    Agreed, in that it breaks closed-loop accounting. Becuz, the Revaluation Account.

    “I am not sure why you then ask “what’s the firm’s value”

    What should we post to the LHS of households’ or “the nation’s” balance sheet?

  285. Ramanan writes:

    Steve Roth,

    True revaluations are important. But it’s not necessary to bring revaluation in highlighting conceptual problems in lines 13-15 of table B.1. You can of course create examples using revaluation to show it but in the latest examples above, revaluation was not all that important.

  286. Steve,

    Revaluation is not terribly mysterious.

    Household revaluation of financial assets matches very well changes in the market value of domestic corporations. And– if in the longer run– market value of domestic corporations does rise with the net worth (before equity liability) of those corporations. Any way you want to theorize about the price of equities, the general IMA net worth of corporations is pretty small.

    Then there is revaluation due to inflation. It is not surprising that the price of, say, land, would rise likewise. If you begin with start of period land in start of period dollars, and end of period land in end of period dollars, then of course some revaluation will be necessary. It is a measurement issue. If the PCE-adjusted price of land does not change, then the market is not deciding simply that land “is worth more” but that dollars are worth less consumption.

    Yes, markets are capricious. But revaluation is in great part inflation and “windowed” real investment and acquisition of net financial assets.

  287. Steve Roth writes:

    I keep asking: What value for corporate firms should we post to the LHS of households’ or “the nation’s” balance sheet? What better estimate should be used in B.1 instead of the market value of equity shares?

  288. Ramanan writes:

    Steve,

    Well, households’ balance sheet in Z.1 looks fine. Don’t know why you ask.

    It’s B.1 which is problematic conceptually.

    But I do not know why you ask about revaluations because it can be understood even without bringing in revaluations – as has been done in the latest comments by JKH.

  289. I keep asking: What value for corporate firms should we post to the LHS of households’ or “the nation’s” balance sheet? What better estimate should be used in B.1 instead of the market value of equity shares?

    That depends on what you are trying to measure. Is wealth tangible assets, plus net financial assets? Do firms have intangible assets or liabilities beyond that? Do households? Does government?

  290. Steve Roth writes:

    >it can be understood even without bringing in revaluations

    Right. Revals — which emerge from outside the flow-of-funds matrix on pages 1 and 2 of the Z.1 — are only necessary to explain the balance-sheet changes from period to period.

    All going back to my point about MMT: since the sectoral balance presentation — like the FOF matrix — ignores revaluation, it has an opaque relationship to balance-sheet/net-worth changes.

    For me, the B.1 is a conceptually intuitive and explanatory top-down pathway into the pyramid of tables below it (so it’s appropriately placed right up front).

    I’ll ask yet again: what should be changed in the B.1, and in what way? What important misconception(s) does it promulgate? Or, is it an impossible exercise that should be excised from the Z.1?

  291. It’s B.1 which is problematic conceptually.

    I still do not understand what the problem is. If we define wealth as nonfinancial assets, plus net financial assets, then B.1 line 13 simply suggests that corporations themselves have no net wealth. In this sense, negative corporate IMA net worth implies that there are uncounted corporate-owned nonfinancial assets.

    B.1 footnote 1 basically says wealth is tangible assets, net financial assets, and intangible assets of domestic corporations. We don’t have to count this way, and you may want to count differently, but the counting does not seem problematic given the offered definition.

  292. All going back to my point about MMT: since the sectoral balance presentation — like the FOF matrix — ignores revaluation, it has an opaque relationship to balance-sheet/net-worth changes.

    This makes no sense to me. What presentation, specifically, ignores revaluation? For example, revaluation can shift the sectoral balances at any point in time, but I do not see how this can cause the financial balance identity to fail.

  293. JKH writes:

    In answer to # 287:

    I would approach the question top down, starting with a higher level choice:

    There are basically two versions or alternative methodologies for looking at the wealth of the US in a broad way.

    a) The household net worth balance sheet B.101

    b) The SNA/IMA/Godley and Lavoie approach to ultimate real asset wealth with comprehensive negative financial asset netting

    Before looking at those two alternatives, I would note that in my view S.1 is an abomination – because as I described earlier it is a hybrid of those two methodologies, resulting in a poisonous conceptual brew of inconsistency and ugliness.

    Your question relates to how to do b).

    If I were doing b), I would use the real asset values that I identified in my comment # 256. The value of non-financial corporate real assets is $ 20 trillion from S.5.a. The value of financial corporate real assets is $ 2 trillion from S.6.a. From what I’ve seen in supporting explanatory text, I think their reasons for using equity claim values instead are very flimsy indeed, and result in the conceptual nonsense I’ve described.

    If the macro accountants want to do a b) version, they should do it right, using available real values, while noting any qualifications they may have about the appropriateness of those values. And if they don’t like those values, they should work on getting better ones. But they should not pursue this hokey mixed menu game that they do in their B.1 version.

    That said, my personal view is that b) is simply not an important version, however it’s done. (It works for Godley and Lavoie as a mere netting device that they use for other modelling purposes.)

    My preferred methodology is and always has been a), using B.101. I’ve explained why this is the case previously.

    So I would continue to use B.101 as the dominant report. B.101 already includes both the net effect of the international position, and the intersect with the government. Note then that the full private sector NFA position is passed through from the consolidated private sector depiction to the household level depiction, where it is embedded in the much large NFA position that is inherent at the household sector level.

    Then do a footnote to that B.101 on the alternative negative equity way of doing the government sector. That footnote would suggest that you have a supplementary choice – do it either as NFA (which is already inherent and embedded in the B.101 household presentation) or do it the negative equity way (in which case you adjust the B.101 net worth position by netting away the “negative equity” measure inherent in the excess of government debt over its real assets. My comment earlier shows this moves household net worth from $ 86 trillion down to about $ 83 trillion, I think. But you do all that in a footnote to B.101., giving people a choice as to how they want to think about the government sector, without wrecking the existing integrity of the main B.101 report.)

  294. Ramanan writes:

    JKH,

    I am not sure how B.101 takes care of the international investment position.

    One can imagine two scenarios.

    1. Households with assets and liabilities and no outside world.
    2. Households with the same assets and liabilities and the government is the only debtor to foreigners. No foreign assets for either the govt or households/private sector.

    The two situations are different from a national viewpoint. Of course the real world is more complicated but the example illustrates that just looking at household balance sheet won’t do.

  295. Ramanan writes:

    “no outside world” as in no foreigners.

  296. Ramanan writes:

    “For me, the B.1 is a conceptually intuitive and explanatory top-down pathway into the pyramid of tables below it (so it’s appropriately placed right up front).”

    SR,

    I’d simply calculate it as nonfinancial assets plus the net international position.

    The reason B.1 does all this juggling is because it thinks its a better way to capture intangibles, since it doesn’t calculate it by other methods. But anyway its method is wrong.

  297. Before looking at those two alternatives, I would note that in my view S.1 is an abomination – because as I described earlier it is a hybrid of those two methodologies, resulting in a poisonous conceptual brew of inconsistency and ugliness.

    To my eye, the IMAs are very consistent. More so than B.1. I think, however, that there is something significant about the fact that S.2.a does not offer a simple total. If we take net wealth to be the total of (recognized) nonfinancial and net financial assets, then U.S. net wealth would be the total of lines 76 and 79-81. Which I believe we have already established as effectively identical to B.1.

    Note also that B.101 line 41 is exactly S.2.a line 76– $83425.7 billion at year-end 2014.

  298. Marko writes:

    I guess the question becomes whether the different net worth aggregation methods could diverge widely in the future , because the past history shows pretty good agreement , whichever method you choose. By pretty good , I mean variations of less than 10% or so. I see the different measures as roughly equivalent to the idea of using of GDI and GDP to measure the same “thing”.

    Here’s four ways of doing it , all set to 100 in 1990. It might be worthwhile to play with setting the “100 date” to other times to see how the divergence changes :

    https://research.stlouisfed.org/fred2/graph/?g=28Ts

    The four methods are :

    Series 1 – blue curve : David’s reproduction of B.1 “national net wealth”

    Series 2 – red dashed : David’s summation of B101 hh net worth with all other sectors’ net worth ( except ROW )

    Series 3 – yellow curve : B 101 hh net worth plus federal and state gov’t net worth

    Series 4 – green dash-dot : B 101 hh net worth only

  299. Marko writes:

    Oops , the description of the series 2 line is incorrect , as you’ll notice , I’m sure. At any rate , I think I’ll try making the same graph using the method David just described in # 297 and substitute that for Series 2.

  300. Marko writes:

    David ,

    Here’s the same graph but with your suggested ( in #297 ) method substituted in the series 2 – red dashed curve. It doesn’t match B.1 as well as you’d expect , I think.

    https://research.stlouisfed.org/fred2/graph/?g=28Uy

  301. Marko,

    I think that is because I had a typo. 78, not 79.

  302. Marko writes:

    Finally , as I’m not sure anyone has yet posted this graph , here’s one with David’s B.1 proxy as series 1 ( blue ) , and all of the line 76-82 S.2.a variables ( red dashed ) , to do with as you will.

    Note that if you simply add 76 thru 81 , ignoring ROW , you get a 2014 net wealth that’s $10 trillion too high , due to double-counting as we determined way back when. You can then deduct ROW which tends to reduce that 2014 discrepancy , but adds deviation overall , resulting in lousy agreement – as shown in this version :

    https://research.stlouisfed.org/fred2/graph/?g=28W7

    To get the best match to the B.1 values using the S.2.a data , you can add gov’t( Fed plus S/L ) to hh , with or without financial sector net worth , and get good agreement. Once you start adding the other sector data , it starts to diverge.

    What this all means , IHNFI.

  303. Marko writes:

    David,

    I suspected as much after I started graphing it. Thus this last post (#302).

    It seems like the proper thing to do is to add all but the noncorp nonfinancial , as you suggest , and to deduct ROW , and to then live with the differences , understanding that we’re getting to roughly the same place using different methods. I’d have more confidence saying that if they hadn’t made that little $ 10 trillion error , however.

    On the other hand , the danger in using shortcuts like hh alone , or hh plus gov’t as I’ve typically used , is in not catching it when those approximations break down at some point because they’re incorrect proxies from an accounting standpoint.

  304. Marko,

    I think it means simply that non-corporate equity is already on the owners’ IMA balance sheets so it needs not be counted again.

  305. Marko writes:

    One thing to keep it mind for those of you who expect things to add up – Zucman has determined that the missing wealth in the U.S. is on the same order of magnitude as the NIIP , i.e. ~$4-6 trillion. So , embedded in all these stats are probably financial assets that don’t match financial liabilities – in big numbers – certainly enough to cause deviations of a $trillion here or there with respect to alternative means of calculating net worth.

    In other words , don’t tear your hair out over this. Luckily , I don’t have to worry about that.

  306. Marko writes:

    David ,

    Yes , I know. I’m the one who set off that avalanche and associated loss of life. See #146 and forward.

    By “error” I mean the Fed should have at least footnoted line 77 in S.2.a to that effect , thus safeguarding against such tragic outcomes.

  307. JKH writes:

    Ramanan # 294,

    Not sure I follow you there.

    If the government debt is held entirely by the private sector, then household sector net worth includes the NFA contribution of that debt either directly or indirectly as a net asset component, other things equal.

    If government debt is held entirely by the foreign sector, thereby constituting a marginal NIIP liability, then private sector net worth does not include that debt as an asset component and neither does the household sector as part of the private sector. So household sector net worth is less than the first case by the amount of the NIIP liability. So the household sector reflects the effect of NIIP compared to the counterfactual.

    That’s what I mean by “B.101 already includes … the net effect of the international position …”

    Whether or not you want to look at the real assets of government is a separate issue. If so, then you’re into ultimate real asset accounting as per SNA, IMA, GL, or perhaps the horrific B.1. But I’m not assuming that conversion to a real asset focus when I say the NIIP is reflected in B.101. I’m saying that NIIP is reflected coherently in the net position of B.101, in the same way that holdings of government debt would be. In other words, both NFA sources are treated coherently in the context of B.101.

  308. Marko writes:

    JKH,

    You reminded me of something I forgot in my post #303 reply to David , above : that ROW is NOT deducted in the published method for combining IMA sectoral net worths , as shown in the series analyzer :

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL892090025&t=S.2.A&bc=S.2.A:FL112090205&suf=A

    As you say , B 101 hh net worth performs pretty well as a proxy for B.1 national net wealth. The main divergence between the two measures has occurred since the GFC , and though it may be entirely coincidental , that divergence can be almost entirely eliminated by subtracting monetary base from hh net worth , as shown here (dotted line ):

    https://research.stlouisfed.org/fred2/graph/?g=292O

  309. Ramanan writes:

    JKH,

    Not like that. Foreigners hold government bonds and nothing else.

    So the comparison is between two cases.

    1. public debt is 40% of GDP and the households hold it directly.
    2. public debt is 140% of GDP. Households hold 40% as in 1 and foreigners hold 100%.

    Almost everything else is the same in the two worlds above even GDP.

    Foreigners’ holding of public debt is not seen in households’ balance sheet. So national net worth in case 2 is bad which household balance sheet doesn’t show. Or not yet.

    In other words, households balance sheet may be in good health but a nation’s balance sheet may be deteriorating.

  310. Marko,

    Yes. I was just clarifying. Also, we need not deduct ROW. We simply do not include, just as with nonfinancial noncorporate. Right?

  311. Marko writes:

    David ,

    Yes , that would also be my understanding of the best combo if you want to sum the IMA aggregates , given what we know so far , anyway. Something tells me it’ll be a while yet before we’re satisfied that everything is properly nailed down. Maybe the Fed will put out a comprehensive guide that will make it all fit together.

  312. JKH writes:

    Marko,

    “That ROW is NOT deducted in the published method for combining IMA sectoral net worth, as shown in the series analyzer”

    I believe that’s the point I made back at my # 216:

    “Line 82 as I described earlier is the international position that is already netted out to arrive at the top line $ 83.4 trillion net worth and therefore irrelevant to the reconciliation”

    And further back at # 187

    “Line 82 is not part of the IMA aggregation calculation. This is the net asset position of the rest of the world as a positive claim on the US. It is excluded from aggregation because its netting effect is already reflected in the residual net worth of those assets that do remain available to US claimants. So this item can be ignored – it does not need to be subtracted in aggregation. To do so would be double counting the netting away of the foreign claim from a household net worth that already takes this into account.”

  313. JKH writes:

    Ramanan,

    And # 312 is essentially the same point I’ve made about NIIP relative to B.101

  314. JKH writes:

    Ramanan,

    Consider this example.

    First consider a 3 sector model.

    And first consider the interface between the government and the private sector.

    The government runs up a debt of 140 per cent of GDP.

    Refer to this as 140 as a value for convenience.

    Let’s hold the foreign sector interface and NIIP unchanged at this stage of the analysis (i.e. “other things equal” for now), so we can assume the government has done this by spending money to acquire goods entirely from the private sector.

    Then the marginal (MMT) NFA position of the private sector with the government is 140.

    Next, overlay on that the assumption that the private sector runs up a 100 net debt to the foreign sector, via trade between the two etc.

    Assume the private sector sells private sector assets to settle this position. Say, it sells US corporate bonds. So the foreign sector holds US corporate bonds as its claims.

    Then the marginal NFA position of the private sector with the foreign sector is (100).

    And the total consolidated NFA position of the private sector is 40.

    So that calculation fully takes into account the effect of both the government debt and NIIP on the private sector.

    And that private sector NFA position will show up also in the household sector as an essential core of the private sector position as per B.101. For example, the income and saving from income generated by the associated cumulative deficit must show up there. Either the income goes directly to the household sector in the form of wages etc. or it goes to business sector capital in the form of capital income on which the household sector has ultimate claim through increased claims on the business sector capital structure. And the cumulative negative saving associated with NIIP will show up in a similar way.

    So everything shows up as an effect on the household sector up to this point. That is the point I’ve made as it applies at this stage of the analysis.

    Now, finally, to deal with the configuration you presented:

    Suppose the foreign sector does a swap in which it exchanges 100 corporate bonds for 100 government bonds.

    That’s just an asset swap. There is no change to net value positions anywhere.

    And the full result now is the same as your configuration.

    So nothing is changed in terms of the issue under discussion. The effect of both the government deficit/debt and the foreign sector deficit/NIIP shows up on the household balance sheet B.101.

  315. JKH writes:

    Ramanan,

    “In other words, household balance sheet may be in good health but a nation’s balance sheet may be deteriorating.”

    In your two examples, the net position of the household sector is the same and the NIIP in particular has been netted out from the household balance sheet. It doesn’t need to be deducted a second time. That has been the point I have made all along.

    If your point is that the government debt is 40 per cent in one case and 140 per cent in another, that’s fine, but that doesn’t make my point wrong.

    One just needs to go and look at NIIP and go and look at total government debt as data searching exercises separate from what appears in the household balance sheet. That doesn’t contradict anything I’ve said. I’m just saying that the household balance sheet takes into account net worth issues as they are telescoped from the consolidated private sector into the household sector. And those net worth effects include the effect of both the government sector and the foreign sector. I mean, the household balance sheet doesn’t show the asset composition of the central bank either, so one can just go and look that up if interested. The issue I’ve pointed to is that the household balance sheet shows the private sector net worth effect through the lens of financial intermediation quite nicely – as an alternative to the SNA/IMA/GL/B.1 ultimate real asset exercises.

  316. JKH writes:

    Separate point on reflection:

    A point I made earlier on in a slightly different way is that the ultimate real asset approach to national wealth is naturally opposed to the notion of MMT NFA – i.e. in opposition to the MMT rationale for private sector NFA generated by government budget deficits. That’s because the ultimate real asset methodology ignores NFA of that origin (different from NIIP NFA). The “net NFA” position delivered by a budget deficit is effectively zero according to the ultimate real asset methodology – because it nets to zero by construction. The private sector NFA effect from government budget deficits is eliminated. That is arguably an anti-MMT methodology.

    And another point I made earlier is that the NFA insurance perspective presented by SRW in his post here is actually separate from the issue of NFA per se. So that point is separate from the issue of choosing between the private sector NFA perspective and the ultimate real asset methodology. Governments and central banks don’t need to run deficits in order to manufacture low risk government assets. They can just do asset swaps with the private sector (e.g. TARP, and qualitative easing as per the Bernanke first round.)

  317. JKH writes:

    SRW,

    Regarding your long ago response # 150 about the depiction of insurance, I believe that cluttering basic balance sheet presentation with the paraphernalia of high risk scenario analysis is dysfunctional and misleading accounting. Things like worst case TARP scenarios should be captured in separate supplementary boxes – not in the basic balance sheet presentation of transaction value (effectively swaps of government debt for high risk assets priced as high risk assets in that particular case of TARP). Jamming extreme contingencies into a standard balance sheet presentation of asset and liability values doesn’t work. It should be transparent but supplementary analysis.

  318. Ramanan writes:

    JKH,

    MMTers confuse the sectoral balance analysis. So just forget them.

    Sectoral balances analysis is good. MMT uses it in a clownish way.

    “In your two examples, the net position of the household sector is the same and the NIIP in particular has been netted out from the household balance sheet. It doesn’t need to be deducted a second time. That has been the point I have made all along.”

    There’s no netting. The NIIP reflects in the government debt position. From just looking at households’ balance sheet it is not possible to deduct a country external position.

    I created two different worlds to show households’ positions are the same but at the same time, the government and external position are deteriorating in the second. There’s no indication by just looking at households’ balance sheet that there is anything wrong going on.

  319. Ramanan writes:

    And not only is there no netting to begin, neither am I saying it should be netted.

  320. JKH writes:

    Ramanan,

    “There’s no indication by just looking at households’ balance sheet that there is anything wrong going on.”

    I think we’re talking past each other.

    Nowhere have I said that just by looking at the B.101 can you can confirm that “there is anything wrong going on”.

    What I said is that the effect of the NIIP on net worth is already inherent in the B.101 presentation (as is the government cumulative budget position), without further adjustment for the NIIP. The marginal effect of a current account deficit or a NIIP deficit is indeed netted out from the household balance sheet when COMPARED to the COUNTERFACTUAL case of a balanced external position. There is no question about that as far as I’m concerned, and I demonstrated an example of that fact above.

    One can then at both NIIP and government balance sheets in a supplementary way to see if “there is anything wrong going on.” I’ve not only acknowledged that, I’ve suggested several times that should be done in conjunction with looking at B.101.

    Finally, and quite separately as it relates to the issue of methodology for net worth aggregation, one has choices as to how to treat the netting effect of government as it connects to the private sector or household balance sheets.

    One option is to adjust household net worth by adding on the net liability position of government according to the SNA/IMA/GL presentation. For example, using very round numbers, suppose government debt is $ 20 trillion and government real assets as identified are $ 15 trillion. That’s a net liability position of $ 5 trillion. The effect of adding that $ (5) trillion to household net worth is to replace $ 20 trillion of government sourced MMT private sector NFA with $ 15 trillion of identified real asset value on the balance sheet of government.

    So that’s an OPTION in terms of government sector treatment as juxtaposed against household net worth.

    But there is no further adjustment required for the NIIP position as it factors into household net worth. It’s already there, as per my example.

    We’re talking past each other because the point I am making has nothing to do with revealing the inner balance sheets of either the government sector or the foreign sector. It just saying that the net position contribution of each is already reflected in B.101, with the OPTION of adjusting the net government effect alone the lines that I just described.

    (BTW, you know that I’m quite aware of the weakness of MMT, and have been roundly critical of some of their expositional techniques. But I’ll reference their core concepts when it suits for illustration purposes, thanks.)

  321. JKH writes:

    Ramanan,

    Simple example focusing on NIIP
    For illustration, assume government in stock and flow balance

    Counterfactual:

    Current account in balance
    NIIP = 0
    Household net worth = X
    No further change

    Factual:

    Counterfactual open, plus the following change:

    Household spends Y to buy T-shirt made in China
    Current account = (Y)
    NIIP = (Y)
    Household net worth = X – Y

    (Household has spent Y amount of net worth on consumption)

    So household net worth includes the netting out of NIIP
    Which is only what I have said all along

  322. Ramanan writes:

    JKH,

    There are implicit assumptions in your comment 321.

    Because once the household buys T-shirt in China, the household and Chinese producer and the Chinese government can do a lot. And not to mention fiscal policy of the US.

    The household can buy government bonds with its saving. Fiscal policy can be expanded simultaneously to reach the previous direction of rise of output in response to a fall because of opening up of the external sector.

    So it isn’t true that “So household net worth includes the netting out of NIIP”. And precisely because of that reason, one needs a nation’s wealth – another table to look at. One can then compare the two and do all sorts of dynamic analysis.

  323. JKH writes:

    By the way, I’m sure you know this but still perhaps worth noting, the configuration in which the foreign sector holds 100 in government bonds as its marginal NIIP position doesn’t mean that the government has run a cumulative bilateral deficit with the foreign sector. It means its run a deficit financed by bonds, and the foreign sector has accumulated those bonds in exchange for other private sector assets it originally held as a result of its own international surplus position in that same example. That’s reflected in my example # 314. I suppose that’s obvious, but again worth noting.

  324. Ramanan writes:

    JKH,

    Yes and my point too.

    Households may have bought imports directly without the government in between but yet the outcome is the second of the two worlds in my scenario.

    Households accumulate 40% of the government debt and foreigners accumulated 100% of GDP because of cumulative budget deficits, even though the government didn’t run a cumulative bilateral deficit with the foreign sector.

    So my underlying point is that households’ balance sheet doesn’t say what the dynamics of the wealth of the whole nation is and for this purpose we need another table in addition to households’ balance sheet.

    About your point “Then the marginal (MMT) NFA position of the private sector with the government is 140.”

    But that’s MMT usage of sectoral balances. The right way to do it is to talk of the domestic private sector’s net stock of financial assets which is 40.

    My other point is that the sectoral balances approach is a fine approach.

  325. JKH writes:

    Ramanan,

    “About your point “Then the marginal (MMT) NFA position of the private sector with the government is 140.” … But that’s MMT usage of sectoral balances. The right way to do it is to talk of the domestic private sector’s net stock of financial assets which is 40.”

    That’s not MMT usage. That’s a fact of composition in any usage. It’s the marginal contribution of the government position to the total position, which consists of the government plus foreign sector contributions. Private sector balance is the inverse of the sum of the government and foreign balances, which is 40 = 140 – 100 in this case.

  326. Ramanan writes:

    JKH,

    Right.

    Going back to my example. Think of the 100% debt held by foreigners (all government bonds) as the factual. Just by looking at the household balance sheet cannot tell you whether it is this case or whether it is the case where the public debt is just 40% and there’s no debt vis-a-vis foreigners.

  327. JKH writes:

    Ramanan,

    Right.

    I’ve tried to make it clear for ages now that I never said anything different, and that the information contained in those balance sheets should supplement B.101.

    But my point has always been that B.101 includes the NET EFFECT of the actual government position and the actual NIIP.

    That must be the case – because the sector balance equation shows the dependence of any one of them on the other two, and because the interface of the consolidated private sector with the other two is reflected directly or indirectly at the household sector level.

    That’s what I thought you were questioning.

  328. Ramanan writes:

    JKH,

    Great, we are on the same page.

  329. Marko writes:

    JKH @ 312 ,

    Sorry if that ( #308 ) sounded like it was a reminder directed at you. You’ve been on top of the ROW issue all along. It was meant to correct my own error a few posts earlier , where I suggested deducting the ROW value in summing the IMA aggregates. That was simply a brainfart on my part. Clearly the ROW value , along with the noncorp nonfinancial value , should not be part of that summation , i.e. they should be ignored.

    The other points you’ve raised ( e.g. on the adjustments to hh net worth that have already been made re: gov’t ) make me also wonder if the other IMA aggregate values ( like corporate and financial ) should be added to the hh value , or whether the best alternative to the B.1 estimate of national net wealth is simply the B.101 hh net worth.

    I wish there was a published value for this item : FL892090035.A

    http://www.federalreserve.gov/apps/fof/SeriesAnalyzer.aspx?s=FL892090035&t=S.2.A&bc=S.2.A:FL112090205,FL892090025&suf=A

    Then , at least , we could see what sort of difference in results between their two methods of national wealth determination that they’d be content to live with.

  330. JKH — I don’t disagree at all that it might be best to keep an accounting of the most straightforward, least contingent insurance-providing assets at the center of a presentation, and then considering more contingent assets in supplements, or alternative “comprehensive” tables, or whatever. If I have a deep meta-point to make in all of this, it’s that there’s no such thing as a true accounting, and good accounting consists of multiple view for different purposes. Sectoral-balances derived NFAs, I claim, do not offer a useful accounting of what people usually mean by savings or net worth. They do offer a useful, but partial, measure of a certain kind of insurance available to individual, unaggregated units within a domestic, private sector (and the insurance might meaningfully affect the aggregate behavior of the sector even though it is not a valid measure of its aggregate net worth). As you say, the traditional analysis is at best a partial measure, and supplements or alternatives might definitely be informative. There’s a lot to be said for leaving simple measures simple, because though they may be “wrong”, they are more likely to be stable in their definitions, and we can learn to adjust for their wrongness when we note relationships between our accounting constructs and the real world phenomena we hope they capture.

    I would revise, a bit, my response to you on bank deposits. Bank deposits are a bit hybrid: they provide NFA-like insurance to the non-bank entities that hold them, but they are leverage, negative insurance to the banks that issue them. I don’t think it’s useful to aggregate these two effects to zero, they are different effects. Effects on the financial leverage of banks and effects on the perceived safety of households might effect different phenomena in different ways, might contribute to risk-bearing via household insurance, financial fragility and risk aversion within the financial sector, or both, with effects difficult to predict “on net”. So, I think agreeing with you, I’d say that if we are interested in looking at financial sources of systematic risk insurance, we’d want the traditional sectoral balances presentation first (which does have a history of some stylized correlation with macro phenomena like recessions, so has some evidence for validity), and then we’d categorize other, more contingent or mixed potential sources elsewhere.

    That’s all prologue. Mostly I want to say, to my dismay, I haven’t found an error with respect to your claim that B.1 is incoherent because it omits corporate real assets in excess of equity value. If you and now I are not wrong, B.1’s net worth estimate will be ridiculously sensitive to capital structure choices of corporations. As you say, the right measure of intangible-including corporate value is enterprise value (defined as mkt value of liabilities + mkt value of equity – financial claims against other domestic sectors here). There’s a real mystery. This is not a subtle point. Unless you and I are missing something, it is a no brainer. Why on Earth would the Fed have adopted this definition? We must be missing something, but i confess that I am missing it.

  331. I’ve just sent the following note to one of the authors of the FEDS note on Table B.1. I’ll report back if I hear something:

    Hi!

    Thank you for your great work, and your recent FEDS note on Table B.1.

    I host a financial blog where sharp-eyed commenter (JKH) pointed out an objection with the definitions in Table B.1 that I find myself unable to address. [ Towards the end of the 300+ comments here http://www.interfluidity.com/v2/6174.html#comments ]

    You define the value of domestic corporations as the market value of equity, rather than looking through to real assets, on the perfectly defensible theory that the market value of claims will capture intangible assets that would otherwise be unmeasurable.

    But JKH points out that market value of equity is not the right measure for this. It ought to be a measure of enterprise value, more specifically something like

    market value of equity + market value of liabilities – market value of financial assets

    If JKH and I are not missing something, using market value of equity alone renders your measure unreasonably susceptible to changes in capital structure. Suppose a firm holds real assets whose comprehensive value is $100 and begins as an all equity firm. Its value, by your procedure, would begin at $100.

    Now suppose the firm issues $20 in debt and buys back $20 in shares. Now its market value of equity is only $80 (to a first approximation, under Modigliani and Miller reasoning, putting aside tax assets and contingent deadweight insolvency costs). Yet the real assets of the firm, its contribution to the net wealth of the country, have not changed at all. We can find no place in B.1 where this loss of equity value might be offset by a gain in some other sector. But it seems entirely unreasonable that the portion of real assets accounted for by liabilities in a firm’s cap structure would not contribute to net wealth.

    Are we missing something?

    Many thanks!

    smiles,
    Steve Waldman

  332. JKH writes:

    Steve,

    “There’s a real mystery. This is not a subtle point. Unless you and I are missing something, it is a no brainer. Why on Earth would the Fed have adopted this definition? We must be missing something, but i confess that I am missing it.”

    Glad you agree.

    Not subtle – quite so.

    I’m guessing wild that the response may observe some (very rough) consolidated macro correlation based on the data (very rough indeed), but that just begs the question – why go to the effort of attempting a better estimate of real asset value by resorting to such a spurious hit and miss relationship. If there is an “invisible hand” that guides macro corporate America to the goal of matching up expected market value of equity to the “true” value of underlying real assets, I’m not aware of it. And just look at the example of Apple as one contributor to that relationship, along with many other minimal real asset holders. It just makes no sense when capital structure is viewed as a conceptual sum of variably weighted parts.

    In addition, as far as equity capitalization just on its own is concerned, I struggle to see why the present value of the future profit – the stream of future profit derived from the future revenue garnered from those who purchase the future product – should come very close at all to the present value of real assets currently held, however that real asset value is measured. After all, the amortization of that real asset value over time is only one contributor to the production process and associated cost as input to the generation of such future revenue and profit.

    On that last point, perhaps Pickety and others have said something directly about this. Is this Marx territory? I don’t know.

    I recall a conference talk given a long time ago by the late Peter Bernstein in which he observed that the economy was getting “lighter” on a secular basis – meaning lower real asset density, so to speak. I don’t know to what degree that’s truly the case, but it seems intuitively associated with increased services penetration and technology advances over time. And this is consistent with equity market capitalizations that become less connected to underlying real asset values, and more connected to the present value of future consumption (or future investment in proportion). And that seems consistent with excess conventionally measured saving desires as a secular feature.

    BTW, it’s been a while since I was schooled on the subject, but I’ve tended to think of enterprise value as the value of the entire capital structure, without taking special account of the asset structure – i.e. without netting away financial assets. But I don’t know.

    Finally, it is perhaps ironic returning full circle that your equation is in effect carves out the “net financial asset” position associated with the firm as an entity – i.e. the firm at the margin displaying a negative net financial asset position (financial assets – financial capital structure) that becomes a positive net financial asset position to the holders of its gross financial capital structure. (I think “net financial liability” is potentially confusing terminology.) This sort of thing is also pertinent to the overall interpretation of B.101 and what lies beneath.

  333. JKH writes:

    Marko # 308

    Marko,

    My views in summary:

    B.101 should be the center of the universe for all of this analysis.

    One should be immediately aware that B.101 can be interpreted essentially as a conceptual “reverse takeover” of the entire private sector balance sheet – because whatever shows up on the business sector balance sheet gets transformed/transmitted down to the household sector by the latter’s holdings of the net capital structure of the business sector both directly as ultimate financial claims on balance sheets with real assets and indirectly though claims on financial portfolios held by the business sector itself. So this includes all financial capital after netting out intra-business holdings of financial capital. BTW, the table B.101.e that shows the household balance sheet with equity details – including a breakdown of direct and indirect equity holdings – is very interesting and very instructive on this perspective.

    I would supplement B.101 with boxes that show the balance sheets for the government and the external sector. Such balance sheets are found in the IMA section and elsewhere in the B. sections. As Ramanan frequently notes, and especially for the international sector, these are important balance sheets. However, as I have also noted several times, their net profile effect in terms of “NFA” is reflected in a complementary way in the profile of the private sector, and by retraction from there to the profile of the household sector and B.101.

    And I would supplement similarly with boxes that show the balance sheets of the corporate and non-corporate business sectors, etc.

    So in this overview, real asset holdings of all sectors – except direct household holdings – become supplementary but interesting data.

    I put things in that perspective because I think that real asset values that exist outside the household sector, while interesting to note and ponder, are really secondary to the overall analysis of what is driving behavior from the household core of a monetary economy. And in particular I think that the volatility of the market value of financial capitalization is something to be observed and respected for what it is in the overall analysis – because of the effect it has directly and indirectly on household wealth perceptions and realities – and not to be effectively ignored as is done with that range of methodologies that prefer financial asset netting in an attempt to get to real asset values.

  334. JKH,

    Why turn tMAs to supplement B.101, rather than start with S.3.a?

  335. JKH writes:

    David,

    One reason is that B.101.e is very instructive, and is reliably congruent with B.101.

    More generally, B.101 has served the purpose extremely well over the years, and I don’t see any immediate value in switching from B.101.

    (Side point – a difference in the final net worth result, which I haven’t been sufficiently in to chase down.)

  336. (Side point – a difference in the final net worth result, which I haven’t been sufficiently in to chase down.)

    What difference? As far as I can see B.101 line 41 = S.3.a line 140 exactly.

  337. In fact, B.101 line 41 and S.3.a. line 140 have the same series code. I do not see how a difference is possible.

  338. Ramanan writes:

    I am not sure about the discussions here about difference in net worth between two calculations and the various FRED graphs linked here.

    As the discussions above show, B.1 has an issue about national net worth calculations because of equities used (see SRW’s comment for summary).

    So why this comparison?

  339. JKH writes:

    David,

    You’re right. I mismatched comparative dates in overlooking B.101 quarterly reporting. Thanks.

    That’s good news for me.

    But as a side point, not relevant to my reasoning prior to that.

  340. JKH,

    I do not agree.

    It looks to me like you either supplement the IMAs with B.101.e, or you supplement B.101(.e) with almost all of the IMAs. Given the equivalence of B.101 and S.3.a, I simply do not see a good reason for starting with B.101.

    The only trouble I can see when mixing B.101.e and S.3.a (at least, trouble beyond mixing B.101.e and B.101) is that S.3.a bumps money market shares out of Deposits and into Equity. That seems a small price to pay for a consistent approach across sectors.

  341. Adding,

    I do think this is an incredibly minor disagreement at this point.

  342. JKH writes:

    David,

    You may be right.

    Perhaps I’m just a curmudgeonly B.101’er.

    I’ll have a closer look at the IMA suite over time.

  343. So, a rather noncommittal response from the Fed economist I wrote:

    Thank you for taking such an interest in our publication. We are always trying to expand what we provide to the public. We have taken note of your question and will think about your comments and how they relate to what we were trying to show.

    I think JKH is basically just right. Which should diminish ones enthusiasm for Table B.1. Although frankly, I think the question of “what is our aggregate net wealth” without specifying some purpose to which one intends to apply the answer, is just plain badly formed. There is no correct number. But there are plainly incorrect procedures, procedures that produce values so arbitrary that for nearly any purpose they’d be hazardous, and I think JKH has correctly identified that B.1’s treatment of corporate equity qualifies as one of those.

    JKH — I think “enterprise value” is a particularly well defined accounting term because, in my experience, it is not consistently defined. It implies a valuation at market value through the liabilities + equity side of the balance sheet, but I have seen it used to refer, as you describe, to the entire L+E (basically total firm assets including intangibles, at market value), L+E-mkt value of financial assets (an estimate of real assets at market value), and L+E-cash_and_cash_equivalents (mkt value of noncash assets). The last one is often used in value investors “free-cash-flow/enterprise value” estimates as a mostly cap-structure invariant replacement for earnings yield (or the informationally equivalent P/E ratio).

    I like its multifarious definition, because I think though things may be plainly wrong in accounting, they cannot be plainly right, and accounting exercises always involve discretionary choices that must be adapted to the purposes of each exercise. There is no one true “enterprise value”, we define it differently as we ask different questions.

    In this case, I think it’s pretty clear that the EV we want is one net of financial assets. Why? Because if we used the full L+E side of the balance sheet, then we would add net wealth to the bottom line of B.1 by having corporations borrow to lend to households. Corporate A = L+E would then expand, while since financial obligations are “looked through” in the household obligations, there’d be no offsetting contraction elsewhere. I hope it won’t be uncontroversial if I assert that we don’t want to include the gross quantity of that sort of operation in our measure of net wealth.

    So, we should net out assets that are financial claims on domestic sectors. Should we net out financial claims on foreign obligors? That depends on how we’ve measured our position with respect to the foreign sector. If we have already included the foreign financial assets of domestic corporations in computing that position, then we should net the out in our measure of domestic corporation value, to avoid double counting (and the same problem of financial operations creating arbitrary wealth). If we have not included foreign financial assets of domestic corporations in our foreign sector position, then we should leave them out here. (There is an analogous issue on the liability side: some of the L+E that serves as our baseline for computing EV may be claims of foreign shareholders or creditors. We’ll have to take care not to double count those. The current B.1 methodology does claim to adjust for that.)

  344. Ramanan writes:

    SRW,

    (So I am going to comment despite saying I don’t intend to comment on Twitter :-))

    “Although frankly, I think the question of “what is our aggregate net wealth” without specifying some purpose to which one intends to apply the answer, is just plain badly formed.”

    I don’t see why one cannot do that. The SNA does that consistently IMO. It is the sum of the value of non-financial assets plus the net international investment position.

    FoF borrows that concept but its calculation procedure is wrong. (It tries to estimate the value of non-financial assets of corporations using the market value of equities, unlike SNA).

    It may be used for example for asking by comparison: Nation A has so much wealth but why is its GDP so low. Or such thing.

  345. JKH writes:

    SRW,

    “Although frankly, I think the question of “what is our aggregate net wealth” without specifying some purpose to which one intends to apply the answer, is just plain badly formed. There is no correct number.”

    I agree 100 per cent and more. Any answer to that question must be context and methodology specific. Moreover, I wince at the idea that such a general and naked question be considered as purposeful in itself. There are only qualified interpretations in attempting an answer – which points to the futility of the desired generality and associated quests for such identifications.

    “There is no correct number. But there are plainly incorrect procedures”

    I agree – VERY important distinction – maybe THE most all-important observation in all of this.

    Regarding “enterprise value”:

    “It implies a valuation at market value through the liabilities + equity side of the balance sheet.”

    If I understand, that was my intended meaning, although I guess I wasn’t clear. In other words, a market valuation of the entire right hand side of the balance sheet instead of equity alone. This would not directly involve a separate left hand side valuation.

    But again I agree regarding the importance of a flexible approach that depends on formulated context and methodology.

    “In this case, I think it’s pretty clear that the EV we want is one net of financial assets.”

    Absolutely.

    And I agree also with your last point regarding potential double counting around the international balance sheet. I think the foreign sector is in some ways the trickiest and most delicate to verbalize accurately. It’s not that easy to grasp to start with, and there can be further misunderstanding in communication. But I think the B.101 methodology should be pretty clean because of this. For example, if Apple has a direct investment overseas, the value of that investment will be reflected through in some way in Apple’s financial capital structure, including debt and/or the market value of its equity. The domestic ownership of that capital structure is reflected on the B.101 household balance sheet – directly through claims on Apple and indirectly through claims on claimants of Apple. Conversely, for example, foreign ownership of Apple stock will simply not be reflected on the domestic B.101 balance sheet. And I think the corresponding NIIP positions will show up as a direct US investment asset claim on the foreign sector, and foreign ownership of Apple stock as a foreign owned equity claim on the US. So that should be a clean reconciliation of NIIP with B.101, where NIIP positions are either translated where required to financial capital versions, or reflected where appropriate as corresponding financial capital values. I hope I have that example right – but if not, Ramanan may well appear shortly to upend it with vigour.

  346. JKH writes:

    Ramanan,

    “FoF borrows that concept but its calculation procedure is wrong. (It tries to estimate the value of non-financial assets of corporations using the market value of equities, unlike SNA).”

    Here’s an important distinction:

    B.1 makes that type of error, as identified.

    But B.101 has never claimed to be estimating the value of real assets held outside of households. Why should it? It’s a household centric view. Households care directly about the value of their stocks, bonds, mutual funds, and pensions. They let the corporate and institutional world worry about the real asset values they are managing.

  347. Ramanan writes:

    JKH,

    Right. B.101 is fine.

    My point earlier was that it’s possible a nation’s wealth is deteriorating but households’ wealth looking good. Or the other way round. It’s an anti-thesis of the efficient market hypothesis.

    I know these could be studied by separately studying the international position but people don’t connect these things. So if given extra table and making it easier for them to analyse is a good step.

  348. Steve Randy Waldman writes:

    Ramanan — Basically, I don’t think that the notion of national wealth is remotely well-formed. The only way you can define it is by the accounting procedure you use to define it. Nonfinancial assets + NIIP is simple and fine from a certain perspective, but it presumes you have a meaningful way of defining the value of nonfinancial assets. I don’t think we do have a generally meaningful way, along several dimensions. One is the dimension that B.1 tries (self-defeatingly) to address, the question of intangibles. Another is the question of how to value even fixed, tangible assets. At cost? Least of cost or market? An estimate of market value? Is it coherent to talk about the market value of “assets in aggregate”, when obviously there can be no buyer of assets-in-aggregate (except hypothetically a foreign buyer, but I don’t think it’s persuasive to say we are defining the price a foreign buyer would pay to buy a nation’s aggregate wealth). If we use cost or market prices, aren’t we making a terribly consequential mistake by substituting prices observed at a margin when small quantities are bought and sold and attributing that as the “value” of gargantuan quantities? In deep ways, I don’t think that the intuitions of ordinary accounting are straightforwardly meaningful when applied to aggregate accounts. (And the intuitions of ordinary accounting are much more slippery than people often appreciate.)

    Even in more narrow, micro accounting questions, I don’t think accounting values are meaningful without matching the accounting procedures that you choose to the questions you intend to answer. There are too many legitimate degrees of freedom in accounting choices to simply call one set of numbers “true”. We can generate one set of numbers, “audited accounts” that is normative according to certain conventions and regulations. But we are making an error if we imagine we can treat them as unproblematically the right values to use for addressing any question.

    For aggregate accounts, these questions are even more urgent. I’m not trying to make a sweeping claim that aggregate accounts are never useful. The post that headlines this comment thread tells a particular story about why a particular accounting exercise might be meaningful, and it’s a story I think is broadly right. But if you tell me you want a measure of “aggregate net wealth”, of the world, of an economy, I’m going to want you to tell me precisely what you intend to use it for. And it’s going to be hard. If, as your example suggests, we are going to use it to make comparisons of wealth to GDP and examine countries that seem to have a high “gross return on wealth” versus a low one, we can certainly do that. But we will either take some existing measure of aggregate wealth and treat that as normative, or else we will have to think really hard about what we mean and realize there are traps embedded in the question.

    For example, consider two countries, one with strong unions and high labor costs, one without. They have identical GDP and identical fixed capital in real terms. However, the price of capital goods in domestic currency is lower in the strong union country (because the returns to capital are impaired by strong labor getting a bigger piece of the surplus). We can say they’re GDP/Wealth ratios are identical, because in real terms, we see the same real product from the same real capital. (We get the same result if we use “world prices” to value GDP and capital.) Or we can say the strong union nation has a higher GDP/Wealth ratio, because if we measure things at domestic prices, that’s what we will appear to see. There’s an underlying reality that remains true in either case, and if we understand the accounting choices we’ve made, we can look through them and see that reality. But the fact of the matter is that these exercises are really complex, and in real life, very often what happens is we make an accounting choice, draw the conclusion that pops out most obviously from that accounting choice, and treat that as “truth”. We either say these economies are very similar (and miss an important difference), or more misleadingly imagine one has more capital depth than the other when in fact the quantity and technology of their production is identical.

    (Note that another of the icks of B.1’s treatment of corporate wealth, now that JKH has set us on this path, is that it is subject to precisely this kind of slipperiness. According to B.1, America becomes poorer when labor unions are strong, because higher incomes to workers aren’t capitalized into any asset while an impairment of stock prices flows into net wealth. Accounting choices have their politics as well.)

  349. JKH writes:

    “higher incomes to workers aren’t capitalized into any asset while an impairment of stock prices flows into net wealth”

    That may be the granddaddy of epistemological accounting conundrums – capital is capitalized and labor is not.

    Or consider:

    Capitalizing the present value of future GDP, in which case quite apart from the problem of labor capitalization, not only present capital stock but future expected capital stock production must be capitalized as at a point in future time, that being net of depreciation on outstanding capital stock at the time, and then each such future tranche being discounted back to the present

    (ouch)

    All of it requiring inquiry context, qualification, and methodological specification for sure

  350. Ramanan writes:

    SRW,

    I understand the point about B.1. In fact I was the one pointing out the debt/equity complication and the error. See my comment #235.

    But that’s a separate issue. The error of the Federal Reserve cannot be used against arguing the point in general about a measure of national wealth.

    The 2008 SNA guide unstats.un.org/unsd/nationalaccount/docs/SNA2008.pdf has sufficient detail on how values of non-financial assets are to be estimated including intellectual property.

    About point on labour … it’s more a question of dynamics. The system of national accounts itself is designed to help in studying dynamics and in a sense there cannot be pure static concepts. In my view, the designers of national accounts themselves had in mind a good picture of the way the world works and designed it to help studying it.

    It’s true in all subjects as well. The mass of the earth is essentially meaningless but acquires meaning when used to study and compare other things such as other planets. In the same way wealth of a nation is also like that. That it is meaningless in abstract is no reason to reject it outright.

    There are several points about your comparing two countries with different labour costs. You are potentially entering the territory of the “paradox of costs”. I do not see how two different countries (non-interacting for simplicity to rule out trade-related complications) can have the same GDP with different labour power and other things equal. The one with stronger labour will have a higher GDP. Higher GDP means higher production and it will have higher wealth. So you have to take into account the paradox of costs.

  351. Steve Roth writes:

    To defend B.1’s treatment of corporate value, by twisting JKH’s words:

    The nation cares directly about the market value of its equity shares as the best possible estimate of its corporate firms’ value. They let the corporate world worry about the real asset values they are managing.

    I think the essential conundrum we’re addressing: two ways to define and measure wealth/net worth.

    A. The ability to produce stuff in the future.

    B. The cumulative value of measured stuff created (net investment) in the past.

    As SRW says: “I don’t think we do have a generally meaningful way [of defining the value of nonfinancial assets], along several dimensions.”

    This is especially true, IMO, of B. “Investment,” as measured, is profoundly arbitrary — a relic, in a 70% service economy, of a time when we could reasonably count horses, drill presses and factories. The recent GDP revisions, now including (some) R&D in investment, only serve to highlight that method’s deeply problematic aspects. The (not unreasonable) notion of education as investment — because it creates real, long-lived though intangible and unmeasurable goods — even more so.

    Which leaves us with A as a measure and definition of wealth, as (best) estimated by the markets.

    SRW’s point about capital structure is well-taken. Increasing corp leverage may largely explain the increasing divergences that Marko has depicted.* But those divergences have never exceeded 4% and are mostly much lower. Bottom line, B.1 and B.101 give us good looks at what we’re worth, using different methodologies — B.101 via telescoping (and “hiding” households’ “ownership” of government/ROW), B.1 via summation (attempting to reveal the government/ROW portion).

    And to note: both B.1 and B.101 are:

    1. Completely dependent on market value of equity shares to estimate (and for B.101, to telescope) the value of corporate firms.

    2. Completely opaque as to the real-asset “value” and structure of corporate firms hidden behind that market valuation.

    In both cases some addenda below “Corporate equities” could provide a look-through into firms’ real and/or financial capital structures, but as addenda they would be inert relative to other measures within B.1 or B.101.

    * I do wonder whether expanded Fed balance sheets have also contributed, per SRWs point about capital structures affecting B.1 wealth. That $3 trillion comes to 3 or 4 percent of net wealth…

  352. Steve Randy Waldman writes:

    Ramanan & JKH — So, I’m very happy to jointly attribute the point to both of you, re the cap structure sensitivity of the equity measure. I’d say you both made significant contributions.

    Ramanan — For any set of accounting choices, if you are lucky enough to have data you treat consistently, you can call them “the national accounts”. They may have something in mind. They may be useful for some purposes, and if you understand the many, many somewhat arbitrary choices well enough, you may be able to pose certain questions and get not-misleading answers. AFAICT, the SNA guide is just that, a very exhaustive list of principles in the hopes of encouraging national accounting schemes to be comparable in some sense. I still have no idea what the “net worth of a nation” would mean, in the context of all that. We all have certain intuitions about what the net worth of a nation might mean, and we might take accounts drawn up according to SNA principles as “that”, but unless we’ve done what you may have done and thought very carefully about hundreds of pages of choices, we’ll miss certain details. Should R&D “investment” be capitalized into the wealth of the nation? You may know what SNA says on the matter. I have know idea, but that may be an increasingly material question. If we have “SNA-compliant accounts”, can we be confident that their accounting of R&D is consistent? There are hundreds of questions like that we have to answer (perhaps implicitly, unthinkingly) in order to generate our “net wealth”.

    The mass of a rock has a meaning applicable to such a wide variety of questions that it’s useful to simply consider it a characteristic. There are few cases where it makes sense to discuss the mass of the rock according to Method A vs the mass of the rock according to method B, C, D, and E. That’s entirely untrue re net worth of nations, in which there are a combinatorial explosion of reasonable methods.

    And when we’re done, we’re still missing the question of what it means. When we write the balance sheet of a firm that we can model as an infinitesimal bit of an economy, we claim that the asset value we attribute is in some sense a price — a conservative estimate of the replacement value of the assets of the firm, if we use the least-of-cost-or-market convention. “National wealth” admits no such interpretation. I just have no idea what “national wealth” means in isolation.

    Again, that doesn’t mean there is never any context where “national wealth” is coherent. Piketty, for example, likes to measure “capital” (by which he really means wealth) in years of income. That reflects and suggests questions that might be coherent and addressable. Suppose I have access to x% of national income, how long would it take for me to store y% share of pre-existing sources of purchasing power? That’s a coherent question, and we could try to come up with procedures to address it, although of course our procedures would be arbitrary and uncertain to some degree. But if you tell me that “national wealth is $15T” without other context, I’ll claim it’s meaningless. Even cross-national comparisons of national wealth I’ll claim are pretty meaningless. Is the national wealth of the US greater than the national wealth of Japan? Japan has Mount Fuji, which I’ll claim is priceless beyond any resource of the US. Am I wrong? (And where is land in all of these balance sheets? Is it there somewhere or not? How is Area 51 valued in the US?) If you tell me that the US has a greater GDP than Japan, I’ll say that’s close to coherent. We can, as we conventionally do, order the wealth of nations by GDP, on the theory that GDP is relatively stable and wealth is productive capacity. But the GDP ordering will sometimes disagree with claims about “national wealth”. You hear this frequently with respect to Italy, which is said to have a high debt-to-GDP but extremely high wealth because of valuable land and artworks and buildings and stuff. Which ordering is more meaningful? I’ll stick with the GDP version. Maybe you disagree? But I’ll bet that an SNA accounting of Italy’s national wealth would miss much of the value that underlies peoples’ claims of its largesse. I suspect that national accounting standards don’t do a great job of valuing the Sistine Chapel. How should it figure in national wealth?

    I think my little thought experiment about labor power is a bit stronger than you think. There is, in an accounting sense, no contradiction is two countries with identical real assets and identical real products pricing the assets and distributing those products differently. You can make behavioral arguments about what is or is not plausible, but that’s imposing a model, an overlay upon the accounting. I don’t want to defend or dispute the plausibility of my thought experiment as a matter of realism. (Obviously, two identical countries in any important sense are very unlikely.) But at the level of accounting, my scenario is perfectly coherent, and we’d draw very misleading conclusions about comparative capital depth if we interpreted “national capital” accounts naively.

  353. Steve Randy Waldman writes:

    JKH — Right, re the convention of not capitalizing labor power into an asset. My only point is that the DCF/market value of “capital”, whose product we may, um, capitalize, cannot realistically be defined independently of norms of labor compensation. We may agree, as a matter of politically charged convention that labor is not an asset, but norms surrounding compensation undeniably affect the market value of the things we do write down as assets. So, we face a fork in the road re valuing “national wealth”. We can choose to treat the same capital, in a physical productivity sense, as having the same value, regardless of where we find it. Or we can choose to value identical capital differently across borders, depending on the degree to which it is “impaired” by labor bargaining power. There is no right answer to the question of which of these choices is correct. If we are interested in drawing international comparisons of productive capacity, the first choice is more correct. If we are interested in measuring the capacity of wealth owners to mobilize purchasing power, the second choice is more correct.

    Plus, if we are really interested in capturing human intuitions about what “national wealth” should mean, might we not want to break the convention that labor isn’t capitalized? After all, increasingly much of our “investment” is education, and we do a lot of talking about “human capital”.

    The conventions of ordinary balance sheets exist not to characterize the value of an enterprise to all of its stakeholders, but to inform primarily shareholders and creditors. That is why labor income and consumer surplus, for example, are excluded. These are not mysterious or arbitrary choices. Corporate balance sheets are what they are because they inform the particular purposes they inform.

    But when we talk loosely about “national wealth”, we intuitively do mean the wealth of all (citizen? resident?) stakeholders. So corporate balance sheet conventions are probably bad choices. But I suspect even the very elaborate guidance of SNA doesn’t deviate form corporate conventions nearly as much as this intuition might suggest.

  354. Steve Randy Waldman writes:

    SR — This is all your fault, you know. You intuitively want something that cannot be provided. You were mad at the MMT-ers, because they kind of claimed to be providing that but really offered something totally different. 350+ comments later, a zillion smart people have weighed in trying to help build you a perpetual motion machine. But the enterprise was doomed from the start. I think you’ll have to content yourself with smaller questions that have better answers.

    I’m not sure you’ve grasped the depth of the JKH/Ramanan complaint re the B.1 treatment of corporate assets. You can say, as a matter of definition that, “The nation cares directly about the market value of its equity shares as the best possible estimate of its corporate firms’ value. They let the corporate world worry about the real asset values they are managing.” But that’s entirely at odds with any reasonable intuition of “what the nation cares about”. The wealth of corporations might reasonably be defined as their productive capacity, or it might be defined as the value (market or otherwise) of real assets. But it is entirely unreasonable, however common, to define a firm’s market capitalization as its “value”. Again, the easiest way to see this is in terms of stability. We want what we call the “value” of a thing to be reasonably stable, unless some event or circumstance really impairs the thing. But a firm carry $1000 in real assets can, pretty much at will, shift its market cap from $1000 to $300 just by borrowing money and repurchasing shares, without changing any aspect whatsoever of its behavior or production. Again, you, and the Fed, can define value however you like, these things are ultimately arbitrary. But a measure with that characteristic strikes me as a pretty terrible measure. Its only possible redemption would be a mix of practical and empirical claims, that it is impractical to estimate the value of more “correct” operationalizations of value, that aggregate cap structure is stable enough that the time series of value would only be shifted, its shape would not be altered, by using equity values as a proxy. Neither of these arguments strikes me as persuasive. Aggregate cap structures do change, we hear in the business press of periods where corporate leverage is high or low. And if the Fed is comfortable estimating real asset values of every other sector, surely it could do so of the corporate sector. I suppose their defense will be the gain in accuracy from measuring intangibles outweighs the loss of accuracy from sensitivity to cap structure. Maybe, maybe not. These are very arbitrary choices.

    I don’t think you are right about similar vulnerabilities in B.101. B.101 includes the market value of debt securities held by households as well as the market value of equities, so it is not subject to the same critique. If firms shift their capital structure (vis a vis households as claimants), the loss of value-to-households of corporate equity will be made up by the gain in value-to-households as corporate debt. I think the B.101 accounting is more coherent and defensible.

    Do you think you have learned anything from all this? Apparently US wealth has grown by ~40% since 2009. Is there any meaningful sense in which we’ve grown 40% richer (other than the sense that if we arbitrarily call B.1-wealth wealth, we are 40% richer)?

    I don’t think the claim that “the US” is 40% richer than it was in 2009 is meaningful at all. I cannot thing of any context whatsoever in which that claim maps usefully to any real world thing. Stock and housing values have recovered, but there is little sense in which that makes “the US” wealthier, that is largely a zero-sum distributional phenomenon, people who hold stocks and homes have seen a growth in purchasing power relative to people who don’t. Real GDP is up about 12% over a similar period, real GDP per capita up by about 7%. It’s hard to know what these measures mean either, but I don’t think it’s ridiculous to claim that we are collectively 12% richer than in 2009, or that each of us is on average 7% richer. (Much of that is skewed to the wealthy, so we might not all feel it, but it isn’t zero sum like pure price appreciation, these should be real gains someone is enjoying not purchasing power at someone else’s expense.) The B.1 net wealth figures (and probably the B.101 household wealth figures as well) just don’t map to what we mean by “aggregate wealth”, because the part of their changes that are just repricings don’t aggregate into what we think of as wealth. These measures are corrupted but what in fact are distributional changes.

  355. Ramanan writes:

    SRW,

    I don’t see how one can I say “I do not know what national wealth means” and at the same time say “I understand households’ wealth”. Imagine two nations with the same household wealth but different government wealths. Isn’t one richer than the other? Or you could say that the other one is more fortunate. Irrespective of how you look at it (and in combination of other tables), there are differences.

    I could argue for example how Greece’s households were getting richer and richer but the national wealth wasn’t showing the same behaviour.

    And I could argue that Germany as a nation is wealthy but its households are not that wealthy as can be possible.

    Basically households are just one part of a nation. Important of course but a nation is greater than the sum of its parts because households get together and form firms, institutions etc.

    About SNA, I do not know how to write it simply. Or maybe I should blog it. On my reading it, it looked fairly consistent though. Yes there are hundreds of questions but I believe that it has reasonably addressed it. Research and development is added in fixed capital formation (“investment”) and also this adds to the stock of non-financial assets.

    There are rules about what is counted in non=financial assets and what is not. So forests etc are not counted because it doesn’t fall in the production boundary. But if someone tomorrow knows how to exploit it, there are rules to include it in the national accounts.

    All your questions are actually sincerely answered in the SNA. I do not know how to answer them simply.

    “You hear this frequently with respect to Italy, which is said to have a high debt-to-GDP but extremely high wealth because of valuable land and artworks and buildings and stuff. Which ordering is more meaningful? I’ll stick with the GDP version. Maybe you disagree? But I’ll bet that an SNA accounting of Italy’s national wealth would miss much of the value that underlies peoples’ claims of its largesse. I suspect that national accounting standards don’t do a great job of valuing the Sistine Chapel. How should it figure in national wealth?”

    But national accounts do not also include things such as unpaid work in GDP. That critique is equally applicable to GDP as it is to national wealth. Why single out national wealth? Churches are part of NPISHs and they contribute to production and their assets are valued.

    Also, Italy’s high public debt is not a problem. That’s because it cancels out in the calculation of national wealth and Italy’s net international investment position is not that bad.

    The SNA is a reasonably well thought out system. B.1’s slipshod ways is not a proxy for it. It’s surprising – publishing it seems to have done hurriedly unlike other things in the Z.1 accounts which is superb.

    I am interested in your example of comparing two countries with some numbers thrown in. As in, what are these numbers? Who holds what and what is it? A machine, software, or what?

    In the meantime, I gave a specific example in my conversation with JKH.

    1. A nation with 40% public debt all held by households
    2. A nation with 140% public debt, 40% held by households and 100% held by foreigners who accumulated this via international trade.

    Almost identical worlds. But obviously, these are different as the nation as a whole is indebted in the second case, even though households are similar.

  356. JKH writes:

    I’d like to elaborate on my intended meaning behind “flexibility” as I used it in my comment to SRW.

    Suppose you are comparing two “systems” of accounting as a starting point in seeking some sort of concept of national wealth, in some broad way – those two systems being the financial capital ex real asset view captured in the core household centric B.101 view, and the IMA/SNA real asset ex financial capital view captured in some manner of core IMA/SNA consolidation. By “flexibility” I mean you have a CHOICE of where to start. To make the point, I would rather start with B.101 and use the government, corporate, and international balance sheets that appear elsewhere as supplementary data, than I would start with SNA and use financial capital data such as appears in the B.101 household view (or something similar) as supplementary data. That’s my own preference according to what I think the priorities should be in the 40,000 foot view. I’ve stated my reasons for that several times.

    Ramanan, I’ve said that I think your government/international example is in effect a red herring – in the sense that finding out that information is simply not a problem from any starting perspective. If I want to know the details of the government balance sheet or the details of the international balance sheet, I go find them in the balance sheet tables. The fact that those gross positions don’t appear in B.101 doesn’t delegitimize B.101 as a view of comprehensive consolidated wealth from the household perspective. It includes and is only intended to include and only can include as an interface item the net financial effects of government debt and international debt – as those positions directly affect the financial wealth of the household sector on a net financial basis. It reflects that net financial effect accurately. B.101 obviously doesn’t include gross balance sheet information any more than it includes gross balance sheet information for the corporate sector. But I’ve said this a number of times now … the starting point of netting on consolidation is not necessarily the end point of investigation into gross balance sheet profiles. One can easily go look for this detail. And when it comes to the government balance sheet for example, one can then consider such stuff as MMT net financial asset views or Ricardian equivalence stuff or go ponder what the meaning of those government “real assets” really is in the context of a common sense liquidity perspective and contingent cash value fiscal scenarios. Their aggregate importance would greatly diminish as a result – I would guess. But I would much rather have this natural flexibility of investigation built-in to the starting perspective than begin with a rigid bean counting approach to the summation of real asset values by artificially netting out the entire financial intermediation fabric of the economy.

    As a contrary extreme, I’d be dumbstruck to view a consolidated real asset report and just declare that to be the endpoint of my investigation, totally ignoring the effects of wealth in the form of financial capital especially as that affects the ultimate household perspective on wealth. Just consider the Apple example. There’s virtually no connection between the magnitude of Apple’s $ 25 billion real assets and the $ 700 billion in current equity value its financial investors see on their computer screens (with a dollop of liabilities value even added to that). Which of these should one really think is more important as an element in getting a sense of aggregate national wealth in a broad sense? (ex the obvious adjustment already included when B.101 nets out foreign ownership of Apple stock) I think the answer is pretty clear.

  357. Ramanan writes:

    JKH,

    I am not debating about placement of the table – whether in front or some place else.

    The point I am debating is the importance of it. Not even saying it’s the end point. But just a table showing the wealth of a nation.

    The starting point for me is exports but it’s not that I wish to place it on Page 1.

    A bit like Don Bradman batting at No. 3 instead of opening the innings for the Australian cricket team.

    Given a choice, of course I’ll start with GDP, then balance sheets and then flow of funds.

    So it’s not like I am suggesting that remove B.101. In that case (if removed), Apple’s equities held by investors don’t make an appearance anywhere. But that’s only in the case when there’s no B.101.

    I am only making the case for statistics for the wealth of a nation for each quarter.

  358. Ramanan writes:

    There is of course the question of netting, although that is different from the netting here about national wealth because I don’t think anyone has any model which has the national wealth in it.

    One can of course make stock-flow consistent models but these run into so many equations and makes it difficult to do modelling with real data. So because of that something simpler is needed and that’s where the connections between stock of financial assets and flows of income and expenditure comes in, with high level of aggregation used.

    There are good things one can see using this. For example if the private sector expenditure relative to income is high and the private sector is in deficit, it will revert its position because its financial assets relative to total wealth is deteriorating and this adjustment will slow down the economy or lead to a recession.

  359. Steve Roth writes:

    @SRW: “This is all your fault, you know.”

    Oh I know, and I frequently cringe at the thought. Mea culpa to any who feel dragged into this.

    I will say, though: those “zillion smart people” have apparently considered this conversation worth having. Very few of the comments here respond to my comments. They’re among others. (And the last time I incited thus, it did give us S=I+(S-I)… )

    When I asked, in #290, “is it an impossible exercise that should be excised from the Z.1?”, it was not a rhetorical question. I was hoping for responses. I’m aware that the best answer might be “yes.”

    So, and, I’m apparently not the only one that feels a need to think about these conundra. Kuznets worried at it in Capital in the American Economy: “…[fixed] capital formation…represents the real savings of the nation.” Piketty does so in his Capital, defining capital and wealth as synonymous (though not consistently).

    Your point about the wealth change 2009–2015 is very well taken. Agreed. My inclination is always to think in the longest time series I can find, though, and over many decades I don’t think it’s crazy to say that market valuations could be the best possible estimate of wealth, of what all our stuff is “worth” — if it’s even worth estimating. Piketty and company, and the national accountants at the Fed, think it is, however imperfectly, and I agree.

    BTW, the most resounding support for a “Yes” answer to my is-it-an-impossible-exercise question arises in a Josh Mason piece I just revisited yesterday, on Germany’s remarkably low measured HH net worth (especially median, but also mean). Greatly simplified, it seems to be basically be a function of how future pension and other social payments are accounted for and capitalized — or not.

    http://slackwire.blogspot.com/2014/04/wealth-distribution-and-puzzle-of.html

  360. JKH writes:

    SR,

    I answered your # 290 question 3 comments later:

    “If I were doing b), I would use the real asset values that I identified in my comment # 256. The value of non-financial corporate real assets is $ 20 trillion from S.5.a. The value of financial corporate real assets is $ 2 trillion from S.6.a. From what I’ve seen in supporting explanatory text, I think their reasons for using equity claim values instead are very flimsy indeed, and result in the conceptual nonsense I’ve described.”

    B.1 is a travesty as is.

    In danger of becoming a long national nightmare – a cancer on the Z.1.

    :) tricky B.1

  361. Steve Roth writes:

    @JKH: “I answered your # 290 question 3 comments later”

    Yes sorry I didn’t respond to that, great answer. I don’t fully agree, but I think I fully understand why you say so.

    >In danger of becoming a long national nightmare – a cancer on the Z.1. :) tricky B.1

    Yes, ;-)

  362. Steve Roth writes:

    In case I haven’t said it recently, the quality of the discussion here is simply stellar. I want to make every macroeconomist read every word of it (except my parts; [blush]).

  363. Steve Randy Waldman writes:

    SR — I hope you know I am teasing you, not scolding you, when I say this is all your fault.

    SR & Ramanan — To be clear, I find household net worth, in isolation, to be precisely as meaningless as “aggregate net worth”. But note the “in isolation” qualifier. I don’t find table B.101 useless, I just find the number on Line 41 of that table to be meaningless except in reference to some other numbers. I think it’s a bad idea to use the “net worth”, either of households or of nations, under SNA or any other accounts, as an unproblematic measure of community wealth, with all the intuitions that might carry along.

    I have, over the years, made actual use of table B.101, primarily to inquire into questions of the household portfolio balance. What share of household financial asset portfolios is held as corporate equity vs fixed income instruments, for example? I’ve sometimes been frustrated, in a practical sense, trying to answer these questions, because it’s hard to look through categories like Mutual fund shares, Life insurance reserves, and Pension entitlements to fixed income vs equity exposures at the household level. (But B.101.e lets you look through those categories. Has that always been there, if so I’m embarrassed for my past obliviousness!)

    Putting aside my measurement quibble with B.1, the ratio of B.101/B.1 net worths might be useful, as a measure of how much “national wealth” accrues to domestic households vs (necessarily) the foreign sector. (Interestingly, just checking, current HH net worth is higher than US net worth, which implies that the US must have a positive NIIP, which is wrong. So maybe not so useful. [Update: probably the govt sector’s negative foreign financial position accounts for this.])

    But in isolation, telling me that “US net wealth” is $79.7T or that “US household net worth” is “$85.7T” are totally meaningless numbers to me. I know how to explain, at least conceptually, at least approximately, what the net worth of a firm is supposed to mean, how it is in some sense a rough approximation of a price. I really have no idea what the dollar value net worth of all households or of “the nation” are supposed to signify. I have some idea of how to interpret the ratio of corporate equities to household net worth means measured in consistent accounting exercise, so that bottom line number is not useless to me. But in isolation, I just have no idea what to make of it.

    You might reasonably complain that the same thing is true about GDP. But, at least conceptually, GDP is not an accounting construct. Its measures, obviously, are. GDP is the dollars actual spent on newly produced goods and services (net of the dollars spent on previously produced inputs consumed in the production of those goods and services). If we had perfect surveillance, we could measure Nominal GDP directly.

    But that’s not really why I think GDP, real or nominal, is a better measure than these aggregate net worth figures. Really, I think GDP is useful because it has proved useful. I think that fluctuations in GDP correlate usefully with other economic phenomena so that it is useful to track it as a measure of macroeconomic health. Fluctuations in versus the nominal GDP track well with my perceptions of periods of boom or bust, or at least relative prosperity vs economic weakness. The magnitudes of conventional real GDP (despite the fact that the “real” in “real GDP” is an incredibly arbitrary accounting exercise) track reasonably well with my intuitions of aggregate prosperity in an absolute sense.

    If you were to persuade me that, in time series, Line 41 of B.101 or Line 1 of B.1 usefully track some “observed” (in a vague, informal sense) macroeconomic reality, I’d concede the usefulness of the measure. Whether I would concede it as a measure of what we intuitively thing of as “aggregate wealth” would depend upon what it seems to track. But I might, if what it tracked and our intuitions regarding “aggregate wealth” seemed to coalign. Ultimately accounting exercises are instrumental. Their merit is in the usefulness of their outputs, rather than anything intrinsic to the procedures used to generate those outputs.

    But I’ve never seen any measure of aggregate net wealth computed in analogy to corporate balance sheets that seems useful to me, in isolation, as a measure of anything. I’d get over the philosophical point (“what does that even mean?“) if there were a practical case, but I’ve not seen a practical case. As we’ve seen here, the FoF measures capture what are really distributional changes as changes of aggregate wealth. I don’t think it’s at all meaningful to say that US households are wealthier because the value of corporate equity held by 1% of them have risen in dollar terms, unless those rises in dollar terms reflect changes in the real productivity of the firms. It’s silly, for example, to think that the Fed’s ZIRP policy has made us “richer” in lockstep with rising asset values. (It may have made us richer, because it is policy that prevents depression, or poorer because it’s provoking financial instability from which chaos will ensue, but those indirect real effects of ZIRP aren’t well modeled by asset price rises that result mechanically from a falling discount rate.)

    The JW Mason example you point to, with respect to differences between German and US households, is right on point. People in the real world make idiotic and misleading claims about comparative wealth because the form in which we hold that wealth collectively is different in a way that makes German households look poorer than US household, when they are arguably as well off or better off than US households. If we wanted to do a particular exercise comparing US and German housing wealth, we absolutely could. We would adjust US and German housing into some sort of comparable basis, argue for the validity of our choices, and make the comparison. But taking standard sets of accounts and expecting them to yield meaningful answers without careful thinking and adjustment is doomed to fail. It doesn’t work between corporations.The first thing any financial analysis worth her salt does in comparing several firms is put their accounts on a comparable basis. There is no reason to expect it would work with much much more complicated nations.

    There’s a point I hinted at above that I really want to emphasize now. The form of the corporate balance sheet was not (despite some structural similarities) handed down from God as tablets on the slopes of Mount Sinai. Corporate balance sheets are specifically designed to measure not the comprehensive value of a firm, but the value and risk of insolvency of the firm to creditors and shareholders. Even for a firm, if we wished to measure its value in ways that comport with our intuitions of aggregate wealth, we’d need to include its value to customers and employees as well. That, after all, is why we want a “business friendly” regulatory environment, for example. We don’t, in mainstream political discourse, claim to like that because we want to see a relatively few number of businessmen and investors become rich. We implicitly understand that the social value of productive, nonpredatory businesses goes way beyond its value to creditors and shareholders.

    Measures of financial vs comprehensive social value may even be inversely correlated. For example, in a B.1 treatment, an increase in the monopoly power of firms will flow directly through to corporate equity. (Even if they fixed that cap-structure-variance problem, an increase in the monopoly power of firms would flow through to mkt-priced enterprise value.) The social cost of rent-seeking is omitted form any corporate-balance-sheet inspired presentation of wealth, while the financial value of rent streams is captured.

    If you really do want to try to invent a concise measure of the “wealth of nations”, a balance-sheet like presentation, however many pages of careful and consistent guidance you might find for its productoin in the SNA, may be entirely the wrong place to start.

  364. Ramanan writes:

    SRW,

    There’s no invention. It is simply a result of SNA. You may not like it, but so did Einstein not like a dynamic universe which was the conclusion of General Relativity.

    Household net worth like $80 trillion don’t seem to mean much but divide it by the number of households or population to get more intuition.

    That Germany versus Greece is quite illustrative in a different sense. Germany is such a rich nation but its households aren’t that rich. Greece is the other way round. Germany is bailing out Greece. And also illustrative of inequality in the US. It doesn’t mean that net worth is meaningless. It illustrates.

  365. JKH writes:

    SRW # 363,

    “Interestingly, just checking, current HH net worth is higher than US net worth, which implies that the US must have a positive NIIP, which is wrong. So maybe not so useful. [Update: probably the govt sector’s negative foreign financial position accounts for this.”

    As you suggest in your update, the reason that household net worth exceeds the larger aggregate is that household net worth includes the positive net worth interface effect of measuring the government sector according to the (“MMT NFA) value of the direct and indirect claims on government debt as a financial asset held, whereas the aggregation in terms of real (SNA/IMA) wealth displaces that measure and replaces it in effect with the value of government real assets.

    Actually, the conversion from the first to the second is: ((real assets – debt) – (debt)) = real assets instead of debt. Either methodology results in the effective inclusion according to that methodology of the government position within the measure of household net worth itself. The noted value differential between household net worth and the larger aggregate has everything to do with this different methodological treatment of the government interface, and nothing to do with NIIP, because the NIIP interface is included in the same way in the calculation of both household net worth and the larger aggregate, so there is no methodological difference. I noted these comparative points earlier regarding these interfaces.

    “But B.101.e lets you look through those categories. Has that always been there, if so I’m embarrassed for my past obliviousness?”

    Yes it’s been there, and I would guess I’ve been referencing it reliably and essentially for several decades now at least (memory is not certain, but I’m pretty sure of that.)

    B.101.e is quite key on seeing just how brilliant B.101 is, in my personal view. B.101.e when considered with some concentrated focus allows you to visualize the deeper guts of what B.101 portrays – what I have attempted to describe in using the word “telescoping” going back some years now in these blogosphere discussions. B.101.e lets you visualize the financial intermediation wiring of how the household sector penetrates the business sector in the connections provided by “claims on claims”. It is the network of business financial intermediation that feeds into the net worth of the household sector – and this is why I said earlier somewhere that the household sector perspective is not just about the household sector alone. It is a snapshot of everything – from the household sector perspective.

    The household sector includes both the B.101.e financial intermediation penetration into the business sector and the netting interface of the government and NIIP. The NIIP interface in particular is the one where I have made the point in conversation with Ramanan a number of times that the netting already exists in the household position, but if interested in the gross balance sheet detail, you can go get it another adjacent table. That supplementary work is no knock against the reality of the netting construction that is a feature of B.101.

    Perhaps these points I have noted above are partly why I was a bit surprised by this:

    “I know how to explain, at least conceptually, at least approximately, what the net worth of a firm is supposed to mean, how it is in some sense a rough approximation of a price. I really have no idea what the dollar value net worth of all households…”

    Because there really is no conceptual difference in the methodology required to read the balance sheets of these respective positions – firm or household net worth. The household net worth balance sheet is just a little more complicated because of the arguably more complex nature of the inclusion of its interfaces with each of the business, government, and foreign sectors.

    In that context, everything is there in the household view. Nothing is omitted when the interface methodology is understood. And because the interface methodology with the government sector can vary between government debt as a direct or indirect household financial asset (MMT NFA as I have described) or directly consolidated instead with additive real assets of the government – as is done in the SNA/IMA version where the debt asset is replaced in effect by the real asset – depends on whether or not one uses that real asset adjustment in the aggregation.

    “You might reasonably complain that the same thing is true about GDP. But, at least conceptually, GDP is not an accounting construct.”

    Half agree. In my view, the sanctity of the GDP definition, up to certain reasonable conceptual limits, is greater than that of aggregate wealth, in the sense for example that I believe one should be able to choose between the household sector view and the aggregate real asset view according to personal preference. My personal preference is that the household view is superior – for a number of reasons that I’ve tried to suggest in comments. But GDP is still an accounting construction, no different in that sense than one’s chosen measure for wealth aggregation. Accounting is a form of measurement and mathematics, applied to real and financial assets. It is ubiquitously necessary. There is far too much accounting-phobia in the strangely elevated land of professional economic thinking in my view.

    “Fluctuations in versus the nominal GDP track well with my perceptions of periods of boom or bust, or at least relative prosperity vs economic weakness.”

    But please witness the long term, violent pulsation of B.101 as it appeared through the financial crisis. This is volatility in capitalism. GDP and B.101 are measurement partners in that sense.

    “But I’ve never seen any measure of aggregate net wealth computed in analogy to corporate balance sheets that seems useful to me, in isolation, as a measure of anything.”

    Disagree, for reasons I’ve start to note above.

    That’s enough from me, for the moment, except that I do wish to say that I believe your concerns about distributional matters while important are quite orthogonal to concerns about general measurement matters.

    I haven’t looked at that J.W. Mason piece yet. I plan to do so.

  366. Ramanan writes:

    “(Interestingly, just checking, current HH net worth is higher than US net worth, which implies that the US must have a positive NIIP, which is wrong. So maybe not so useful. [Update: probably the govt sector’s negative foreign financial position accounts for this.])”

    SRW,

    It’s not wrong. Household net worth higher than the US net worth suggests NIIP is negative and that’s indeed the case – US NIIP is negative.

  367. JKH writes:

    Ramanan # 366 + SRW # 363

    I completely misread SRW’s update there – moving to my explanation in haste as soon as I saw the word “government”.

    My explanation is correct though.

    The difference has nothing to do with NIIP, or the government’s “negative foreign financial position”. It has to do with the difference between financial asset consolidation and real asset consolidation in the treatment of the government position.

    And notwithstanding the apparent bilateral optics between the government and the foreign sector, it has nothing to do with which sector owns the government bonds. Those bonds in theory can easily be swapped for domestically issued corporates, and nothing would change regarding the wealth difference under question.

  368. JKH writes:

    The Two Steves:

    Regarding the Mason piece:

    “In short: The biggest reason that German household wealth is lower than than elsewhere is that less claims on the future output of the housing sector take the form of assets. Housing is just as commodified in Germany as elsewhere (I don’t think public housing is unusually important there). But it is less capitalized.”

    I can’t make sense of this. If he’s talking about wealth distribution, then lower ownership at lower income levels explains it. I think that’s pretty simple in terms of depressing wealth medians. But if he’s talking about capitalization per se, those houses being rented out are owned by somebody and therefore capitalized at some level. This is completely different than the pay as you go/capitalized pension difference. So I don’t understand what he’s saying here.

    Whereas, this is more interesting:

    “the ”control right” or ”stakeholder” view of the firm can in principle explain why the market value of corporations is particularly low in Germany (where worker representatives have voting rights in corporate boards without any equity stake in the company). According to this “stakeholder” view of the firm, the market value of corporations can be interpreted as the value for the owner, while the book value can be interpreted as the value for all stakeholders.”

    So he’s saying the stock market is relatively depressed in Germany, other things equal. It’s a matter of figuring out how institutional arrangements drive market capitalization. So after that, the fact remains I guess that although income levels are relatively elevated, stock market wealth is relatively suppressed.

    Fine. It took some accounting to figure that out though, didn’t it? I see no general indictment of wealth measures in that. It’s a direct matter of explaining the income/wealth configuration, which he’s done. Impressive post (although somewhat sprawling). But his conclusion totally depends on the availabilities of income measurement and equity market capitalization measurements.

    I read it quickly. Have I misinterpreted in any way?

  369. JKH writes:

    Regarding the NIIP thing once again:

    Let me offer a highly simplified example.

    Case 1 net asset:

    NIIP consists only (gross and net) of the ownership of BMW stock by some unit in the U.S.
    If that unit is a household, that wealth shows up directly in B.101.
    If that unit is a business, that wealth shows up indirectly as a financial claim held by households on the U.S. business that owns the stock.

    Case 2 net liability:

    NIIP consists only (gross and net) of the ownership of Apple stock by some foreign domiciled unit.
    Then that stock is not available for direct ownership by households in B.101.
    And neither is that stock available for direct ownership by a U.S. business, which means it isn’t available for indirect ownership by households in B.101.

    All NIIP positions consist of compilations of such asset and liability positions.
    So B.101 takes full account of the netting effect of NIIP on wealth.

    And because the same type of NIIP financial netting still exists (by deliberate construction) in the consolidated real asset methodology – as an adjustment to real asset calculation results otherwise – there is a similar addition or subtraction effect in what ends up being available as wealth domestically – essentially the same as is the case for B.101.

    So NIIP is a non-issue in terms of the net effect being fully reflected in either methodology. It is still very interesting though to explore gross positions by looking at the NIIP balance sheet in its gross detail.

  370. Ramanan writes:

    JKH,

    Hmm. The comment and SRW’s update is his thought process. Whatever the case maybe, doesn’t disprove any inconsistency of a national net worth.

  371. Ramanan writes:

    JKH,

    I am again not sure of your point in 369. Simply by looking at 369 doesn’t give an idea of the US net international position.

    So we have

    Private sector net financial assets = Net government liabilities + NIIP.

    Household is a part of the private sector. Just looking at the household balance sheet no way indicates the various combinations possible on the right hand side.

    So not clear by “So NIIP is a non-issue in terms of the net effect being fully reflected”

    Anyway. That’s deadlock. Maybe should discuss some other time.

  372. Ramanan writes:

    Oops.

    “Simply by looking at 369″ should be “Simply by looking at B.101″

  373. JKH writes:

    Ramanan,

    “Simply by looking at B.101 doesn’t give an idea of the US net international position.”

    I’m not sure how to say it differently than what I have done so far several times.

    So try, try again:

    B.101 reflects a net financial effect from NIIP. That is my direct claim about NIIP.

    It reflects that effect in combination with additional net financial effects from the government sector and business sector.

    That means that the net financial effect is included – it doesn’t mean one sees the gross or net NIIP balance sheet directly in size or texture. You look at the NIIP balance sheet to get that information. I’ve qualified that several times.

    This additional information drill-down is also what is needed to understand how the government balance sheet has a further incremental effect on the household balance sheet in addition to or in combination with NIIP. And both of these outside effects can be evaluated further for their directness or indirectness on B.101 by studying the profile of B.101.e, which helps us understand the way in which business claims on financial assets are passed through to household claims on business.

    The point is that the consolidated effect of all these positions orbiting the household balance sheet (in a household centric view) is in fact necessarily captured in the household balance sheet. The household balance sheet is not an island in this regard. Everything affects it. That’s the reason why it’s a legitimate point of overall consolidation, provided one understands the netting methodology against the various sectors that it interfaces with directly or indirectly.

    This is really just sector balances in a 4 sector model, where there is an explanation of how all of the government, NIIP, and business positions taken together affect the B.101 position.

  374. Ramanan writes:

    JKH,

    Of course households’ balance sheet is affected by how they have transacted with other sectors and the assets they have accumulated, liabilities incurred in the past and how they have revalued. But I feel a bit lost here about what is being debated and so on. I guess the Fed’s Z.1 report starts out with a table on the national wealth and moreover has conceptual errors in it. But that’s the Fed’s publication. I’d generally start just like the system of national accounts: production, generation of income, distribution of income, use of income, capital account, financial account, other changes (such as revaluation) and balance sheets and then flow of funds. What I did not understand is why you saw the example I gave as a red-herring. It’s also consistent with the fact that household sector is not an island.

  375. JKH writes:

    Ramanan,

    Maybe think of it this way:

    2 approaches:

    SNA – which is more or less top down, providing consistent balance sheet information across the economy (I think) – your preference, understandably. It is top down by construction, which tends to be a clean way of looking at things.

    B.101 – which is more or less bottom up, providing a view of the wealth being generated from the household sector looking out – my preference.

    I prefer B.101 because it shows both the real and financial portfolios of the household sector. It does not give up anything in terms of comprehensiveness, because all net wealth gets channeled from other sectors via financial intermediation. It does put real assets held by the non-household sectors in the background, but that is inherent in the wealth aggregation methodology for the core report. That real asset information along with additional financial intermediation detail is available however in supplementary reports, particularly now with the IMA as part of the Z1. But it is the initial aggregation approach of B.101 that I prefer as a starting point.

    In both cases, if you want detail on specific sector balance sheets, you can find them. In the case of B.101, you go to adjacent balance sheet tables in Z1. And I think the same holds for SNA although I don’t use it. I would think you look at more than one table to get detail.

    I mean red herring in the sense that NIIP detail that is obviously not contained in B.101 can be found by looking at the NIIP gross balance sheet somewhere in an adjacent table in Z.1. That gross balance sheet detail obviously doesn’t have to appear on B.101 itself in order for the net position to be reflected in B.101 (the point I have made repeatedly about net impact).

    I’m not sure how you use SNA yourself, but I find it hard to believe that you can find NIIP gross position detail in a SINGLE table which at the same time gives you a comparable level of detail for every sector of economy. That strains credulity – along with eyesight, I should imagine. So I see little difference about accessing a comparable level of detail in choosing between two systems that can be roughly characterized as top-down or bottom-up.

    But the main reason I prefer B.101 is that it doesn’t obscure financial intermediation in the depiction of wealth, and that this is the view of the world considered rationally (in my opinion) by the ultimate consumer in the economic system. I think this vital and natural information to be interested in.

    B.1 is a disaster. I wouldn’t even associate it with what I would characterize as the B.101 approach. In terms of supplementary information that might be useful in adding to the initial B.101 perspective, B.1 is literally the last report in the entire Z.1 that I would go to for anything useful in that context.

  376. JKH,

    I’m sorry, but this still makes no sense to me. Is there anything you have said here about B.101 that cannot be said of S.3.a?

    I find it hard to believe that you can find NIIP gross position detail in a SINGLE table which at the same time gives you a comparable level of detail for every sector of economy.

    I do not know what, exactly, that you are requesting– that does not appear in S.9.a.

  377. […] remarkable discussion on “national wealth” and the national accounts that is running over at Interfluidity (373 comments and counting…) in response to my last post prompts me to recount an anecdote […]

  378. Steve Roth writes:

    Quite a few things to say re: SRW’s, if I can find the time, but had to write this up that I’ve been meaning to for a while, echoing some of what he said:

    http://www.asymptosis.com/?p=8413

  379. JKH writes:

    David,

    Perhaps you can convince me to use S.3.a instead of B.101.

    Why?

    What advantages will I experience in doing that?

    I’m all ears and open mind.

  380. JKH writes:

    David,

    Regarding that quote – I’m not requesting anything.

    It’s a contention that Ramanan can refute if he likes.

  381. Ramanan writes:

    JKH,

    Yeah B.1 is a disaster.

    However, if you check the annexe of the SNA guide, there’s a table with all sectors of the economy and their assets and liabilities, though in less details. So the columns are sectors and then the totals and rows are assets and liabilitis. Then one can go to flow of funds to view more details. That’s my starting point in looking at balance sheets.

    When you say view of the world considered rationally by the ultimate consumer, what do you mean exactly?

  382. JKH,

    You have it backwards. You are talking up B.101 as if it had something that S.3.a does not. I think they are both fine if you only care about household balance sheets.

    I guess I was unclear about S.9.a, though. I do not understand what your contention is, if not immediately and obviously refuted by the existence of S.9.a. You seem to suggest something cannot be done, and yet there it is. So either I misunderstand your contention, or your contention is wrong. What am I missing?

  383. JKH writes:

    Ramanan,

    I mean consumer demand drives the economy (with investment as a factor in future consumer goods production)

    That demand comes from the household sector.

    And households are more interested in the value of the Apple stock they own or that their mutual funds own than they are in the value of the real assets Apple owns.

    etc.

  384. JKH writes:

    “You are talking up B.101 as if it had something that S.3.a does not.”

    Bullshit.

    You’re the one that raised the S.3.a. issue. Not me.

    Keep it straight and stop trolling.

  385. Something else I do not get is why B.1 is a “disaster.” Is there really anything beyond the addition of corporate intangibles (that is, the subtraction of IMA corporate net worth) that is at issue?

  386. JKH,

    I am sorry if you have misinterpreted something I wrote, so let me be clear. If you are only interested in household balance sheets, then B.101 and S.3.a are to my eye equivalent. If you are interested in integrating the household balance with household flows and other sectors of the economy, then it makes sense to start with S.3.a.

    The main “issue” I brought up regarding S.3.a was that the net worth was exactly the same as in B.101. So If you have a problem with anything I wrote, please cite specifics so that I might clarify.

  387. Steve Randy Waldman writes:

    Ramanan —

    There’s no invention. It is simply a result of SNA. You may not like it, but so did Einstein not like a dynamic universe which was the conclusion of General Relativity.

    Well this is the heart of our dispute. I don’t dislike SNA. As fas as I know, it’s as fine an attempt as there is to do what it tries to do. But it is entirely an invention. The choices that it adopts may be defensible and they may be consistent, but they are not facts of nature, and accounts that made many alternative choices consistently would be just as defensible. National accounts do not capture or approximate nature, like general relativity or quantum mechanics. Each exercise in national (or any other kind) of accounts represents one of many lenses we might employ to understand… well, what precisely? The lens we choose ought to be motivated by whatever it is we mean to understand.

    Household net worth like $80 trillion don’t seem to mean much but divide it by the number of households or population to get more intuition.

    I think meaningless household net worth divided by the number of households may be just as meaningless. The issue isn’t whether the magnitudes are intuitive. It is a question of what meaning the resulting number has, how we should interpret it. US household net worth, for example, rises with stock appreciation that results not from expected growth in real production but from reductions of the discount rate, perhaps as a policy choice. Calling that change an increase in “household wealth” strikes me as an invitation to bad intuitions. If anything, in my view, “unbacked” stock appreciation makes us poorer in real terms because it heightens social cleavages in ways that are likely to confuse and impair economic activity going forward. But if you think of it as “household wealth” you’ll totally miss that and reason poorly about “wealth” changes.

    That Germany versus Greece is quite illustrative in a different sense. Germany is such a rich nation but its households aren’t that rich. Greece is the other way round. Germany is bailing out Greece. And also illustrative of inequality in the US. It doesn’t mean that net worth is meaningless. It illustrates.

    So, we’ll disagree pretty sharply about this. The claim “Germany is such a rich nation but its households aren’t that rich. Greece is the other way round.” strikes me as a meaningless statement, unless we tautologically define “rich households” as whatever the accounting exercises tell us. Is it true, in any practical sense, that Greek households can stabilize their consumption expenditures through hard times better than German households can because they have high asset wealth? Can Greek households, in practice, whether voluntarily or under coercion by a court, liquidate this alleged wealth to satisfy their debts and still have net worth to spare? I think these questions answer themselves. You can chalk this up to distributional claims: It’s just the Greek HH net wealth is held almost entirely by Greek oligarchs, so typical Greek households aren’t wealthy but the average numbers are redeemable. But you can also point out that these estimates of net worth are hostages of estimates of value that may not reflect what we mean by “wealth”. For all of SNA’s attention to comparability, German households may have unmeasured assets in their ability to draw on different kinds of rights re the rental contracts into which they enter, in the very different labor contracts that determine their wage perpetuities, in the different degrees of insurance provided by their respective welfare states. (These were the sort of points I recall JW Mason made.) That might be the answer to why Greek households don’t “seem” by ordinary understandings of what asset wealth means in practice to be wealthier than German households. To see these kinds of things, you have to look for them, but if you include, say, the a capitalized measure of labor income discounted by relative uncertainty under different legal regimes, you’ll lose the resemblance to corporate accounts that gives these exercises much of their face legitimacy. I don’t say SNA-style accounts lack legitimacy. For some purposes they may be great. But I don’t think they are particularly useful as estimators of “net wealth”, over all sectors or households.

  388. Ramanan writes:

    SRW,

    It’s not an invention as it is not forced upon. The national net worth is simply the sum of net worths of all sectors.

    In all cases – if you are interested in the measure or if you think there is another measure or if you think no such thing can exist, you have to make the case that aggregation by additivity is not a valid thing to do.

    In that sense it is a consequence of the SNA. It is not forced.

    “US household net worth, for example, rises with stock appreciation that results not from expected growth in real production but from reductions of the discount rate, perhaps as a policy choice. Calling that change an increase in “household wealth” strikes me as an invitation to bad intuitions. If anything, in my view, “unbacked” stock appreciation makes us poorer in real terms because it heightens social cleavages in ways that are likely to confuse and impair economic activity going forward. But if you think of it as “household wealth” you’ll totally miss that and reason poorly about “wealth” changes.”

    This is behavioural stuff. Households are likely to increase their consumption if they feel wealthy. This is likely to increase production. So a boom can cause more production. I am obviously not saying this always happens but you have to rule out such things. I don’t know what unbacked stock appreciation is. To me that seems like production is given by a production function or something.

    The Greece/Germany story became popular in media when the ECB report came out. It became popular because it looked like Germany’s households are that rich and yet Germany is being asked to bail out Greece. The answer obviously is that Germany is a rich nation, a huge creditor and household wealth is not a correct proxy for it.

  389. Marko writes:

    Many of these arguments about ” what we mean by wealth” could apply to gdp or gdi , i.e., ” what do we mean by income” , “what do we mean by output”. Just the distributional differences within and across countries make the aggregate measures less than fully satisfying.

    Still , we have a lengthy historical database of gdp , and it does seem useful for many purposes. I don’t think we should stop measuring “national gdp” just because the measure has some flaws , but we can certainly add to our information about the economy by supplementing with other measures , and wealth is the most obvious first choice , because a dollar of wealth spends exactly like a dollar of income , once converted ( at the speed of light , oftentimes ) into the appropriate paper or digital format.

    If we now have two accounting methods that give us reasonably close agreement on “national wealth” , both of which attempt to measure essentially the same aggregate , I can’t help but see that as about as useful and as reassuring as knowing that gdi equals gdp. It’s not the end-all and be-all , but it’s progress.

  390. Ramanan writes:

    SRW,

    Also, there’s income inequality and wealth inequality. These are different things but one would expect a good correlation between the two.

    But if you argue that because there is wealth inequality, aggregate wealth doesn’t make sense, you also have to argue the same for income inequality. Do we then not take line 3 of F.3 i.e., “compensation of employees” seriously?

  391. The value of non-financial corporate real assets is $ 20 trillion from S.5.a. The value of financial corporate real assets is $ 2 trillion from S.6.a. From what I’ve seen in supporting explanatory text, I think their reasons for using equity claim values instead are very flimsy indeed

    But of the nearly $30 trillion in market value of equity, only $24 trillion is owned domestically. It’s not so very far off, really. It is probably a bigger difference (with opposite sign) if you go back a few decades.

  392. Ramanan —

    Proofs are in puddings. I’ll take any measure that reliably correlates with something interesting as a measure of that thing. I’ll take no measure as interesting based solely on the internal coherence of its production. The Greece/Germany example is just one example of how aggregate-accounting “net worth” comparisons simply fail to correlate with very much that seems meaning, and correlates particularly poorly with the intuitions we normally associate with phrases like “wealth”. There are other measures that I think are very problematic.

    Aggregate employment compensation doesn’t fall out of nearly as uncertain an accounting exercise as aggregate “net worth”. That it may be included as an input into the computation of flows into the latter doesn’t make the latter any more reliable or useful. I understand what aggregate employment compensation means, in a very straightforward way. Like NGDP, with perfect surveillance, we could measure it close to perfectly (although there would remain some complications regarding noncash compensation). It simply isn’t analogous to “net worth”, because “net worth” implies valuation and decisions about what does and does not qualify as an asset or liability an order of magnitude more complex than anything involved in measuring or estimating employment compensation. Plus, I have no idea what aggregate net worth means or is supposed to mean: it certainly does not mean the price at which a sector could sell its assets (who would be the buyer), or a discounting of cash or value flows from those assets to the sector (as many of the household sector’s financial assets are partially claims back against the households themselves — future flows from the assets imply flows to the assets).

    Again, if you persuaded me that SNA net wealth correlated with something I could count as meaningful, and make the case that it would consistently do so, I’d accept it as a measure of that. But I know that aggregate net wealth measures deceive people who understand them to be analogous to, say, firm wealth. Greek households are not richer than German households, and a measure that suggests they are is flawed, at least for that purpose. German incomes are larger than Greek incomes by lots of measures that don’t remotely depend upon the net worth computation they might help inform. GDP-based measures, on the other hand, are dull, but they actually do correlate fairly well with lots of stuff of interest, so I’m glad to accept them as useful. Survey- and tax-based measures of income and wealth inequality within a nation are imperfect, but they are what they are, they are not meaningless. SNA net wealth may not be meaningless either — I’m agnostic, I don’t want to draw a hard line. But SNA net wealth does not remotely or accurately capture the kind of intuitions people jump to when the look for an overall wealth measure.

    “Unbacked repricings” — by which I meant repricings of financial assets unrelated to the productivity of whatever real assets they lay claim to — are just an example of their failure to match. When US stock markets rise, US “household balance sheets improve”, and we end up with Alan Greenspan telling us that easy credit is no problem, because “household balance sheets have never been stronger” (or something like that). But changes in stock prices reflect a shift in the distributional of purchasing rather than a meaningful change in the value of the productive assets of firms. When stock prices rise, the household-sector balance sheet looks good. But when stock prices fall without loss of productive capacity, representing a shift of relative purchasing power from shareholders to people reliant on the uncapitalized asset of labor income, the household-sector balance sheet deteriorates. Very arbitrary choices in the definition of “net worth” lead one group’s welfare to be taken as the whole group’s welfare. (And that group’s welfare is taken as the whole group’s welfare way too much.) A similar story can be told with respect to housing. A mere repricing of existing home and rents, unbacked by any change the the quality of housing services provided, is a benefit for existing homeowners and a cost to people short a house. Yet on the household balance sheet, it shows up as net wealth. That’s just bad.

  393. Marko —

    Here’s a graph of your synthesis of B.1 wealth (from a comment way above) against conventional real GDP, indexed at 100 at 1980-01-01. I’ve divided your measure by CPI, to compare both in “real”-ish terms.

    https://research.stlouisfed.org/fred2/graph/?g=2bRO

    Over the long term, they match fairly well (much better than I expected, frankly). But the net worth measure is much more volatile. It’s pretty clear what the noise is in the synth B.1 measure: It’s asset market (stock and housing) fluctuations. In fact, if we treat GDP as the “trend”, this looks like a fairly nice picture of the “mean” to which asset prices revert. Which, as a sometimes speculator, strikes me as a pretty interesting graph to have!

    But as a “wealth” measure, the divergences between synth B.1 wealth and GDP don’t strike me as very helpful. Again, asset price gains get temporarily captured as “wealth gains”, and then revert. (Going forward, I wonder will they still revert? I think that the much of US economic policy is now devoted to “stabilizing asset prices”, that we’ve grown accustomed to levels of financial asset and housing wealth higher than would be justified by the GDP trend since the 1990s, and are trying to stabilize around a new multiple.)

    What I see is a recapitulation of Shiller’s point that asset prices fluctuate more and with higher frequency than “fundamentals” can justify.

    I won’t claim this new measure is uninteresting! But if we are going to interpret one as “aggregate wealth”, I’d choose the one that’s much more stable. I think we would (and did!) make consequential errors if we interpret the 2007 synth B.1 peak as something like durable wealth. If we graph the two series in logs…

    https://research.stlouisfed.org/fred2/graph/?g=2bS0

    …we get what appear to be periods of incredible growth — in 2002-2007, very recently — during periods when it doesn’t seem our real economic performance was all that great, but when prices of some important asset classes did rise briskly.

  394. Marko —

    Same as the first graph above, but with the longest available start-of-year base, 1952-01-01.

    https://research.stlouisfed.org/fred2/graph/?g=2bSa

    That doesn’t look very different from the 1980 choice of basis, but that’s not true in general, the picture is sensitive to coice of base. Basing at 1962, for example, puts the synth wealth graph generally below the GDP graph, for example.

    https://research.stlouisfed.org/fred2/graph/?g=2bSi

  395. SRW,

    This seems pretty extreme to me. I do not think this is supposed to be so complicated. Just as GDP is not an all-encompassing measure of welfare (if it is even an adequate measure) neither is net wealth. It does not bother me in an accounting sense that household net wealth will rise with house prices. Rather, it is important to understand the effects of such a change.

    I do not think there is tremendous import to the net worth figure. It does give some sort of scale to say the market value of nonfinancial and net financial assets comes to about 4-5 times the market value of one year’s production. I am disposed to think positively of Piketty’s general framework.

    I also disagree that “perfect surveillance” would yield us a close to perfect measure of NGDP– at least in a sense that net wealth does not. As the recent revisions show, there are questions about what should and should not count as production. Are housing services really consumption produced by residential investments? If we want to count intellectual “property” should it go in investment, or should we count the consumption which flows from? As Piketty (among many) observed, much capital income is actually mixed.

    Regardless,

    I have no idea what aggregate net worth means or is supposed to mean: it certainly does not mean the price at which a sector could sell its assets (who would be the buyer),

    is a critique which baffles me utterly. Are you suggesting that the problem with our measures of sectoral net worth is that an entire nation of households cannot in practice sell their assets and dispose of their liabilities and so it is meaningless to add up the marginal contributions at current prices?

    or a discounting of cash or value flows from those assets to the sector (as many of the household sector’s financial assets are partially claims back against the households themselves — future flows from the assets imply flows to the assets).

    Why is this a problem? Net wealth includes liabilities as well as assets; it accounts for claims in both directions. Hence– I do believe– the “net” in net wealth/worth. It seems to me that this is well accounted-for.

    It is always possible that I have misinterpreted your concerns.

  396. But as a “wealth” measure, the divergences between synth B.1 wealth and GDP don’t strike me as very helpful. Again, asset price gains get temporarily captured as “wealth gains”, and then revert. (Going forward, I wonder will they still revert?

    If by divergences, you mean fluctuation in the ratio of wealth to GDP, then we’re close to talking about Piketty. I think we observe capital income as more stable than wealth and hence fluctuations in the wealth to GDP ratio associate negatively with fluctuations in the average rate of return to wealth. On a less broad scale, this kind of thinking led to accurate guesses that stock and real estate markets would correct to match income, rather than the other way around. (E.g, thinking that rent prices would not rise rapidly to raise the lowered rate of return on housing, or that corporations would not become hugely more profitable to raise the lowered rate of return on equities…)

    To my eye it is less useless than it is insufficiently explored.

  397. Marko writes:

    SRW,

    I agree with you about the distortions due to bubble valuations. That’s a factor that needs to be kept in mind , more so since the dotcom bubble days than prior to that. Here’s a graph like the ones you presented using the pce deflator rather than CPI , to normalize both gdp and wealth to 2009 ( instead of 1982-84 with the cpi deflator ) :

    https://research.stlouisfed.org/fred2/graph/?g=2bTs

    I’ve set the comparison to 3.8 times gdp , a baseline to which wealth seems to revert post-bubbles , as you noted.

    As Piketty’s work has shown , over time even the “stable” multiples of wealth over gdp can change , and those changes can be indicative of important distributional shifts. My relative satisfaction with the new B.1 measure is mainly due to my hope that it’s a sign that we’ll start paying more attention to the double-barreled distributional blasts that occur as wealth plus income gaps widen , rather than looking at only one or the other in isolation.

    There are two pies , income (gdp , roughly) , and wealth , both growing at more or less the same rate over the last 70 years. It’s important for people to understand that the wealth pie is ~3.8 times bigger than the income pie , so our concerns about inequality are misguided to the extent that the focus remains locked onto the income component. The upward redistribution we’ve seen over time has involved both components , and the lucky recipients at the top are largely the same group of people for both components.

    I won’t be entirely satisfied until we have an administrative database showing a Haig-Simons-type income measure for every individual , so I don’t think this new wealth measure means the job is complete. Far from it.

    https://anticap.files.wordpress.com/2014/11/wealth-same.jpg

  398. Ramanan writes:

    SRW,

    Without going into details … there’s nothing wrong with accounts .. the inconsistency arises because one might be making so many implicit assumptions in the way the world works.

    It’s certainly possible for person A to be richer even if his/her income is less than person B. Person A just saves a lot.

    Another example. Person C may have just started working but with a burden of a student loan and negative net worth and Person B with far lesser income may just be retiring and having a lot of wealth.

    So do I then reject the notion of net worth at an individual level?

    In the same way it’s possible for Greek households’ to have more average net worth than German households for many reasons.

    I am not sure about your point about the stock market. My point was that stock market booms can lead to consumption booms because people feel richer and may increase their expenditures. Is it incorrect to say that consumption can depend on both holding gains and net worth (in addition to depending on income)?

    In this example, there’s no backed or un-backed anything. Boom created demand and firms produced more to meet the demand. What is this backed-thing? Purely a psychological thing as much as a stock market boom feeding on itself leading to more earning for firms.

  399. David —

    The “net” in “net wealth” is not very good.

    Suppose that some units in the household sector issue debt to other units in the household sector. The net wealth arising from that arrangement would, properly, be zero. Now suppose that some units in the household sector sell water to other units in the household sector. The sellers hold shares in a water company, which we’ll assume issues only equity. The value of those shares becomes household net wealth. And to the degree that that value represents the water producers’ capacity to discover, obtain, purify, and distribute water, that accords with our intuitions about wealth.

    However, suppose that the water company is decides to take advantage of its monopoly power, and it sets the price of water very high. The value of the equity shares, “net household wealth” in this accounting, would rise dramatically! But in every meaningful sense, households have been impoverished by this operation.

    I think this kind of thing is what Steve Roth means in his post when he refers to a need for a “zero rent” assumption. But I don’t think that’s remotely realistic. I think very much of what is captured by household net wealth represents, for example, the capacity of some property owners to exploit monopoly power conferred by location and restrictive zoning. The implicit liability of nonproperty owners to cough up a stream of high rents or an inflated purchase price is not netted out of “net wealth”.

    Similarly with stock markets. When interest rates experience a secular fall, that is a redistribution of returns from future holders of financial assets (who will enjoy lower returns) to current holders of the asset (who experience a capital gain). It is not meaningfully an increase in wealth. Yet it is accounted as “net wealth”.

    I want to clarify, I am not “against” any of these net wealth measures for any and all purposes. As long as we are cognizant of all these things, I don’t deny that they might be useful. Piketty-esque analysis requires I think a measure of gross wealth, rather than net wealth, so I don’t think that these measures are quite appropriate. But, despite having read (well, largely listened to) that whole very long volume, I don’t think it’s precisely clear what is the best measure of what Piketty means by capital, and my sense that gross measures would be better could be wrong. I’m certainly open to FOF/SNA net wealth measures representing useful quantities as saying something about the claims we hold against one another.

    What I do not think they reasonably represent though is what I took Steve Roth to be pining for at the long ago beginning of this conversation, a measure of collective wealth in the normative sense of being well or wealthy, a summary statistic of how well off “households” or “the private sector” or “the nation” is. I think, by virtue of how they are constructed, that they are likely to be poor measures of that. I think that GDP is imperfect but not terrible at that, and given the exercise Marko and I have gone through, I have been surprise that these net wealth measures are not as terrible as I’d expect if we use GDP as a standard. But to the degree that these measures deviate from a multiple of GDP, I think that what they reflect are things that ought not be interpreted in any normative sense as “wealth”.

  400. JKH writes:

    Net wealth is obviously zero – because somebody pays the price that creates the marginal income to capital or the marginal subsidization of present value by future value. For every income or capital winner there is a whirlwind of non-winners. This corruption of distribution is encouraged by accounting. So let’s outright ban this output and income value from the accounting books. Declare interest and profit and dividends of zero everywhere. There is no such thing as net wealth, and those who have been left behind by others pretending it is otherwise are mocked by an outrageous accounting illusion that doesn’t yet net to zero. And God dammit – let’s ban wealth volatility by eliminating net wealth itself.

  401. SRW,

    I think it is your example which is “not very good”

    To begin with, “in every meaningful sense, households have been impoverished by this operation” is misleading as the owners are enriched. If demand is inelastic, then this is largely a transfer from some households to others.

    But what of the accounting? In the IMAs, U.S. net worth does fall– to the extent that foreign holdings of the water company stock rises and the U.S. NIIP falls. This suggests an increased flow of income out of the country, which also suggests an overall loss to households one way or another.

    Of course, much of this (“demand is inelastic” “suggests and increased flow”) is modeling and not accounting. Does the increased wealth of the company owners increase their contribution to demand by more than the loss due to lower real income of the consumers? That’s a modeling question. But the IMAs do show a loss of national wealth.

    Piketty-esque analysis requires I think a measure of gross wealth, rather than net wealth

    Do you mean Piketty should have used gross rather than net? That strikes me as quite strange. If I lend you $10, that should increase our combined wealth for purposes of computing our combined net income? I think this goes against what you suggested with your water company.

    In short, if I read JKH at #400 correctly, I second this modest proposal.

  402. I should have said “modeling out combined net income” and not “computing”

  403. Marko writes:

    One of these days we’ll notice a new table in the flow-of-funds – B.1A++ , perhaps – that provides a more comprehensive measure of “national wealth” , including such things as gov’t land values with all mineral and natural resource components tabulated. Same for coastal resources. Unused but potential energy sources , like solar , wind , geothermal , wavepower , etc will all be assigned some value. Human capitalization will include such things as happiness and wellness. Everything you can think of will get a number.

    My question is this : How many posts will be in the thread that follows that release ?

    Surely more than 400. What , maybe 4000 ?

    Anyway , I can’t wait.

  404. JKH writes:

    excellent “pro accounting identity” post by Michael Pettis:

    http://blog.mpettis.com/2015/10/how-to-spend-thin-airs-endogenous-money/

  405. JKH writes:

    David Glasner also has an excellent post dated October 20 in response to Pettis’ post above.

    (For some reason, this site is rejecting the link)

  406. JKH writes:

    regarding # 405:

    uneasymoney blog

    keynes-and-accounting-identities

  407. JKH writes:

    btw,

    the Glasner post may be loosely described in context as “anti accounting identity”

    this is a Glasner theme, in context, in contradistinction to the Pettis take

    my favorite passage from Pettis is from the section titled “Violating Identities”:

    “The important point about accounting identities – and this is so obvious to logical thinkers that they usually do not realize how little most people, even extremely intelligent and knowledgeable people, understand why it matters – is that they do not prove anything, nor do they create any knowledge or insight. Instead they frame reality by limiting the number of logically possible hypotheses. Statements that violate the identities are self-contradictory and can be safely rejected. Accounting identities are useful, in other words, in the same way that logic or arithmetic is useful. The relevant identities make it easier to recognize and identify assumptions that are explicitly or implicitly part of any model, and this is a far more useful quality than it might at first seem. Aside from false precision, my biggest criticism of the way economists use complex math models is that they too-often fail to identify the assumptions implicit in the models they are using, probably because they are confused by the math, and they would often be forced to do so if they weren’t so quick dismiss accounting identities on the grounds that these identities don’t tell you anything about the economy.”

  408. Marko writes:

    Pettis’ stuff is always worthwhile , but he’s in dire need of a better comment system.

    You could establish a more spontaneous , conversational , and reliable stream if you transmitted all the comments by carrier pigeon.