Insure depositors, not banks

Raghuram Rajan offers a thoughtful comment on regulating big banks (hat tip Mark Thoma). The upshot is that bigness, interconnectedness, and proprietary trading are difficult to define, and hard caps are likely to be gamed and evaded. Rajan advocates more subtle strategies “such as prohibiting mergers of large banks or encouraging the break-up of large banks that seem to have a propensity for getting into trouble.” That strikes me as weak tea. I’d prefer that regulators explicitly target a diffuse market structure under which any bank can be let to fail, and that they impose graduated-to-prohibitive taxes on a wide variety of markers of bigness and badness in order to meet that target. Successful banks should grow by division rather than accumulation.

Like Ezra Klein, I was especially intrigued by Rajan’s concluding paragraph:

In reality, proposing limits on size and activity is just an attempt to diminish the deleterious effects of another previous and now anachronistic intervention — deposit insurance. When households did not have access to safe deposits, deposit insurance made sense. With the advent of money-market funds, households gained access to near riskless deposits. Money-market runs can be eliminated by marking them to market daily; they do not need deposit insurance. To encourage community-based banks, deposit insurance may still make sense because small banks are poorly diversified and subject to bank runs. But for large, well-diversified banks, deposit insurance merely contributes to excess. We will bail out these banks anyway in a time of general panic. Why encourage the poorly managed ones to grow without market scrutiny by giving them deposit insurance along the way? Why not phase out deposit insurance as domestic deposits grow beyond a certain size? That would be far more effective in reducing risk than size or activity limits, and far easier to implement.

I’ve a bunch of nits to pick with this: money-market funds are not perfect substitutes for state-guaranteed assets, mark-to-market doesn’t limit aggregate risk or prevent runs (it just makes the race to the exit a bit fairer, and the collapse a bit sharper, as everyone learns when to get nervous at the same time). If large numbers of ordinary households, swayed by yield or convenience, hold funds at an uninsured Chase, Citi, or BoA (cue the television commercials claiming “large, well-diversified banks” to be solid as Gibraltar), that fact would make each of those banks individually too big to fail even in the absence of a “general panic”. No system that expects sales clerks and schoolteachers to monitor financial firms is reasonable or politically sustainable. (If ordinary households can’t be persuaded to do without FDIC insurance, Rajan’s proposal would amount to an end of large depository institutions, which wouldn’t be a bad thing. Depositors would migrate to smaller banks, and large institutions would fund themselves as pure-play investment banks, unless regulated by other means.)

Of course I have a better idea. Rather than insuring banks, the government should insure depositors individually for losses they suffer on deposits at any FDIC-approved bank, up to a pretty high limit (say $1 million, indexed to inflation). Ordinary households would be unaffected by this change, as most families hold balances far less than the insurance cap. They could continue to deposit funds at the FDIC-approved bank of their choice without fear. Affluent households would no longer be able to play the wasteful game of evading insurance limits by splitting funds among different types of accounts and institutions. The affluent would be expected to monitor and help discipline the firms in which they invest. This is both fair and politically credible. It’s fair, because pushing wealth forward in time requires hard information work, and those who wish to push a lot of wealth forward (and earn interest on top!) should contribute to the effort. It’s credible, because ex post facto bailouts for underinsured depositors would be a hard sell when the underinsured include only wealthier depositors, who would not be reduced to penury but, at worst, to a level of affluence most households never achieve, simply by maxing out their government insurance.

Insuring depositors rather than banks wouldn’t, and doesn’t purport to, resolve the too-big/bad/sexy-to-fail problem. It would be a modest change that would eliminate some of the gaming that permits affluent investors to shirk their duty of discipline, and that would improve incentives to monitor by reducing the likelihood that notionally uninsured depositors get bailed out. But all of this matters very little in a world where even junior, unsecured creditors of some banks enjoy an implicit state guarantee.

Update History:

  • 27-January-2010, 5:15 a.m. EST: Touched up the text a bit, fixing some awkward sentences. No substantive changes.

145 Responses to “Insure depositors, not banks”

  1. This is what that money market suggestion looks like to me (I may be misinterpreting it):

    People would only leave small amounts of money in a bank’s checking account, because it’s not insured. Any more money would not go in a banks savings account, or in a CD, anymore; it would instead go in a money market fund, like one of U.S. T-Bills, super safe, perhaps through the bank as an intermediary, like what Vanguard does.

    If this is true then doesn’t this greatly change the nature of a traditional bank?

    They won’t have peoples savings account money and CD money to loan out anymore; all that money will now be placed with the government via T-Bills, or corporations via commercial paper? Where will traditional banks replace this big source of loanable funds?

  2. […] By insuring depositors to the losses of banks, we solve the too-big-to-fail problem.Close […]

  3. […] Steve Waldman isn’t too happy with that paragraph either (”No system that expects sales clerks and schoolteachers to monitor financial firms is reasonable or politically sustainable,” for instance). Waldman instead proposes to insure depositors rather than bank, which is an interesting but as he points out, modest, idea. […]

  4. a writes:

    Was the justification for deposit insurance to help depositors out? Or to stop bank runs?

    Expecting the affluent to look after the banks is not going to work, since the affluent won’t be able to, and simply won’t, do it. The risk averse will simply put 1M (or whatever the limit is) in a bank, and the rest in TBills – end of story. And those who seek yield will put their money in the same troubled institutions which can only attract money by paying high interest rates, thinking that they will probably get their money back because, if the bank gets in trouble, they will be the ones to pull their money out before everyone else. And that kind of belief is precisely what leads to bank runs.

  5. winterspeak writes:

    Raghuram Rajan, who I have met, has no idea how the banking system works. Otherwise he would not be advocating any market discipline on the liability side of the balance sheet. This is pure Chicago School ideology talking.

    A bigger problem is his targeting the deposit section of the liability sheet. He believes that banks somehow lend out deposits. This is pure academic Economics talking, and therefore, the Fed and Treasury. It is equally nonsense.

    SRW: I’m afraid your suggestion is weak tea mixed with thin gruel. The correct way to treat deposits is to FDIC “insure” then with no limit, and fee “insurance fee” on banks for doing so. This should get rid of money market accounts, which are a direct reaction to the FDIC cap, needless, and horribly destabilizing.

    The actors who need to regulate banks are investors (either on the debt or equity side), regulators, the board, and management. Not depositors.

  6. […] Insure depositors, not banks.  (Interfluidity) […]

  7. csissoko writes:

    re: money market funds: “mark-to-market doesn’t limit aggregate risk or prevent runs”. It seems to me that mark-to market does prevent runs, because if you join the run assets will be sold to honor your claim and you will get the market (fire sale) value of those assets. Everybody who doesn’t run still has a claim on unsold assets and can benefit from their recovery in price when the fire sale is over.

    Whether the typical money market fund investor would be willing to put up with this kind of volatility is an entirely different question. But I think Rajan does propose an environment where we can put deposit insurance to the market test. If you value deposit insurance you go with a small bank, if you don’t, you can put your funds in an uninsured bank.

    What happens to large banks in this environment will depend on how bank runs are handled. I think that (i) if there is a standing RTC facility and the assets of a large bank that faces a run are immediately auctioned to the smaller banks that have probably absorbed all the escaping deposits (that is, there is a carefully considered and speedy resolution process for the inevitable bank runs) and (ii) deposits have super-priority over all other liabilities — including derivatives, the system could work. I wouldn’t expect to see many large banks, however.

  8. Steve Randy Waldman writes:

    Richard — If deposit insurance were eliminated by ordinary depositors turn out to be pretty averse to uninsured deposits (and behaviorally capable of expressing the preference despite the sweet persuasions of bankers), as you say they’ll insist on putting their savings in full-faith-and-credit mutual funds and holding only small transactional balances in banks. Rather than saying this will sink the loanable funds market, I think it’s better to say it will make explicit that the idea of a loanable funds market is pretty much inoperative under an interest-rate setting central bank. As the Chartalists (e.g. winterspeak) like to point out, accurately, at the level of the banking system as a whole, the Federal Reserve has to supply whatever loanable funds are necessary to prevent its target policy rate from rising. If there are good projects that earn more (in risk-adjusted terms) than the Federal Funds rate, banks will lend, borrow on the interbank market, forcing Fed purchases of bonds to hold down the FFR (unless there are already excess reserves available paying approximately the FFR), ensuring there are sufficient funds in the system to make the loans.

    a — The runnability of banks is not a binary thing, and not managed under the present system solely by deposit insurance. The only way to make banks immune to runs with deposit insurance only is to offer blanket, unlimited deposit insurance (as Winterspeak advocates). I think that’s a bad idea. We want banks to face runs as their solvency decreases as part of the market discipline they face. We just don’t want banks to be so “runnable” that there are “sunspot” runs — irrational stampedes out of banks that would be solvent absent the run. Limited deposit insurance is intended to reduce the likelihood of runs so that they only occur when people have a pretty good reason to believe their funds are at risk. With a $1M individual deposit insurance I think the fraction of insured deposits would be pretty high. There’d be no lines in any street to freak people out. “Runs” be the the result of considered judgement by multiple individuals that their bank is insolvent. It’s possible, of course, that there is just some rumor bad enough that, even with a high high fraction of deposits insured, provokes a run by the underinsured. But that’s why modern banking systems have a second backstop against runs — the lender of last resort facility. A run on a bank no longer leads automatically to its closing. Instead it triggers an option whereby regulators may 1) supply liquidity unconditionally, if they deem the run irrational; 2) supply liquidity in combination with regulatory corrective action, if they deem the run to have been a “canary in the coal mine” for a still salvageable bank; or 3) shut the bank and payoff insured depositors. Allowing socially nondisruptive runs to trigger regulatory scrutny is a healthy and clever way to let markets inform the ever-gamed regulatory process.

    winterspeak — I knew we were gonna be at loggerheads on this one. I disagree with you (and even more with @wbmosler) that the liability side of bank balance sheets is the wrong place to discipline banks. Of course, it’s true that historically, market discipline on the liability side has always failed eventually, because banking systems always fall into crisis. But then regulatory discipline on the asset side has always failed as well. To the degree banking systems have worked, in times and places and for a while, I think liability side market discipline has made a positive contribution.

    In general, you seem to agree that liability-side market discipline is important, because you expect bank equityholders to discipline banks. That’s as “liability-side” as deposits. (Mosler seems to emphasize only asset-side regulation.) The way that I think about it is in options terms. I’d be content with a banking regime that was 50% equity financed, but with other liabilities immune from loss (and therefore capital provided solely by the state.) With 50% financing, equityholders would have incentive to make prudent loans, rather than maximizing option value via volatility. But if banks are going to be levered 5x, 10x, or 20x, the option value to equity holder is too high to rely solely on equity (even with regulation) to discipline the bank. It becomes too tempting, in rational, financial terms, to maximize option value by shooting for volatility. Except during the height of credit booms, private creditors don’t provide that kind of leverage without covenants, which effectively grant creditors the right to run and potentially shut down the firm when various triggers that might signal gambling or distress are breached. The limited runnability of banks does the same thing. Again, this is not a binary question: one doesn’t want banks to be so runnable that they are brought into crisis by a butterfly flapping its wing on Mars. And one wants a lender of last resort, so that runs (like covenant breaches) trigger scrutiny, not automatic failure. But one does want some discipline to be provided by large depositors.

    Re money market funds, I mostly agree. There structure is explicitly designed to hide the fact that they are risk investments. They are made to look cash-like, which encourages investors to treat them as cash, which essentially forces a state guarantee since most people fail to risk-manage “cash” holdings. I don’t think it’s too much of a burden of liberty to require that investment vehicles not automatically dividend or dilute to maintain a NAV of 1. That strikes me as a form of false labeling. Let mutual finds be mutual funds, and let value changes appear in price, to help people understand that fluctuations in value can go both ways. (Unfortunately, investments in commercial paper will make the price seem nearly always to go up: it’s a tail-risk kind of strategy. But at least the price will change, not always be 1 like a dollar bill.)

    I would point out that besides evading FDIC limits, many people who could hold Federally insured funds opt for money market funds simply in order to reach for yield. Interestingly, the investment industry was able to persuade even fairly risk averse savers that these uninsured savings vehicle were safe enough. (I don’t think many money-market funds participants expected ex ante the guarantee they eventually received post-Lehman, although looking forward there’s now a bad precedent.)

  9. Steve Randy Waldman writes:

    csissoko — It’s not true, with money market funds or with banks, that investors with confidence in the underlying assets can sit out a run and wait for asset values to recover. Assets aren’t segregated by depositor/investor, and first-runners get payed full nominal values, regardless of the fire-sale prices their running has forced. Depositors/investors who don’t run may end up with nothing at all, even though the hold-to-maturity value of the underlying assets was fine before the run began. That’s why it’s critical for banks to have a lender-of-last-resort or to be shut very promptly (at regulatory discretion) when runs occur, because blind selling to satisfy liquidity needs transfers wealth from “slow” claimants to external parties.

    [In practice, this is mitigated somewhat for money-market funds by the liquidity of the securities that they hold. But even commercial paper sometimes becomes illiquid and price-uncertain, and neither daily mark-to-market nor dividending to $1 NAV affect these issues. It is simply the case that a bank holding liquid assets is less vulnerable to forced loss of capital during runs.]

    I don’t think we could rely upon a system where bank sold assets into a standing RTC-like facility to satisfy runs, and depositors essentially repurchased them at other banks, because of price uncertainty in the short time frame necessitated by the press for liquidity.

  10. winterspeak writes:

    SRW: You are right, I don’t take as extreme a position as Mosler on freeing the liability side entirely from market discipline. But I absolutely do not believe that depositors should be responsible for any market discipline at all, no matter how big their checking accounts get. Investors (both debt and equity) are a different matter, and I am open to suggestions here.

    What do you mean by 50% equity financed banking regime? That you have 2x capital requirements? (so you have only have $2 of assets on every $1 of equity on your balance sheet?). That’s a very strict capital requirement, if that is what you mean. Our financial system is massively more leveraged than that, and would need a huge slug of equity to meet such a requirement and maintain AD.

    In such a regime, and in any capital requirement regime, you need to regulators to say whether $2 of assets are really worth $2. You cannot get around needing regulation on the asset side. I’d be interested in seeing many shadow banks at least return to private partnerships. That way the equity liability plus the asset valuation get tightly re-coupled. Banks will always lie to regulators. Banks will lie to themselves if agents get paid end of the year. Banks will lie to themselves least when they get paid only when their loans do.

  11. csissoko writes:

    SRW: So are you saying that MMMF withdrawals take place in real time, rather than end of day like standard mutual funds? I had assumed that when MMMFs switch to daily marking, they would like mutual funds only allow redemptions at end-of-day values. Perhaps I was wrong about that, but it doesn’t sound like the SEC has finished writing the mark to market regulation for MMMFs yet.

  12. David Merkel writes:

    Have you seen my proposal on money market funds? Neat and simple. Even Felix liked it.

  13. Indy writes:

    Let’s say you have a wealthy individual with $10 Million he wants to keep in safe, mildly interest-bearing nearly-perfectly liquid assets, and under the policy of Waldman’s $1 Million individualized insurance proposal. He can choose between:

    1. Putting all $10 Million in bank accounts and monitoring his banks to the best of his ability for fiscal propriety and soundness, and moving his money around accordingly, as per his incentives, or

    2. Putting $1 Million in one bank, and buying $9 Million in marketable treasuries (as Richard Serlin spoke of).

    Consider the monitoring costs imposed under option 1. We might ordinarily want this person to diversify and spread his assets over a good number of institutions (under assumptions of uncorrelated risk), but the more institutions in which he places the deposits, the more time, money, and energy he’ll have to spend monitoring.

    He may also not believe himself or even hire-able experts and advisers to be particularly good as monitoring banks comprehensibly and with a decent capacity to predict future solvency. (Weren’t most of our experts very, very wrong on this precise matter a few years ago?). In other words – and despite his best monitoring efforts – he still has to live with a certain amount of uncertainty.

    On the other hand, the monitoring costs of option 2 are zero (assuming he doesn’t have to monitor the government too!). He is insured to his $1 Million, and the rest is backed by full faith and credit – so he is in effect fully insured by the government.

    I find it doubtful that the benefits of option 1 over 2 would ever exceed that difference in monitoring costs and risk. But I could be wrong, let me know if I’m missing something.

  14. winterspeak writes:

    David: Your proposal is ridiculous, for obvious reasons (and no, I don’t mean Felix Salmon’s approval).

    Indy: Your example illustrates the problem very nicely, thank you. If you want to just park your cash, the right place to park it is with the Govt, and as banks are public private partnerships, FDIC insured accounts solve this simple problem 90% of the time. To get to a 100% fix, just lift the cap on FDIC insurance.

    This really is the only part of the financial system that almost, sort of, works right. That it ends up in the cross hairs so often reflects some truly bizarre ideas about how the world is, how the world could be, or both.

  15. Steve Randy Waldman writes:

    csissoko — No, I’m not making any claims that depend on the frequency with which redemptions can be honored, or at least I don’t mean to be. My understanding is yours, that money markets work like mutual funds and permit purchases and redemptions at end-of-day NAV, although I’ve never paid enough attention to be sure about that. But I don’t think my claim that runs harm those who sit them out in money market funds is sensitive to the frequency of redemptions (except in the case where redemptions are suspended so that all are effectively simultaneous or selling is slow enough there is no fire-sale/liquidity discount attached to the realized values of fund assets).

    David — I like your proposal. Generally speaking, I hate the version of finance that tries to be magic, and pretend that asset are more liquid or less risky than they are. Banks and money markets, even when assets are fundamentally sound, are a means of providing liquidity insurance: they pool liquidity risk. Just like any insurance company, too many simultaneous claims can render them unable to easily meet their obligations. Investors in banks and money markets (and other mutual funds too) should know that. Banks and mutual funds are also investment companies that take on asset value risk as well as liquidity risk. Investors should know that as well. We choose as a matter of public policy to hide these facts from people who do not wish to consider themselves investors, and for whom becoming investors would be an inefficient use of time: ordinary depositors for whom banks are primarily facilitators of transactions (or lenders), not savings vehicles. I think that’s good public policy. But once one becomes an investor — if one reaches for yield in a money market, or one manages substantial savings, I think one does become an investor — one should understand the actual characteristics of the vehicles in which one is investing.

    Your proposals are clever, in that they explicitly introduce valuation risk into vehicles where vehicles have been permitted to lazily ignore it, and they potentially flip around the dynamic which rewards hair-trigger panic selling of potentially good assets and punishes those willing to “sweat through” on an independent valuation. There are devils in details: when the fund manager adjusts down individuals’ number of units, she sends a mixed signal. If I see that my holdings have dropped steeply, say from 100 to 95, I don’t necessarily know whether that means that the NAV behind my position has fallen 5%, or the manager is building in a large premium to discourage redemptions. In your proposal, you suggest a fixed formula, which resolves the problem, but might not be appropriate for all assets and conditions. If commercial paper is suddenly very volatile due to a credit crunch, the “maintenance margin” built in to the fund should be large to reflect that volatility, whereas during ordinary times a 2.5% margin might be more than sufficient. I don’t think these are at all insurmountable issues (and the fixed formula might be fine, or at least better than current practice), but they are interesting to think about.

  16. Steve Randy Waldman writes:

    Indy — You’re mostly right, and I consider that a feature, not a bug. The thing I think you are missing is yield. Banks compete for deposits, and they pay higher than risk-free rates to attract (term) deposits. Right now, lots of risk-averse affluent investors ply the moral hazard trade, seeking out the highest-yielding FDIC insured CDs and divvying their funds among them to stay within insured limits. That’s paying people to do socially useless work (searching out differentially priced CDs, tracking and managing many accounts).

    In the world I’m describing, affluent savers would have to make a choice: follow the strategy that you suggest dominates, and accept risk-free Treasury rates on their savings beyond the insured limit, or seek the extra return offered by banks, but bear some of the risk that generates that return. Those who want the return can employ whatever risk-management strategy they please: they can blindly diversify (as has become sadly popular), or they can choose a few banks whose solvency they monitor. I think that’s a much more socially productive choice than letting people earn risk-yields for (nominally) risk-free investments, if they are willing to do a lot of useless busy work to make it happen.

    It might be true that everyone would value your strategy 2, and put all their uninsured wealth in Treasuries. I’m fine with that (well, mostly, I’m a bit of a radical in that I think people’s access to Treasuries should be rationed). It offers no subsidy to bankers to waste real resources by seeking volatility in order to maximize the option value associated with poorly monitored leverage. (Alas that problem remains for leverage offered by the insured balances, but it renders the banking system substantially more transparent by eliminating a significant chunk of the obfuscatory finance, which would have to be replaced either by other private or more explicitly public source.)

    It might also be true (in fact I think it would be true) that banks would offer sufficient yield differential to attract a lot of uninsured deposits (as they currently do now in mutual funds), so some risk-averse savers would go for strategy 1 and other would go for strategy 2. (Many retirees, however risk averse, are very very sensitive to yield. As a rule of thumb—an unwise one to the incautious—they try to “live off the income”, or annuitize their withdrawals in a manner which is very sensitive to yield.)

    Winterspeak — I gotta run to make the SOTU. More after.

  17. csissoko writes:

    SRW: If redemptions are at end of day, all the SEC needs to do is require that the fund sell assets from 4 to 5 pm to honor all cash withdrawals and allow “end-of-day” in the particular circumstance where this is necessary to be 5pm (one hour after the last withdrawal request can be placed) or some similar standard. In short, it becomes a trivial problem to solve the “bank run” aspect of MMMFs.

  18. csissoko writes:

    Probably it would be easier to simply require that all withdrawal requests that are valid for the current day be placed by 3pm — one hour before markets close. Any withdrawal requests placed after 3pm are automatically rolled to the next day. As long as the SEC approves this policy — and I can see little reason not to — the whole “bank run” problem disappears completely.

  19. Rogue writes:

    Steve, I seem to be getting that you are proposing to insure banks, not depositors, (and not the other way around, as it is today). If banks are the ones insured, depositors have no incentive to split funds among different types of accounts and institutions, because some banks will not be insured at all , while other banks will be insured at a very high limit.

    In any case, if this is what you are advocating, this is a powerful stick to use against banks that continue to game the system to their benefit.

  20. JKH writes:

    The Mosler architecture is as fascinating as it is wild. The plan is to manage risk through both assets (banks as slimmed down utilities) and capital. He doesn’t emphasize the importance of capital, but it’s importance is there implicitly in spades. Reduced asset risk means capital as option volatility is lower, but the moneyness of the option is still sufficiently high that the tail risk for deposits is minimized. Return on capital for banks appropriately would be very low because of that. One of the interesting aspects is that the system deposit base would roughly double in size, as a result of jamming the cumulative government deficit into bank deposits and reserves, displacing current bond placement outside of the banks. Deposit rates would be anchored to a risk free rate structurally set at zero, which as the core pricing parameter is very much consistent with the objective of well in the money capital and very low tail risk for all deposits.

    I have no idea what the optimal configuration is. You can’t redesign the financial system in pieces. All roads lead to specifying the objectives of capital coverage.

  21. winterspeak writes:

    JKH: Which Mosler plan are you talking about? He has a few bits and pieces floating around.

    SRW: Just as I think you know savings (deposits) do not enable loans, I think you know that savings do not fund investment in a way that generates return either. While savings must equal investments, savings do not create investment any more than deposits create loans.

    When banks seek deposits, it is because deposits lower their cost of capital compared to other sources. So the interest earned on your savings account is not a reflection of the yield that an associated investment is enabling, it’s a reflection of how much the bank is willing to pay for that liability (as part of a broader bank business plan). Interest isn’t a reward for taking on risk, it’s more like a salary.

  22. Steve Randy Waldman writes:

    winterspeak — We’re unlikely to agree on he question of whether high value depositors ought to be in the business of monitoring banks and supplying market discipline, but for reasons we’ve covered before. Recall that I think paying claimants more than the risk-free rate for zero-risk investments is corrosive of incentives to perform the information work of investing (which for banks means monitoring strategy and solvency). Further, I think access to risk-free assets should be rationed, because I think carrying wealth forward in time is hard work, while a risk-free asset + a price stability commitment mean that holders of risk-free assets get guaranteed real returns at the expense of others who do the work of making those returns possible. Presuming the price stability constraint is solid, if the economy fails to perform at the risk-free rate by virtue of a supply shortfall (rather than a shortfall in aggregate demand), an individual spending returns on her risk-free investment must provoke taxation of other agents. If the price stability constraint is loose but supply is tight, then the stock of risk-free assets is related to the likelihood of inflation. If the economy underperforms due to a shortfall in demand, the nominal returns on the risk-free asset are antideflationary and helpful in a sense, but they amount to a regressively distributed stimulus which is both less effective and less just than other means of increasing real economic flows. Broadly, I think that people really value and need a certain level of the insurance that risk-free assets provide, and so we should make those assets available, but beyond a certain level the marginal benefit to individuals of the insurance is less than the marginal cost to the economy of blunting incentives to steward real capital well. I think stewarding capital is a burden a very rich man ought not be able to shirk if he wishes to remain very rich. Anyway, we’ve not persuaded one another before on this.

    Re the division between discipline on the asset side and on the liability side, I think we are largely on the same page. I agree that there always needs to be some regulation on both the asset and the liability side. But the degree and instrusiveness of regulation on the asset side is inversely related to the level of equity on the liability side. If there is a dollar of equity for a dollar of deposits (which was intended as an example of very low leverage), the interests of creditors (the state) and equityholders are fairly well aligned, and the only regulation necessary on the asset side is to verify that losses don’t bring equity levels below some threshold beyond which the interests of the two parties materially diverge. (Obviously, there is no fixed threshold — creditors have to choose some place in the slipper slope from aligned to radically divergent interests.) However, if the capital structure starts out at 20:1 deposits:equity, then the interests of equityholders and those of creditors are different right from the start: Equityholders will prefer more volatile to less volatile investments for any expected return, and will even pursue negative expected return assets if the volatility is sufficient. At very high degrees of leverage, the interests of creditors and equityholders are almost directly opposed, so creditors have to entangle themselves in the control of the firm to avoid being expropriated. This is true in general of any kind of firm, but with respect to banks, whose main creditor is the state, entanglement in control means regulation. Anyway, that was overlong, and it’s a story I know you know, the optionality of the equity of a leveraged firm. The point is that the balance between market discipline on the liability side and regulation on the asset side is determined by the degree of leverage. From the state’s perspective, for a bank, both equityholders and all credibly uninsured debtholders (including uninsured depositors) are equity, since those private parties take losses before the state. If the state feels compelled to insure all creditors and there is only a sliver of formal equity, then it must regulate banks extremely intrusively. If the state is able to carve out a class of uninsured depositors, those depositors can protect their interests by threatening to run, and the state can regulate the asset side less intrusively. The key is that the stakeholders junior to the insured depositors must aggressively assert the control rights necessary to protect their own interests. If they fail to do so, because they view themselves as insured or view bad outcomes as too far along a tail, then market discipline will fail. Having uninsured stakeholders occupy much or most of the liability structure is necessary but not sufficient to let market discipline substitute for intrusive regulation. The other piece that is required is to normalize failure, so that uninsured creditors feel like they are in jeopardy and observe with reasonable frequency peers at other institutions bearing losses when they fail to discipline their firms.

    I agree with you that banks and bankers will always lie (and sometimes make themselves believe pretty obvious lies) when it is in their interest to do so. I do think, contra Mosler, market incentives on the liability side will exercise better control (both limiting bad risks and pursuing worthwhile ones) than intrusive but outgunned government regulators. Therefore I think the core task of regulators should be to maintain the preconditions of effective market control: lots of nonstate stakeholders who credibly lack any guarantee and who observe failure frequently enough to actively and consistently strive to avoid it.

    I’m with you on partnerships, on stuff like unlimited or double liability of controlling stakeholders. That fits into this framework. If partners have unlimited liability, their full net worth is effectively equity of the firm (it is at risk and will bear firm losses prior to creditors). Diminishing liability is equivalent to reducing firm leverage, with some added “oomph” from the fact that controlling shareholders have so much of their net worth at risk that they have very strong incentives to exercise prudential control.


  23. csissoko writes:

    SRW: I am still trying to understand your view of marked to market MMMFs and I think where I’m losing you is this: “first-runners get payed full nominal values, regardless of the fire-sale prices their running has forced”. How is it possible for early withdrawers to protect themselves from fire sale prices if the mutual fund is marked on a daily basis and thus sales must be made to honor redemptions on the day of the redemption — and the whole fund will have to be marked to the prices of those redemptions? While there’s no reason to believe the process would be perfectly smooth, I really don’t see how those who run avoid swallowing the cost of fire sales.

    Example: If 25% of the funds are withdrawn on one day and the sale of assets drives prices down by 50% end of day, then what is actually distributed to the 25% who withdraw is 25% of 50% or 12.5% of the fund’s value on the day before the run.

    What’s your understanding of marked to market MMMFs?

  24. Steve Randy Waldman writes:

    winterspeak — Insured depositors don’t bear risk, so by definition, the interest they are paid cannot be a reward for risk. You can call it a salary. I’ll call it a subsidy. Insured depositors aren’t working any harder for their money than they would have if they invested in formal risk-free assets.

    That banks are willing to pay more for customer deposits than the central bank policy rate (or for term deposits the OIS rate) reflects institutional factors that render the elegant account you and I both like to tell about bank liquidity only approximate. Banks cannot costlessly rely on the central bank or interbank market for all their funding. Regulators dislike it, and counterparty banks will add a credit risk premium to the policy interbank rate (or not lend at all) if depository institutions don’t have a deposit base. I don’t know if this is a good thing or not — there’s something to be said, perhaps, for imperfections, sand in the gears. But there’s something to be said for comprehensibility as well, and I’d probably prefer insured deposits and central bank reserves to be costlessly substitutable. Our story is true in aggregate of the banking system, and in practice for most banks individually who do not find their lending constrained by deposit or credit concerns. It’d be better, I think, if we just rationalized the system and normalize bank borrowing of reserves directly from the central bank, subject to capital requirements.

    Uninsured deposits are a different species. As long as they are really, truly, and credibly uninsured, they reduce the state’s exposure to bank losses, and therefore represent a form of private capital. If we rationalized the system and let banks borrow reserves for routine liquidity, then uninsured depositors would be junior to the state and could count as a form of capital. Banks would seek uninsured creditors to loosen the capital constraint on lending rather than the liquidity constraint. Regulators should like uninsured creditors, because by running they can serve as early warning that state interests are in jeopardy, although there’d need to be substantial capital junior yet to uninsured depositors, and again, failure would have to be normal enough that they run early, not late, or else by the time they run the state can be stuck with stiff losses.

    In the current, unrationalized world, uninsured depositors are (i think) pari passu with insured depositors. They represent a form of loss-sharing, rather than a first loss tranche, so they shouldn’t really qualify as regulatory capital. Uninsured depositors should demand higher interest than insured depositors, so to the degree banks don’t find deposits to be scarce (per our stylized, mostly true, description of banks), banks would mostly fail to attract them. Per Indy above, if retail deposits are valueless to banks, banks would offer insufficient advantage for the risk and people would keep their funds in Treasury assets. But, if “core” retail deposits are valuable to banks and interbank reserves are an imperfect substitute, then banks would be willing to pay a yield premium to attract uninsured deposits. I don’t understand well enough the institutional and regulatory factors that compel banks to seek out and pay extravagantly for retail deposits rather than use interbank funds to make good predictions here. But I think that in the real, current world, externally liquidity is necessarily available but its use is stigmatized and invites regulatory action, and banks voluntary pay a higher-than-achievable cost-of-capital in order to avoid that negative attention. So expensive uninsured deposits would be sought.

  25. winterspeak writes:

    SRW: I fully agree with you that the CB should normalize lending to banks directly, and kill the overnight, interbank market. It is a terrible design that does one simple job in a needlessly complicated way. Those retorting that the problem isn’t “too big to fail” but is “too interconnected to fail” should identify the overnight interbank market as being one web of connections, and simply sever them all. If you’re short reserves, you borrow from the Fed, unsecured. End of shadow bank runs.

    This would well and truly put a stake through the heart of liquidity risk. All that is left is solvency risk, and that exists, as always, at the prerogative of regulators.

    Here’s what I cannot understand about your position against unlimited FDIC insurance. There’s your millionaire who has $10M of income that he decides not to spend, so he deposits it in a bank. The bank makes loans, which are investment decisions. Somewhere, you seem to believe that the $10M directs, or should direct, the investment decisions in order to truly earn or deserve its interest rate. But that’s not true.

    The bank can make as many loans whether the millionaire puts his money under his (large) mattress or deposits it at the bank. In my toy model, the millionaire can even torch the money, and the bank’s ability to make loans is unaffected. So the credit decision is entirely distinct from the deposit decision. (I’m assuming the $10M is a drop in the bucket to overall AD btw).

    The $10M deposit decision does impact the bank’s cost of capital, which would impact its profitability, and maybe change the portfolio of loans it chooses to make. The millionaire created room for $10M of investment when he chose not to spend his money. The $10M of investment which actually happens and fills that void is not directed by how the millionaire decides to store (or torch) his $10M pile. Investment is enabled by the “not spending” decision, the “how I store what I decided not to spend” is irrelevant.

    So why do banks want that $10M deposited with them? It lowers their cost of capital, which helps their profitability. They may pursue profitability by hiring smarter, more creative loan agents able to make better credit decisions on the asset side, or by having better customer service so you want to bank with them, or by offering higher interest rates to lower CoC on the liability side. All of these are perfectly legitimate business tactics.

    “I don’t understand well enough the institutional and regulatory factors that compel banks to seek out and pay extravagantly for retail deposits rather than use interbank funds to make good predictions here”

    Is your question now answered? Given the role of retail deposits I have described above, do you still believe that they should not be fully FDIC backed? Do you really want to introduce DD unstable equilibria into the financial system??!

  26. winterspeak writes:

    Sorry. To be clear “having better customer service so you want to bank with them, or by offering higher interest rates” would lower the CoC if it could attract sufficient deposits that the lower CoC generated by the deposit heavy liability side of the balance sheet more than made up for those higher costs.

  27. Steve Randy Waldman writes:

    csissoko — I’ve been thinking about your comments, and I may or may not be mistaken. Here’s the issue as I see it.

    Each evening an MMMF marks its accounts to market and publishes client positions in terms of units representing $1 of value. That evening or the next morning, a client (who conventionally treats her money market as a bank account) requests a “withdrawal”, really a redemption of a certain dollar amount, based on the most recent statement.

    That redemption request represents a kind of liability to the fund. It’s not a net liability — conceptually, the fund replaces an indefinitely rollable claim with an overnight or one-day claim. The question is, in what units are those two claims denominated.

    My thinking, perhaps naively as an ordinary sometimes-user of MMMFs, is that the denomination of my redemption request is in dollars. So, based on last night’s statement which showed a balance of 1000 (dollars? units?), our client requests requests a $100 withdrawal, to be fulfilled this evening. Suppose that the market value of the underlying assets drops precipitously over the afternoon, so that the NAV falls by 10% (extreme, but for the sake of argument). There are three possible ways this could be handled. 1) 100 units are redeemed, so our client only receives $90, not the $100 she thought she had “withdrawn”; 2) $100 are delivered to our client, but 111 units are deducted, representing the current cost in units of that money; or 3) $100 are delivered and 100 units are deducted based on the most recent published statement when the redemption request was made.

    In case (1) or (2), you are absolutely right: the client’s redemption effects only her own assets, and our “runner” takes her full share of the loss. In case (1), she ends up with $90 cash and a remaining position in the fund worth $810. In case (2), she holds $100 cash, and her remaining position is worth $800.

    I was imagining case (3). In this case, our client receives $100 and holds 900 units. Assuming the fund is large (so we can ignore the per-unit loss in NAV from the redemption), those 900 units are worth $810, and she has shifted $10 of losses to the rest of the fund by running earlier. (If she had liquidated entirely, she would have been paid $1000 and would have shifted $100 of losses to other claimants.

    I don’t know which of these three policies obtain. If (1) or (2), you are absolutely right. For an unleveraged fund and ignoring transaction costs, clients who “sweat through” other peoples’ run will eventually see their NAV return when prices normalize or the paper that backs the fund matures (assuming that the fund managers do not themselves run from the assets when things look dire!). If (3), runners shift losses to slower claimants.

    Note that if the fund is leveraged, financially or behaviorally, so that selling at depressed prices by clients triggers more than proportionate selling by managers, clients who stay are harmed by the run. But my claim that liabilities are nominal so losses are shifted rested on my implicitly imagining case (3), which could very well be incorrect.

  28. Steve Randy Waldman writes:

    winterspeak — In a rationalized system, that $10M of credibly uninsured deposits would count as private bank capital. (It would be junior to central bank borrowing and senior to everything else.) It would directly enable the bank to make more loans and take more risk. It would not be the same as cash under a mattress. Uninsured deposits would be senior risk capital that happen to be attached to liquidity, not mere liquidity. Above risk-free interest rates would be payment for risk.

    Putting that aside, I don’t understand how a bank could significantly reduce its cost of capital relative to direct policy-rate borrowing by attracting sizable retail deposits. You argue that clients would accept a lower cost of capital for the convenience of easy access customer service, etc. It’s certainly true that small retail clients offer deposits at an interest rate of zero to run checking accounts. But the cost of maintaining glorified spreadsheets and running a branch is basically fixed. Suppose a bank offers free checking with a minimum $1000 balance (that’s about my current experience for my most expensive checking account). The actual cost to a client for the convenience and liquidity of that account, in terms of interest foregone, is anywhere from $0 (now) to $70 per year. Let’s call the market price for really posh convenient banking with very good customer service $500 per year, 70% of which is profit. On $10M of deposits, $500 is 0.005%, half a basis point. Unless we rely upon saver stupidity, and assuming that the central bank rate and Treasury risk-free rate are similar, the gross savings a bank would realize is 0.5 bps, and net savings would be 0.35 bps. Fundamentally, at a narrow bank the business of providing convenient access to government deposits would cost users of any size order-of-magnitude hundreds per year. That’s just not enough to be material to the cost of capital on millions of dollars.

    In real life, of course, a bank pays, it does not save, relative to the central bank rate to attract retail deposits. If I have $10M that I wish to cautiously save, and I don’t spend nearly that much in a year. Right now I could buy a 6-month CD for 1.067% (average according to; i could do better at a moral hazard bank). A bank that could borrow without stigma from the central bank at its policy rate would at this moment be paying no more (much less) than 0.5%. That 56 basis point spread represents 56K on $10M. So my very posh bank is willing to pay me >$56K for the privilege of holding my deposits. We live in strange times, but this usually holds: banks pay well above the risk-free or expected compounded central-bank (OIS) rate to attract customer term deposits [ ]. I don’t have a good explanation for this beyond institutional and regulatory frictions. I do know that I think there is no sense in which an insured depositor “earns” that 56K. The CD rate spread is a subsidy to the relatively affluent, nothing more and nothing less.

    I don’t see any way, in any world, that large retail deposits would be much cheaper than central bank policy rates (+ swaps to match term structure), when retail depositors always have the option of earning close-to-policy rates by investing directly in Treasury securities. If a 10M saver likes the convenience of the bank next door, she can keep 50K there as petty cash and put the rest into Treasuries.

  29. Steve Randy Waldman writes:

    Rogue — Good point, implicitly, in the ambiguity if trying to describe exactly who is ensuring what under the current regime. It’s really the intersection if banks and people (with people oddly multiple via various tricks) that are ensured.

    I’m proposing that only people be insured: it doesn’t matter what bank you bank with, or how many accounts you have. You are insured up to a certain limit (perhaps over a lifetime, perhaps over some shorter period). My proposal doesn’t give insured depositors any more incentive to discriminate between banks (and therefore penalize poorly run or TBTF banks). Instead it tries to define a class of credibly uninsured depositors who would have reason to discriminate, and who could signal a need for regulatory attention by running, without interfering with ordinary commerce and transactional use of banks.

    Your proposal, insuring banks up to a certain limit without reference to depositors at all, could be used to penalize TBTF banks (or banks that provoke any other negative externality through their practices). It’s a gradualized version of Rajan’s proposal, I think.

    It’s an interesting idea, and I’m all for finding ways to differentially tax TBTF banks, but it has a side effect I dislike. Suppose your bank is only insured up to 90% of its deposits. Suppose there is some rumor that calls the bank’s solvency into question. Each depositor has every incentive to withdraw 100% of their cash, as they stand to lose up to 10% of their balance if the bank is insolvent. (This is like Northern Rock — Britain’s deposit insurance originally was set up to reimburse less than 100%, so depositors just took all their money out when it became perceived as risky.) In theory, after the first 10% of depositors take all their money out, there’d be no incentive for the rest to do so, but I wonder if in practice, given that small retail insured depositors would start the run, people wouldn’t know when or would be psychologically reluctant to stop.

    In theory, this needn’t matter: a central bank can provide indefinite liquidity support to lean against any run. But in practice a visible run by all depositors is devastating in a way that I claim a quiet run by relatively few very affluent investors would not be. I could be wrong, but my thinking is that the contagion associated with sharp, unnecessary runs has something to do with the physical interference of runners with the ordinary activities at the bank, and that small depositors disproportionately provoke that. Large depositors won’t empty the ATMs and jam the branches: they’ll arrange electronic transfers between institutions. (Many small depositors would do the same, but many small depositors would also hit the branches.) By acting less visibly, I think potential runners’ judgments may be more independent, leaving runs longer, slower, and more remediable affairs. Also, the fraction of the balance sheet subject to runs would be smaller than if all depositors are potentially in jeopardy, limiting the pressure to liquidate in distress while potential funders, regulators, and the central bank work out a resolution.

    But again, that’s all speculative. Your proposal is very clear in being able to differentially penalize banks, and that’s a real advantage. If depositors insist upon never investing in banks that are near their FDIC limit, then it would be a very powerful tool that could limit bank size. (But if that’s the result we want, why not just cap deposits?)

  30. Steve Randy Waldman writes:

    JKH — Reading Mosler on banks, I often like many of his specific proposals on the asset side (which mesh quite well with Winterspeak’s). Despite my preference for more market-disciplined banks, at present substantially all bank capital is arguably provided by the state, so the asset side is everything. Under that circumstance, I agree that the sort of restrictions he proposes, like originate-and-hold-to-maturity; no secondary trading or nongovernment securities; no subsidiares, SPEs or organizational complexity of any kind; strict controls on derivative use (no credit derivatives); etc. make a lot of sense.

    When you talk about limiting asset volatility, though, I wonder if these proposals mesh with that. Sure, they’re very restrictive, but they also prevent some kinds of diversification and hedging and leave the bank exposed to idiosyncratic risk. If that’s right (perhaps it isn’t), it leaves the risk to depositors, which is really the risk to the state, fairly high from volatility, unless there is a lot of genuinely private capital junior to depositors. (It would be expensive capital, given the restricted-by-regulation opportunity set.) It isn’t necessarily see a big problem, if we adopt MosleroWinterspeakian heavily restricted banks and let the state guarantee almost everything — the effective subsidy to equityholders might be worth paying if banks’ incentives as far as real investment become genuinely aligned with the public’s interest in high quality, real economy investing. But the stakes become very high if the public ratifies the emerging status quo of fully public liabilities but is unable to adequately exercise control to ensure managers and equityholders don’t too much game things.

    I have to think about the question of how cumulative deficits might affect bank capital structure. It’s not obvious to me that it should: if the banking system generates its own deposits subject to a capital constraint, why should net government spending matter? Government bonds, or central bank reserves become assets. What would affect the liability side? What would render deposits cheap so that they’d displace bonds?

  31. JKH writes:


    Mosler’s proposals all amount to the same thing. And there as much MMT proposals (i.e. Kansas City and Australia) as they are his. They all promote the same thing.

    The banking system retains plain vanilla credit risk. FDIC insurance is extended to everyone for all deposits (that’s one I actually had to point out to Mosler that was implicit in his proposals before he documented it formally).

    Put those two together and you’ve got lower risk weighting than now, but you’ve got to have capital and it has to be fat capital even relative to that risk weighting.

    But by far the most pivotal of all the MMT proposals is that of a “zero natural rate of interest” and everything that flows from that. It’s only on that basis that the Fed becomes a utility and a mere clearinghouse. And it’s on that basis that the Fed loses its independence because it no longer has any policy responsibility. And it’s only on that basis that the Fed can lend unsecured – because the reason it lends secured today is to maintain its policy independence and ranking with respect to its creditor position in FDIC wind ups. And the overarching implication of ZNR policy is that monetary policy in terms of rate setting dies, and monetary policy by other means is subsumed by fiscal policy. The entire proposal is “heroic” in its ambitions for fiscal power.

    The “zero natural rate” has huge implications for another risk area of banking which is that of interest rate risk. Under the assumptions of ZNR implementation, there is no longer any interest risk – because by presumption there’s no longer any risk of future policy rate changes. Unless of course the regime reverts – as in the failure of any fixed rate regime, which is what ZNR would be, even though the MMT’ers haven’t yet acknowledged that. I put the question to Mosler about regime reversion risk and its relationship to interest rate risk, and he didn’t respond.

    The interest rate risk aspect is terribly important because the vast bulk of banking system deposits is time deposits, and because most people conflate interest rate risk with liquidity risk when they are analytically separable. The favourite Austrian concept of “maturity transformation” is such a conflation.

    I probably shouldn’t have interrupted your discussion here. It’s a terribly messy topic because of various dependencies on larger institutional parameters.

    But it’s all capital centric. There isn’t a single aspect of the global crisis that can’t be reduced to the effectiveness of risk analysis and corresponding capital allocation.


    SRW – just noticed your latest; I’ll have a look and respond a bit later

  32. JKH writes:


    I’m glad you brought up the issue of the private/public capital split, which you’ve noted previously.

    I disagree with the idea that substantially all bank capital is provided by the state. Substantially all bank capital insurance is provided by the state. There’s a difference. The purpose of cash capital is to cover likely (implicitly – confidence level specific) capital losses. Capital insurance is to cover extreme tail risk. The fact that capital is provided as a derivative (e.g. guarantees) rather than cash reflects the disproportionate relationship between notional exposure and effective exposure at the tail. That notional guarantees were gargantuan is no more indicative of the effective level of capital provided than is the notional amount of derivatives outstanding indicative of their risk. If one were to market to market conceptually the risk associated with those notional guarantees, the number would be much, much smaller, and considerably smaller than the original private capitalization of the banking system (especially when including ongoing internal capital generation capability). While state capital insurance was essential to the survival of the banking system, that doesn’t translate to substantially all bank capital being provided by the state, in my view. It was the tail insurance that was essential, and once that was in place, the effective risk associated with that tail insurance began to recede. The outcome is that the net cost of TARP et al will be a tiny fraction of original private capitalization levels. The dynamic of tail insurance is an example of Soros’ reflexivity principle, I think. Perhaps I agree essentially with your idea, but would somehow phrase it differently for technical purposes.

    From my comment posted just prior, you can see that I would disagree with your second paragraph. The overall risk profile is reduced due to the elimination of casino type volatility businesses. Substantial credit risk diversification would be achieved via industry differentiation in plain vanilla lending. But credit risk is still credit risk. It is risky, and it is subject to cycles. So it requires capital. And capital levels would be higher than they are now for a comparable subset of risk, in order to justify the free put to all depositors. And because capital levels are high for the amount of risk taken, the cost of capital would be correspondingly relatively low (less leverage). Again, the concerns you raise are handled by insisting on high in the money capital levels for the level of risk taken, notwithstanding that the level of risk taken would be lower than it would be for a previous universal bank structure with casino departments.

    The ZNR structure without bonds works as follows:

    Net government expenditures create new bank deposit liabilities and new bank reserves. The government does not issue bonds to drain those reserves. The reserves remain at the Fed depository. The Fed infrastructure gets rolled up into Treasury. The reserves become the immediate “funding source” for the deficit and just sit there in perpetuity (unless the government starts running surpluses). The government pays zero interest on reserves. The government can mature or buy back outstanding bonds until the entire cumulative deficit that is currently publically floated (about $ 8 trillion) ends up as bank deposit liabilities and bank reserves, instead of bonds. The deposit base of the bank system roughly doubles (unless banks start to issue non-deposit liabilities).

    The ZNR structure merely crystallizes the initial stage of money creation via deficits that takes place in the system as currently structured. The government spends and creates new deposits and reserves. In the current system, those reserves and deposits are drained from the banking system by issuing bonds to the non-bank public. The ZNR structure merely stops at the first step of the current process.

    The reason its zero rate is that the government effectively traps the banking system into being the exclusive conduit for liabilities associated with deficit creation, including its own liability in the form of reserves. Because the government is the monopoly provider of reserves, it sets the price where it wants to. And the MMT prescription is that it should be set at zero, so that all monetary policy is then subsumed by fiscal policy (e.g. easing via deficits; tightening via taxes.) And there is no impact on capital constraints because reserves are zero risk weighted.

    (This has been an explanation; not an endorsement.)

  33. JKH writes:


    Although I say that the state doesn’t provide most the capital, I agree that the tail exposure this time around was obviously unacceptable. Substantial reform is required to lessen that tail risk, including crisper differentiation of intended institutional risk profiles, more conservative risk quantification of those profiles, and higher levels of equity capital for given risk weightings.

  34. Rogue writes:

    A powerful stick all the same. My idea is to threaten to take away deposit insurance (of depositors themelves) who continue to put money in banks who continue to avoid helping solve the crisis, such as doing more modifications where it would actually lessen the possibility of more delinquencies/foreclosures down the line.

    The mere threat that depositors will abandon them in droves, due to lack of FDIC coverage, could coerce any unhelpful banks to bring their full attention towards solving the problem.

  35. winterspeak writes:

    JKH: All good points. I don’t understand all the implications of the MMT zero natural rate of interest policy. For example, suppose the Fed lent and borrowed unsecured at the discount window, at 1%, and shut down the overnight interbank lending market. It also stopped issuing Treasuries, since there was no longer any point. Wouldn’t that get us to the same place, except at 1% and not 0%? I don’t see why 0% is the magic number.

    SRW: In your response, you argue that banks overpay for deposits. Maybe. What I’d like you to clarify is why you think banks need, or want, deposits at all? What purpose do you think deposits actually play in the financial system? I’m not asking what role you think they should play, I’m asking you what you think they actually do play, regardless of whether banks are overpaying or underpaying for that thing.

  36. JKH writes:


    Good question.

    The key is that the government basically traps the “funding” of the deficit within the banking system. From the banks’ perspective, they end up with trillions in excess reserves. (The idea of a “required” reserve would remain to be defined, but it may as well be zero. It doesn’t matter.)

    The only way the banks don’t arbitrage the short rates down to zero is if the central bank artificially intervenes by setting a non-zero rate floor – exactly the way it does now – by setting a non-zero Fed funds target, paying interest on reserves, or by issuing non-zero interest paying bonds where the non-zero market rate is basically anchored by the first two types of rates, or some combination of all of these options. In the ZNR policy, no bonds are issued, so reserves are the only outstanding government liability, other than currency. Therefore, unless the government intervenes with explicit non-zero rate support on reserve pricing, competitive forces among the banks will drive the least risky market rate pretty well down to zero. The risk free rate becomes zero, at least by default, if there are no other government liabilities to trade, other than the trapped reserves, which may or may not be traded in the interbank market (acknowledging that banks can only “lend reserves” amongst themselves, not to non-banks).

    Therefore, absent “artificial” government intervention at a positive rate, the short term risk free rate goes to zero. It’s pretty much the same argument that used now where central banks intervene to support rates, but here the counterfactual is used as the desired outcome, and the fact that no bonds are used puts even more downward pressure on rates. Hence, the “natural rate” is zero. It’s essentially an argument by contradiction, defining any rate other than zero as an artificial interest rate propping exercise by the central bank.

    This puts reserve pricing on par with currency pricing, where no interest is paid. I fear Scott Sumner would have mystifying things to say suggesting the natural rate become negative. I won’t go there. Although pricing arbitrage between reserves and currency would be a come back on that, I guess, if a comeback were deemed warranted or worthwhile.

    I have queried the MMT’ers along the lines of what is so natural about trapping the entire deficit within the banking system, basically eliminating non-bank holdings of government liabilities, other than currency. That is a structural imposition after all. I don’t recall the response.

    It is interesting though that directionally the ZNR policy is sympathetic to the notion that investors shouldn’t be earning an outsized annuity on risk free assets, which is a reason I thought it might appeal in some way to Steve in particular, and perhaps to yourself.

    Finally, note that the zero natural rate as I said earlier is part of a package whereby the central bank gives up its policy rate setting autonomy and where monetary policy becomes surrendered to and embedded within fiscal policy. But that’s why your question is a good one. That approach presumably could be achieved by setting the policy rate at one per cent forever. But then in either case you have the issue of “regime change risk” for the market to figure out. I’m not sure why some future George Soros wouldn’t bet the bank that a future President Mosler or his type might be forced to cave in on a natural policy rate of zero as a fixed rate feature of the monetary system, at some point in the future, in some scenario of extraordinary inflation threat that wasn’t being handled so well by fiscal policy at the time.

    If you’re interested in more from the horses’ mouths, see:

  37. winterspeak writes:

    I see. So if the goal is to shut down the Fed, then you do it with a zero interest rate, because then you don’t even need the small but ongoing intervention to run a 1% interest rate (fixed). Makes sense. Still not sure if 0% is the right risk free rate, it sounds bizarre to me, but it now also sounds bizarre to say that 0% is the WRONG risk free rate too!

    Mosler and SRW have (I think) different issues with the rentier class. In the UK, this group are called pensioners, btw. I think Mosler wants to increase the supply of labor, and thus, real output. I’m not sure about SRW since I remain baffled by what impact he believes the chosen vehicle of saving has.

  38. Steve Randy Waldman writes:

    JKH — Boy there’s a lot here. I’ll start with a bit of a digression, the question of “how much” bank capital is provided by the state.

    The first thing I’d point out is that we agree that notional values of asset guarantees are largely irrelevant to the question. It is not interesting to know that the state has guaranteed 200B of BAC assets or whatnot. (OK, it is a little relevant, but because it is a credible signal of state commitment, not because the values are very meaningful per se.)

    I disagree with you strongly, though, that what the state provides is “tail risk insurance”. I would have agreed with you prior to the crisis, but by revealed preference, governments around the world simply refuse to let any class of claimant above common equity to take other than mark-to-mark losses, and they cushion the losses to common equity as well. To call the de facto insurance provided by the state tail risk insurance, you have to define bank capital structures as 5% bofy, 95% tail. I think you can draw a meaningful distinction between idiosyncratic and correlated risk, and claim that the state bears much less idiosyncratic risk (i.e. it might let an individual medium-sized bank struggle if the rest of the world is healthy), but even there big bank problems are inherently systematic, and the government simply will not let big bank creditors who don’t scare off suffer. The quantity “substantially all” is difficult to define. Obviously, the government formally provides the economic capital represented by the part of the balance sheet called “insured deposits”, but to quantify capital in equally weighted dollar terms is misleading, given the seniority of deposits. To “measure” state capital, we’d have to come up with some number representing the likelihood and degree to which the state will take risky actions that blunt losses to bank claimants up the capital structure. I think we have learned that the likelihood is very high states will protect very junior claimants at the big banks that represent the majority of most countries’ banking systems. I don’t have a number, but I stand by my “substantially all” claim with respect to big banks. (I have an idea for how to come up with a number. Maybe I’ll get off my lazy ass and try it sometime.)

    Just like I don’t think notional guarantee values are very interesting, I don’t think that notional performance of individual programs is interesting either. TARP will perform reasonably well on the books, ex-automakers and AIG. But a substantial portion of TARP repayments involved replacing state funds with private funds at a much higher cost of capital in order to make possible large bonus payouts. In economic terms, those TARP repayments are losses to the state relative to the alternative of more cheaply funding banks and curtailing compensation, because the paid-out version leaves the banks less well capitalized (after compensation and capital costs) than before, while it remains just as necessary for the the state to protect creditors now relying upon a semi-explicit guarantee.

    Further, in order to measure the ex post fiscal cost of the interventions, it is far from sufficient to measure the losses that will be taken from all the various support and guarantee programs. (Unless the Fed, GSEs, FHLB, etc. are very thoroughly audited, we may never know those costs anyway, they can be hidden as offsets to future income.) We also have to count the cost of a wide variety of subsidies, including the steepening of the yield curve, excess reserves on which interest is paid, and less-than-aggressive value protection in deals like unwinding AIG portfolios, purchasing mortgage-backed securities, etc. that have been conducted en masse by state agencies.

    It will be a talking point of both banks and government officials soon (actually, it already is) that “the bail-outs didn’t actually cost very much”. Besides the obvious fallacy of confusing ex post luck with ex ante costs, that will be a lie. The scale of subsidies is, I think, in the hundreds of billions, and the means by which they have been supplied is corrosive to the public trust.

    OK. Enough on that. In sum, I do think the government pretty clearly provides substantially all large bank capital, and that my case is pretty strong. (Even if you disagree with my estimates of ex post costs, it’s hard to dispute that junior unsecured creditors at big banks have been unburdened of most of their risk.)

  39. Steve Randy Waldman writes:

    JKH — Your description of the ZNR revolution is fascinating. Much to think about. I certainly understand now why, in the world described, deficits would drive capital structure, as all deficits would be reserves, and all reserves would be offset by deposits. Add zerp-cost liquidity provision by the central bank and a regulatory scheme that let’s only equity-like securities drive capitalization and balance sheet size, and you have a world in which all bank debt would be zero-interest deposits or borrowed reserves.

    (There would be no capital constraint on deposits backed by reserves, but there would still be capital constraints on balance sheet expansion via loans on the asset side, right? So even though in aggregate reserve expansion would not promote or restrain lending, individual banks would still have to modulate their asset share between reserves and loans, or the size of their overall balance sheets, in order to remain well capitalized, no?)

    I actually think this is a very intriguing proposal. (It’s too pure and radical to be very likely, but still…)

    As you foresaw, I like that it provides a unified default-risk free rate, ending the incentives to wasteful arbitrage among risk-free vehicles. Zero is a pretty good nominal rate because it is easy to defend in a world where central bank reserves are redeemable for physical cash. Just make sure reserves are in excess, and the rate is defended. No balancing act is required. As you point out, monetary policy disappears and is subsumed by fiscal policy. In theory, I like this a lot, because fiscal policy choices are more thoroughly vetted and transparent than monetary policy that hides behind delusions of wise technocracy but has profound allocational and distributional effects. Bank capitalization could be made very clear (let bank balance sheets be restricted to a binary split of common equity and central-bank lending / insured deposits). Trade-offs between liability-side regulation and asset-side regulation could be analyzed pretty easily. We could have different classes of institutions, some highly capitalized and lightly regulated (call then “hedge funds”) and others highly leveraged and intrusively leveraged (call those “banks”). All in all, it’s clean and quite beautiful, at least as a thought experiment. As you say, it is very radical.

    A big, huge, monster question one would have to ask is, would we, collectively, be able to handle any truth consistent with good allocation choices? This comes up throughout the crisis in all kinds of ways, and people like me want to take a clear position based on ethical ideas rather than practical certainty that transparency and accountable choice is good. If macrofinancial choices are made largely via public fiscal policy, rather than a cabbalistic Fed and convoluted, plutocratic banking system, would we make those choices well, or devolve into a system where people are drawn out of the real economy in order to agitate for stimulating themselves and taxing others? My answer to that would be that the range of institutional choices is not one-dimensional, and that we can surely come up with better alternatives to the existing clearly corruptly subsidized banking system that don’t devolve to a public contest for spoils. But the idea of simply telling Congress they can tax and spend as they please to regulate the macroeconomy, without offering some kind of ideology and institutional framework to inform their choices, gives me pause. I don’t like the ideologies and institutions we’ve been operating under, but that doesn’t mean an alternative choice (like naked self-interested horse-trading) would be better. Don’t get me wrong: I think we can come up with better choices. But when I read, say Bill Mitchell (whom I like very much), I think the guidance he gives (spend until full employment or “capacity constraints”, hire all comers to do “useful” work) may be insufficient or realistically unsustainable.

  40. Steve Randy Waldman writes:

    JKH & winterspeak — One last thing, on risk, “casino finance”, and “vanilla lending”: As I’ve been doing a lot of lately, I think we need to be cognizant of the difference between idiosyncratic risk and systematic risk, and different kinds of idiosyncratic risk one can take.

    There is “vanilla lending”. It is inherently risky, but supports real economic projects carefully evaluated as potentially productive and worth the capital. Vanilla lending by a bank exposes it to idiosyncratic risk, the risk of its own projects. But both in evaluating and funding these projects, banks serve their social purpose.

    There are all the accoutrements of “casino finance”, securitization, derivative arrangements, secondary loan markets, etc. Banks can use these tools to speculate, to increase their idiosyncratic risk, and under some circumstances they clearly do so. (Those circumstances are pretty straightforward: bankers use these tools to speculate when they can get away with it in their institutions and the rewards to large successes are much greater than the personal costs of large failures. Banks as institutions speculate in this way when they are heavily leveraged and their creditors fail to exercise sufficient control to prevent them. When banks are heavily leaveraged, “casino” speculation is in the interest of the bank’s well-diversified shareholders.

    But the tools of casino finance also can be, and are, used by banks to hedge and diversify idiosyncratic risk. Derivatives really can help banks manage interest rate risk to credit risk, and can also increase the rate which with they can offer loans. Fundamentally, derivatives allow risk-sharing, which diminishes banks’ exposure to idiosyncratic risk arguably at the cost of increasing systematic risk.

    I’m not interested in increasing the rate. As we Chartalist-sympathizers know, financial capital can be manufactured at will, so we can expand equity as necessary to get the pace of lending we want without offloading the function to mindless quasibanks.

    However, I think we should acknowledge that, especially among small banks, forbidding the tools of “casino finance” will leave them more exposed to their own idiosyncratic risks, and therefore more volatile and likely to fail. That will implies a fiscal cost — banks that would otherwise have hedged an exposure and survived collapse, and we’ll need to pay their insured deposits. However, this fiscal cost is worth bearing, because we want banks to be exposed to their own idiosyncratic risks, because that creates sharper underwriting incentives and avoids contagion of judgment by herding financial markets. (With derivatives, banks underwrite the risks they can most easily hedge, rather than the risks they think best independently.)

    This added fiscal cost to relying solely on vanilla finance will be offset by diminished fiscal cost from banks and bankers using them to escape risk-management by creditors or supervisors and speculate. My guess is that the fiscal savings from avoiding speculative finance will exceed the fiscal cost from forcing banks to hold undiversified idiosyncratic risks. But that’s a judgment, and derivatives-boosters will claim that good risk management can mitigate speculation costs so we shouldn’t throw out the baby with the bathwater. I think it’s important to respond that even if we can magically improve the quality of supervision and stay on top of gamesmanship, so that these tools are only used wisely to hedge, we wouldn’t want banks to use them because banks are in the business of taking informed idiosyncratic risks in the real economy, and we want to preserve their incentives to do that well. The cost of bank failures provoked by independently underwritten projects that turn out to go sour is offset by the benefit of independently underwritten projects that succeed and contribute to the real economy. The cost of bank failures provoked by speculative financial activities is offset by benefits that are indirect and questionable at best.

  41. Steve Randy Waldman writes:

    winterspeak — I suspect that JKH might know the answer to the question of why banks pay for deposits, but here’s my guess.

    I think that liquidity management has become something like an olympic sport, a vestigial competition among banks. Once upon a time, it was the core of their business. A well-managed bank first and foremost had to always have the funds needed to meet its obligations. As banks evolved from independent institutions to participants in a large system that in aggregate supplies liquidity at will, they never ceased to manage liquidity, and the culture that associated “good management” with internally raised and managed liquidity never went away. In fact, not only is it embedded in industry culture, but it is embedded in the regulatory culture that now surrounds banks. Although liquidity management is a useless function for a well-capitalized bank, and deposits aren’t referenda on bank soundness, regulators still judge banks in part by their ability to manage liquidity without overmuch recourse to borrowed reserves, or (God forbid) the discount window. Accepting central bank liquidity provision invites regulatory scrutiny, even though given the nature of the system, paying for overpriced deposits rather than borrowing from the central bank is useless and wasteful. I don’t think it’s very rational, in the world as it is, that banks compete for deposits. The best you can say for it is, just like Olympic success tends to correlate with economic success, perhaps skill at the liquidity management competition correlates with good banking generally. There’s no necessary link, but perhaps.

    Regardless, it is clearly the case that even though liquidity is always ample to the banking system as a whole, and even though individual lending decisions are not coupled to liquidity positions, banks are willing to bear significant costs to ensure that a substantial fraction of their liquidity comes from a retail deposit base. Old habits die hard, especially when they are deeply embedded in business and regulatory culture.

    BTW, I’m not suggesting that it’s irrational for banks individually to pay for liquidity: I’m sure the costs they perceive in terms of added regulatory scrutiny for lacking deposits, or for relying upon volatile liquidity source that might require them to borrow interbank or central bank reserves, are real. It’s irrational collectively that regulatory and cultural incentives should be bound up with insured liquidity. But for individual banks can’t opt out of those incentives.

  42. Steve Randy Waldman writes:

    winterspeak — i think when the chosen vehicle of savings is at risk, people exert effort to ensure those risks are wise or at least discipline by exit. When the chosen vehicle of savings is not at risk, people shirk. I think the information work associated with perpetually reproducing, let alone expanding, our capacity to produce for one another is exceedingly difficult, and that monitoring or contributing to that work is genuinely useful. I think “savers” who do monitor or select investments are not rentiers at all, but are either producers (when they invest well) or insurers (even when they invest poorly, they privatize some of the costs of bad investment).

    i don’t claim, by the way, that there is any meaningful distinction between saving via Treasuries and saving via insured bank CDs, except that savers via bank CDs are, for no useful reason, subsidized more heavily than savers via Treasuries. neither group is contributing to the reproduction of their savings. i want to eliminate that excess subsidy, because it leads to wasteful real practices (elaborate account splitting schemes, markets CDARS, etc.) But I also want to ration access to Treasuries, because one should only be able to save so much without helping along wealth reproduction or bearing some of the costs of its nonreproduction.

  43. winterspeak writes:

    SRW: Wow — competing for deposits as an appendix! I’m glad I asked you to clarify the role they play, I was not on the right track at all.

    Given that you believe that competing for deposits is a useless activity whereby banks give up potential profits to the saver, your hostility to it confuses me even more. I’ll spend time with your other responses, maybe that will make this clearer to me. I really was holding my breath for this one.

    JKH: Do you have thoughts on why banks compete for deposits? My understanding was that it reduced CoC (and I think I know why this happens), and I assumed that the reduction in CoC more than made up for the expense of trying to get deposits through a profit maximization argument.

  44. JKH writes:


    We’ll have to disagree on tail risk insurance. The US banking system held more than $ 1 trillion in equity capital prior to the crisis and generated another $ 250 billion per annum pretax – very roughly speaking. It has pretty much recapitalized itself internally and externally following government preferred injections and paybacks. The corresponding global numbers were larger. Also remember that a good deal of the damage was outside the commercial banks as defined, including Bear, Lehman, Fannie and Freddie, and AIG. I would leave it optional as to whether that’s considered part of the tail risk or not. And I’m not saying it’s a good thing that junior creditors were covered. The fact is that the banking system first loss equity covered most of the direct hit on a net basis. And I’m not saying there weren’t costly externalities borne by the rest of the economy as well, but externalities occur in crises. The government doesn’t end up paying for all things bad, notwithstanding that bad things happen in an unfair world. And not all things bad were directly due to banks. It all needs to be cleaned up with a new risk, capital, and regulatory framework.

    “(There would be no capital constraint on deposits backed by reserves, but there would still be capital constraints on balance sheet expansion via loans on the asset side, right? So even though in aggregate reserve expansion would not promote or restrain lending, individual banks would still have to modulate their asset share between reserves and loans, or the size of their overall balance sheets, in order to remain well capitalized, no?)”

    Yes. There are still capital constraints on the non-reserve portion of bank balance sheets.

    I thought you might like the ZNR policy (BTW I haven’t seen any MMT’ers use that acronym, which I adopted.) It is very elegant in its own way.

    “But the idea of simply telling Congress they can tax and spend as they please to regulate the macroeconomy, without offering some kind of ideology and institutional framework to inform their choices, gives me pause.”


    “But the tools of casino finance also can be, and are, used by banks to hedge and diversify idiosyncratic risk. Derivatives really can help banks manage interest rate risk to credit risk, and can also increase the rate which with they can offer loans. Fundamentally, derivatives allow risk-sharing, which diminishes banks’ exposure to idiosyncratic risk arguably at the cost of increasing systematic risk.”

    You sound a bit like Greenspan there (did anyone ever say that to you before?) – But agreed – I can vouch for how useful and relatively safe interest rate derivatives have been over the years, and they have been reasonably stress tested in actual macro environments. Credit derivatives were more recent developments in the evolutionary history of derivatives creation. Credit cycle risk intensity usually ends up trumping interest rate risk damage. Also, credit derivatives were the domain where investment banking types dove in head first into what had theretofore had been a commercial banking risk expertise, nearly destroying the world as a result. Just like them to do that.

    I’ve said elsewhere that the ZNR idea somewhat intersects with the idea of 100 per cent reserves on certain kinds of deposits – possibly that part of the system balance sheet that is potentially asset-liability “matched” with respect to that very large asset reserve position. But I haven’t thought that through.

    SRW and Winterspeak – I think SRW has it pretty well on why banks compete for deposits. Also, this may sound trite, but banks compete for deposits because they’re available. At a macro level, the notion that loans create deposits extends to the notion that assets create (liabilities plus equity). The macro balance sheet is always in balance. And the central bank always (well nearly always) ensures adequate system reserves to meet that balance without system borrowing. So banks compete for deposits because they’re available (because they’ve created them). The system treated as a whole is a pretty much a monopoly supplier of its own deposits, so as a whole it can price them consistent with some target return on equity, via their effect on interest margins. And then individual banks can deviate from that pricing at the margin, depending on their marginal strategies. Borrowing from the central bank is simply frowned upon in normal times. It’s intended to be a periodic option, not a continuously used resource. It’s not supposed to be part of the primary purpose of banking. The idea is to stay out of the central bank, not be at the trough all the time. so in normal times, the central bank imposes a very strong moral suasion whip to calibrate discount window utilization. Remember that it only reversed the direction of this moral suasion during the financial crisis. So banks are not in the business of borrowing from the central bank, but they are in the business of competing for the deposits in the same way they are in the business of competing for the loans that create the deposits – Yin and yang. Or asset liability management, as it is known in banking. And liquidity management as part of asset liability management just fine tunes the process by controlling exposure to cash flow mismatches. Retail deposit gathering is a profit center by the way. And retail deposits are a cheaper source of funds than wholesale deposits. Banks typically price retail deposits as a rate discounted off the wholesale curve. That way they get a spread on both liability businesses and asset businesses. Banks with a strong retail base are considered to have a lower cost, more stable source of funding from the private sector, and their cost of capital will reflect that favorable impression.

  45. winterspeak writes:

    SRW: I want to distinguish between the decision to “not spend” and the decision to “save”. After all, you can not spend some of your income and set fire to the money instead. This is not “saving” in any sense, yet the amount of investment enabled is the same. “Not spending” is the action, post receiving income, that enables investment. Putting the money in the bank is one option you have after the not spending decision is taken, as is putting it in a vault, or burning it in your back garden. But all those decision to do impact the quantity of investment that initial “not spending” action enabled, nor which specific investment opportunities are pursued.

    I think it is useful to give people the option of “not spending” without forcing them into an investment decision. When I say all bank liability holders should be on the hook for the quality of a bank except depositors, it’s because equity holders and bond holders have made an actual investment or credit decision, so yes, some stewardship is called for. But the depositor is a different beast, he is not directing capital in any way, and so let’s be honest about what’s happening and engineer sanity into the solution. Unlimited FDIC. It stops the lie that is the money market fund. Also, SRW, it improves the quality of real investment decisions in the economy as it screens out noise (I assert. I’m just trying to convince you.)

    I am fine with depositors having zero interest, actually, and that is pretty much their reality today. If the US goes down Japan’s path, it may be the reality for quite some time to come. Does this make you more comfortable with unlimited FDIC? Suppose I underline how unlimited FDIC makes ZNR work better?

    It kills me to read about how the Govt has “made money” on its TARP “investments” because the Govt doesn’t need any frickin’ money. It’s a goddamn currency issuer, so it should grow a pair and do what a frickin’ sovereign is supposed to do. Rule. Obama’s SOTU described the bank bailouts as a root canal, but that isn’t the case, since after a root canal, your roots are canaled. We have had all the pain, but the financial system is exactly the same as it was. The dentists drilled holes in your jaw, but your molars are still there, festering. WTF.

    Therefore, I give even less credence to whatever money the Govt did or did not make on its TARP operations. The output I want from them is good Governance, not “profit” and I haven’t seen any of that yet. Actually, if the Govt made money it means it continues to drain net financial assets from the private sector, and that is bad.

    JKH and SRW: Wow — both of you see competition for deposits as vestigial. I guess it must be true then. Certainly the Fed frowning on discount window use is vestigial, as is Academic Macroeconomics, so why not this too?

    JKH: You say “Banks with a strong retail base are considered to have a lower cost, more stable source of funding from the private sector, and their cost of capital will reflect that favorable impression.” OK, but is this entirely a matter of belief, or is it true?

  46. JKH writes:


    It’s true, due to spread pricing and pricing that is less interest rate sensitive compared to wholesale deposits.

  47. winterspeak writes:

    JKH: Thanks. Does the lower CoC then make up for the cost of competing for deposits? Or are banks pursuing a non-profit maximizing strategy for anachronistic reasons? Or does the anachronistic Fed attitude towards the discount window force a non-profit maximizing strategy on the banks?

    It is ironic that part of “too interconnected to fail” are interconnections forced by the Govt itself.

  48. Steve Randy Waldman writes:

    JKH & winterspeak — First thanks, both of you, for a characteristically fantastic conversation.

    While my wife was wondering why I was so spacey when we were out for dinner, I was thinking of deposits. Here’s another way of thinking about it: Perhaps deposits are coveted precisely because depositors impose no discipline, and therefore provide banks with strategic options that other financing sources do not. As JKH put it, retail deposits are viewed as a “more stable funding source” than wholesale deposits or other sorts of debt. Why might that be? First and foremost, retail deposits are insured, so they stay put when other creditors might try to exit. They impose no covenants as other creditors might. (Do banks enter debt agreements with covenants?) Their willingness to rollover their advances is a statistical property of their own collective lives, largely uncorrelated with the circumstance of the bank. Retail depositors are less price sensitive than other sources, because there are economies of scale to shopping and moving deposits that they simply can’t exploit. The costs outweigh the interest they stand to gain.

    Creditor discipline is like competition, something business people love in theory, especially for the other guy, but that they do everything to avoid themselves. And that’s perfectly rational: creditor discipline restricts the options of managers and equityholders. A bank that can sweat through and earn its way out if a few projects drive it to what market participants might view as negative equity is more valuable than a bank whose creditors protect their interests by intervening and shutting it down or wresting control from equity. From shareholders and managers perspective, a less disciplined bank is simply a more valuable bank. The indulgent financier is really the government, but retail depositors vote with their deposits which banks will get the gentle finance.

    Of course the “indulgent financier” examines and regulates banks, but the strictness of the regulatory process is not related to the proportion of a bank’s assets supplied by retail depositors (and to the degree there is any sensitivity, deposit-funded banks are favored rather than hyperscrutinized). So the incentives to acquire deposits aren’t blunted by the regulatory process. Effectively, every dollar of retail deposits comes with a valuable strategic option of creditor forbearance. The same is true with insured wholesale deposits, who don’t monitor for soundness, but with wholesale depositors banks have to pay for the option, since that group is price sensitive. Effectively the state writes an option to retain finance despite potential solvency concerns. Retail depositors have no pricing power, so that option is free to banks and this group is very valuable. Wholesale (and high-value retail) depositors have to be wooed with high rates, so wholesale depositors earn much of the premium on the option supplied by the state. (Effectively the spread between Treasury rates and CD rates is a transfer from the state to price-sensitive depositors: that state writes an option, the depositor gets paid for it.)

    So it makes perfect sense that banks would compete for deposits, and especially for small retail deposits, which are cheap in absolute dollars and confer the strongest options. It’s not exactly clear that any useful social purpose is served by this kind of competition (although it does create incentives for banks to offer pleasant service to individuals of modest means). There are obvious pathologies that can result from this dynamic, c.f. weak banks drawing depositors with heavily marketed high yields to buy time to gamble for redemption.

    It still is all vestigial, from the perspective of a system where reserves are borrowable on demand to any bank regulators deem solvent at the policy rate. In such a system, the state agrees to write depositors’ indulgence option to all comers for a preset low price. But as JKH points out, we’re not quite there yet: central banks typically discourage discount window borrowing, and they require collateral that retail depositors do not. So deposits are valuable.

    It makes sense that retail deposits could reduce the cost of capital to banks even relative to the policy rate, as the marginal cost of new customers is low, and small customers demand almost nothing in interest on their transactional balances. But it remains true, I think, that to attract large and wholesale depositors banks have to offer a premium over same-term Treasury rates.

    So to offer an answer to Winterspeak before JKH’s more informed one, I’d suggest that small retail depositors can reduce banks cost of capital in absolute dollars, and especially when the stability option is taken into account. Large and wholesale depositors can’t reduce the cost of capital relative to the central bank policy rate, but they can offer valuable options, and the central banks policy rate is really not on offer except with steep hidden costs.

  49. Steve Randy Waldman writes:

    winterspeak — there’s a difference between burning a dollar and saving it in a bank: burning it forecloses ones claim to the produce of current investments. that means, ceteris paribus, others have more incentive to perform the work of production if one burns the dollar rather than saving it.

    At the level of a dollar, of course, that connection is so indirect as to be meaningless. At the level of the economy as a whole, if making good on desired purchases by past savers would pressure on the price level, then people will legitimately expect taxation or inflation, blunting current incentives to produce. Further, if those past savings balances were generated to easily, without sufficient attention to savers’ contributions to the production process, the transfer represented by the moblation of savings is arguably unjust.

    Of course, there’s another side of the case: if people do not know they’ll be able to store the wealth they save, they’ll have little incentive to produce beyond momentary needs. And if people who have no ability or inclination to do so are forced to contribute to the investment process, that may, as you suggest, add noise that is harmful to the process of reproducing or expanding real wealth.

    So, I think we need to carefully balance things. I do think the state should try to offer a guaranteed price-stable asset. (Price stability might mean slow decay — what matters is committed predictability and a good enough deal that producers consider it worth their while to produce and save. Note that historically, with the exception of the 1970s, the US government has offered pretty complete price stability for savers who earn short-term Treasury rates. That’s the price-stable risk-free asset, not paper in your pocket.)

    But I think that beyond some level of savings, not asking savers to take responsibility for the reproduction of their wealth creates too much risk for current producers, and ex post produces unjust transfers. So I think we need to find a balance: the state should offer access to risk-free assets up to a certain level of wealth, but beyond that, savers should be responsible for carrying their wealth forward.

    That doesn’t mean every high net worth saver has to become Warren Buffet. Even the state’s risk-free investment is risky, because the commitment to price stability must be broken. A risk-averse high net worth investor can provide overcollateralized secured financing on very safe terms, or lend to Berkshire Hathaway. High net worth investors can pay to hire managers, and diversify among them. Of course, even these low risk investments could tank in a systemic crisis that leaves the real economy less productive than anticipated. We want people to invest carefully to help avoid such crises. But if such crises do occur, and there is a real supply shortfall, it would be unfair if risk-free savers lost no purchasing power while current producers bear more than their share of loss (via unemployment, reduced wages, and reduced returns to real capital, if the state succeeds at stabilizing nominal prices). When there is a shortfall of supply (rather than an aggregate demand malfunction), the state has a much harder time keeping its price stability commitment, so you can argue that risk-free savers will be hit by inflation to share the pain. Maybe so, but past risk-free savings makes the price stability commitment harder to keep, so the overstock of risk-free claims helps cause the injury it suffers, but doesn’t suffer the injury alone. Also if price stability will fail when supplies are tight, that reduces the cost of restricting access to the risk free asset, since rational savers should want to hedge that with some claims on real assets.

  50. winterspeak writes:

    SRW: That was my point. Once the “not spending” decision is made, burning that dollar or burying it in the ground does not impact the quantity or quality of investment at all (I’m assuming you aren’t impacting demand materially to address your objection). The “not spender” creates space for investment by forgoing consumption, but they do not direct it.

    This is a good thing, because you want investors and entrepreneurs to direct investment. But they need space in which to do so, and that space is created by “not spending”.

    SRW: “But I think that beyond some level of savings, not asking savers to take responsibility for the reproduction of their wealth creates too much risk for current producers, and ex post produces unjust transfers.”

    In Gods name why? They don’t have the power to do this. Inflation (or deflation) is a product of aggregate demand, which is governed by fiscal policy (managing vertical money) and financial regulation (managing horizontal money). Not-spenders, even at very high levels of not-spending, just cannot move these numbers. You’re saying that the guy with the hammer should also take responsibility for turning the screws, even though he is ill equipped for it and there’s a guy with a screw driver sitting right there. Is this a moral position with you, Steve? Because it isn’t sound engineering.

    And we don’t have a real supply shortfall, we have a real demand shortfall. Maybe that is what you meant. Either way, rectifying the shortfall is entirely within the realm of Government, but they choose not to do it for a variety of reasons.

    Also, and this may be straw that breaks the camel’s back, I don’t think price stability should be the top goal of a Government. It’s certainly nice to have, but other things are more important.

    You are correct, btw. If you want to protect your wealth moving forward, then you will need to buy real assets. At that moment you cease to become a “non spender” of course, and the onus is on you to pick wisely.

  51. Steve Randy Waldman writes:

    winterspeak — I guess i’ve not communicated well; you’ve misunderstood me.

    I think that burying the dollar does adversely impact the quality and quantity of investment relative to burning it, because aggregate “buried dollars” build claims that diminish the upside of investing relative to saving, and therefore dissuades others from making real investment. When there is a shortfall of aggregate demand, I think “not spending” in general does no social good whatsoever. Obviously it doesn’t “make room” for real investment (since resources are underutilized), and it hides information about preferences that even consumption spending would contribute to the production process. Rewarding “not spending” during underutilization with claims in future wealth is in a sense perverse; it pays people to hide their preferences, and leads either to loss of economic capacity or (if the Keynesians and MMTers have their day) spending by governments to offset the saving that is less well-informed than individual saving would have been In other words, today’s “savers” are arguably causing social harm. We should not reward that with the most certain claims to future wealth we have on offer.

    [I want to qualify that by saying today’s savers arguably cause social harm to the degree that they save more than is necessary to adequately insure a reasonable level of their & their family’s own future consumption. The social harm of letting oneself get homeless is far worse than the social harm of failing to inform the market. But savers with large balances do no social good by not spending or investing. The retain a claim to the future for nothing or worse than nothing in the present.]

    It’s certainly true that right now there is a shortfall of aggregate demand, not pressure on the price level, so those savings balances aren’t putting pressure on real resources at this moment. But, unfortunately, those savings balances are nonperishable, while real wealth and capacity decay without continual real investment. By overbuilding financial savings now, we foreseeably create the circumstance in which investors will see returns damaged by taxation or inflation, depending on which outcome the state views as most destructive. That is, at this very moment, harming the quality and quantity of aggregate investment. There are lots of people like me, a former tech investor holding claims to gold and sitting out the sort of analysis I used to do, because we think that claims to default-risk-free financial assets have been allocated in ways so unmatched to future production that a time will come where giving up on the price level or taxation will be the only choices.

    I agree with you that, in that decision, taxation is not always the right choice. In fact, I think inflation is the only reasonable choice now: too many dollars have been allocated to people who harmed rather than helped ongoing real capacity, often by persuading people to take on debt, so that a transfer from creditors (of the state and of individual debtors) to debtors is the least harmful way out. But that does damage to money: The value of fiat currency may in the end be supportable by compulsory taxation at threat of imprisonment or violence, but in a decent society, those concerns are very far removed: a money is stable because nothing shakes a coordination Nash equilibrium. Inflations do shake that equilibrium, and limit the capacity for the state to manage aggregate demand with its patterns of and net quantity of spending.

    The point is that it is not enough for the state to manage present aggregate demand. The state also must ensure the real economy and accumulated deficit are such that it will also be able to manage future aggregate demand without destructive or politically implausible taxation or a large inflation. I think the state has already failed to do that, and we will face the consequences. And I think we should understand why: because we were insufficiently attentive to matching allocation of financial savings to successful reproduction of real wealth.

  52. JKH writes:

    SRW, winterspeak,

    I generally agree with SRW # 48 above. Retail deposits result in higher net interest margins than do wholesale deposits. From those fatter deposit margins, the bank must deduct additional operating costs associated with retail deposit gathering. After that, it gets messy in accounting for retail deposit gathering infrastructure costs – e.g. the additional capital cost sunk into branch real estate, or lease costs. And there are joint infrastructure synergies and economies for loan and deposit businesses considered together. All such costs are typically well covered by typical retail deposit margins and that it ends up being a profitable source of funds relative to the wholesale alternative. Regarding stability, the effect of deposit insurance subsidization is one factor. But also consider portfolio management behaviour of the typical wholesale depositor versus the typical retail depositor. Simplifying, wholesale money is “hot money” and retail money is “cool money”. This translates to higher sensitivity of wholesale money movement as a function of day to day basis point changes in money market interest rates. And it translates to lower sensitivities in the retail market. Moreover, banks set retail rates generally in lagged, jump fashion with respect to corresponding movements in wholesale rates of the same term structure, and they move asset and liability retail rates in conjunction. Looking across a commercial bank balance sheet, the entire asset liability rate structure tends to ratchet up and down like a step function while wholesale rates demonstrate continuous volatility around trends. The result is that for a retail bank with a reasonably matched book, interest margin behaviour tends to be more stable that it is for a wholesale funded bank – more stable and fatter. In other words, commercial banks with both loan and deposit retail operations are to a degree hedged against the vagaries of daily money market rates, certainly more so than with pure wholesale funding operations. Such stability reduces net interest margin risk, which is worth something in terms of the (risk adjusted) cost of capital.

    Part of Winterspeak’s question assumes a premise where I might disagree with both of you. This is the aspect of “vestigiality” or “anachronisticity” (Hello, Sting). I noted earlier the generally proposed MMT architecture, which Mosler makes specific but that is also accepted and promoted by the others. One of those architectural features is this idea that the Fed should be lending unsecured in unlimited quantities, and that the interbank market should be eliminated. This seems to be one of Winterspeak’s favourite features. Central bank borrowing capability of course is essential to the MMT idea that bank lending is not “reserve constrained”.

    My impression, perhaps wrongly, is that you have interpreted this feature in a way in which it is not intended. While unlimited and unsecured central bank borrowing is intended to be an option, it is NOT the case that it follows that central bank borrowing is intended to be an ongoing viable alternative to deposit gathering. And the status quo is not an anachronism in that specific sense.

    Consider the banking system where as usual loans create deposits. Then consider any amount of unsecured lending by the Fed. That lending creates new reserves. Since Fed borrowing is matched by an equal increase in reserves, there is no net effect on the rest of the banking system balance sheet. The deposits that were there prior to borrowing from the Fed are still there. It’s just a question of where they are; i.e. how they are distributed. In no way has central bank borrowing substituted for deposit gathering at the macro level. So when you ask the question as to why do banks compete for deposits, again I answer by saying it’s because they’re available, and now to which I would add it’s because they’re unavoidable when they’re available in this way at the macro level. Any individual bank that goes to the discount window is essentially foregoing deposits available elsewhere in the system that could have precluded the necessity of discount window borrowing. The supply of deposits as a result of discount window borrowing therefore has been distributed among fewer banks than without discount window borrowing. And that means that the rest of the system can price them more cheaply, other things equal.

    There is only one way in which central bank borrowing can become a substitute for deposit gathering. It is a very extreme case, not assumed in the MMT architecture proposal at all, and not a terribly productive case to consider on its own. It is the case where there is a massive run from bank deposits into central bank currency. Apart from the fact that this is very unlikely, due to deposit insurance, it’s worthwhile looking at the macro balance sheet effect that results just the same. Assume the extreme case where all depositors run into currency. As banks pay for currency “purchased” from the central bank with reserves, persistent net currency demand will run down aggregate system reserves into a deficit position at some point, which is untenable. The first central bank response would be to attempt to replenish reserves by putting more deposits into the system with asset acquisitions of various types. But if nobody wanted new bank deposits, newly issued central bank cheques would simply be presented for newly issued currency, and the process would continue feeding on itself. So the central bank would finally be forced to lend money to the banks directly in order to facilitate massive currency issuance. The end result of this limiting case is that the central bank would provide all of the funding for the banking system apart from capital. The central bank effectively becomes a conduit for funding commercial bank credit with currency issued to the public. That’s a very roundabout way of getting deposit insurance for the public. It isn’t necessary. And the only reason I illustrate it is that it’s the only way in which one can interpret the MMT prescription of unlimited and unsecured lending as a substitute for deposits. And even then, such lending is forced by systemic deposit hemorrhaging dictated by depositors, rather than the banks’ own efforts to replace deposit funding with central bank funding. So the MMT prescription includes a viable role for bank deposits. The purpose of unlimited, unsecured lending is to resolve reserve and deposit distribution anomalies in such a way that any overflow credit risk at the origin of the trouble is handled by an integrated central bank/FDIC/treasury government infrastructure for such risk absorption when necessary. It’s simply a mechanical arrangement to avoid the current balkanization of risk management functionality within the entire government sector. It’s not an invitation for banks to start funding themselves from the central bank rather than with deposits. So given that, the idea of deposit banking functionality is no anachronism, and there is no intention to change it in the various MMT proposals.

  53. winterspeak writes:

    “aggregate “buried dollars” build claims that diminish the upside of investing relative to saving, and therefore dissuades others from making real investment.”

    How? By building up “inflationary forces”, a deep lake of dollars held back by a dam, ready to gush forth and drown all value in a sea of hyperinflation?

    I was dead wrong making assumptions before, so please tell me how you think this happens.

    “When there is a shortfall of aggregate demand, I think “not spending” in general does no social good whatsoever. Obviously it doesn’t “make room” for real investment (since resources are underutilized), and it hides information about preferences that even consumption spending would contribute to the production process.”

    The villain is the shortfall of aggregate demand, not the “not spending”. Don’t blame the rain for getting you wet if you refuse to open your umbrella. A high not-spending rate enables more Govt spending AND lower taxes. You’ll get 50% of the electorate to cheer either option. And an increase in “not spending” REVEALS a preference — a preference for not-spending.

    It is also obvious that, in this environment, not-spending has made room for real investment, and that real investment is inventory. It’s not a “good” investment, but it is as real as a mountain of containers in Long Beach.

    “But savers with large balances do no social good by not spending or investing.”

    How can you say this? Savers with large balances create room for investing (which is good) and enable more Govt spending (which can be good) and lower taxes (which can be good) without triggering inflation (also a good thing). The less “not spending” you have, the less “investment” you can have. And you want investment, and you want that investment to be good!

  54. winterspeak writes:

    JKH: The people who argue against “to big to fail” and for “too interconnected to fail” have a point (not that I don’t believe size alone isn’t a problem). Lehman was essentially kicked out of the interbank market when counterparty credit risk “infected” it. This is clearly one destabilizing point of interconnection, and the Fed has, by relaxing requirements so dramatically at the discount window, brought us closer to a system of uncollateralized lending even if we aren’t there completely yet. This, along with relaxed capital requirements (via TARP), and payments on reserves (bypassing the traditional rate setting mechanism entirely) was what restored “market function”.

    I think the MMT point here is that we have a traditional interbank mechanism, which is brittle, and which gets replaced by a more central CB role when things get tough. It’s like not having FDIC, except when there’s a bank run and you instigate FDIC to stop the run. When the run ends, you take FDIC away again. Does that make any sense? What’s the Fed going to do? Go back to its old collateral requirements until the next shadow bank run, and then relax them again after another firm goes under?

  55. winterspeak writes:

    Sorry, I re-read your post, and I think we actually agree.

  56. Steve Randy Waldman writes:

    JKH & winterspeak — Before I go back to slugging it out with Winterspeak over the economic value of risk-free savings, let’s take a moment to observe we’re all pretty much in agreement on the role of deposits. Kumbaya, people!

    Addressing JKH’s post a bit, what would be “vestigial” under a system with freely borrowable reserves at the policy rate isn’t deposits, but competition for deposits (and there only for large deposits). Assuming (catastrophically!) that I don’t get my way and all savers have the option of holding government obligations that pay approximately the policy rate directly, price-sensitive deposits and government obligations become perfect substitutes to banks. Banks still create deposits by lending, and hold them where they have them, but they address any imbalance between the liquidity offered by deposits and their loan plans by borrowing from central banks (or lending to them — holding reserves). The deposits are still out there, and banks still hold them, but they don’t compete for them because: 1) large depositors won’t accept less than the policy rate, since they can hold government obligations; 2) banks won’t offer more than the policy rate, since they borrow from the central banks uncollateralized and without rollover risk, just like deposits. So the CD rate will be no more or less than the policy/government obligation rate (for any given term), and banks would be indifferent to financing via loans or deposits (assuming there is no regulatory or industry stigma attached to CB borrowing).

    The Mad Max case, where all private borrowers flee from holding deposits and therefore all bank finance is from central banks, is I think best subsumed by the label “total collapse of the currency and the value of financial assets.” (Since, presumably, deposits remain redeemable for cash, so people would only behave this way if they think cash will be worthless and wish to spend their deposits, or if they think the government will fail to fulfill insurance guarantees. In either case, that’s a guns, gas ‘n go-go girls kind of scenario.)

    The competition for small, retail depositors is not vestigial, since as JKH points out, they offer capital at sub-Treasury rates even after the costs of branch and service overhead is accounted, because it is impractical for small depositors to insist upon earning the full risk-free rate when they require transactional balances. Even when banks can borrow freely, without stigma, without collateral, and without fear of termination from the central bank, they’d prefer to borrow from retail customers who are willing to do without interest payments on their aggregate float.

    Now back to our previously scheduled swaldman/winterspeak slugfest…

  57. Steve Randy Waldman writes:

    Winterspeak —

    “aggregate “buried dollars” build claims that diminish the upside of investing relative to saving, and therefore dissuades others from making real investment.”

    How? By building up “inflationary forces”, a deep lake of dollars held back by a dam, ready to gush forth and drown all value in a sea of hyperinflation?

    No, I don’t image magic “inflationary forces” or crumbling dams beneath the national money hole. The mechanism by which burying dollars alters the behavior of third parties relative to burning dollars is simple: it is public knowledge that the dollars are buried, not burnt, and that the buried dollars can be retrieved by their erstwhile owners at any time.

    It is the same as the difference between taxation and savings. At time zero, right now, there is no difference. That other dollar is unspent either way, my purchasing power right now is no worse, and perhaps better, than it would have been had it been spent. For present-tense consumers, there is no difference between a dollar burnt and a dollar buried, a dollar taxed or a dollar lent to the state.

    However, for people considering spending money to endow future consumption, say by purchasing a factory, there is a huge difference. If trillions of dollars have been buried, I have to wonder whether a year from now, when my factory starts producing and selling widgets, those savers will suddenly unbury and try to mobilize those dollars. If they had burnt the dollars I wouldn’t have to worry about this. But since they have buried them, I do. If the quantity of buried dollars is large relative to the size of the economy (how large is an important question!), then their attempts to remobilize of dollars will put pressure on the price level. I’m actually fine with that, as a factory owner, if I can be certain that there’ll be an inflation. After all, I’ll hold valuable real goods to trade or not for their depreciating dollars. But, there’s that government with its price stability commitment. If it takes that seriously, than it will have to tax somebody in order to sustain the value of those suddenly-too-plentiful dollars, and in aggregate that somebody has to be me. Because if those dollars are remobilized in quantity faster than the economy has grown, the only way to maintain price stability is to tax those who hold or produce real goods and services and (indirectly, intermediated by money) transfer those goods and services to those trying to make use of their savings. Foreseeing that possibility will reduce my willingness to build a factory, and increase my willingness either to consume today, or to become a risk-free saver myself. And the more people risk-free save, the greater the incentive for the next guy to do the same rather than build a factory. Obviusly, this becomes a tragedy of the commons, where everyone’s individual incentive is to hold money, but the aggregate effect is a collapsed real economy and also the state’s ability to live up to its purchasing power commitment.

    Again: There’s no magic. If people credibly claim to have burnt their money when they have in fact buried it, so I think that the stock of buried money is zero, then I’ll invest just the same as I would have if the money had really been burnt. But when that money is suddenly unburied and the state comes to tax me, I’ll know I was a sucker, and be pretty unlikely to make forward looking investments the next time. (The real world analogy would be the Fed lying about the deficit. Chartalists like to use words like “deficit terrorists” and such. If you really think there’s no difference between “burning” and “burying”, you should be in favor of obscuring or lying or perhaps simply not keeping or publishing accounts on cumulative deficits, because they are of now present concern whatsoever. Do you believe that?)

    The villain is the shortfall of aggregate demand, not the “not spending”. Don’t blame the rain for getting you wet if you refuse to open your umbrella. A high not-spending rate enables more Govt spending AND lower taxes. You’ll get 50% of the electorate to cheer either option.

    Ah, but you presume a chartalist utopia, and then you gloss notions of quality.

    Let’s suppose the chartalist utopia part. When there is a shortfall of private demand owing to an increase in the risk-free savings rate, hooray!, the government steps in and makes up the shortfall. What do they buy? Mitchell & Mosler want them to buy labor at $8/hr to do, well, what exactly? Clean the streets? Build highways? Invent transporter beams?

    For the present, sustaining aggregate demand at preexisting comfortable capacity is sufficient to eliminate macroeconomic pain. So, the chartalist / vulgar Keynesian notion of stimulus by any means, purchasing jobs, whatever, works fine for a while. But real economic capacity depends on more than just the quantity of expenditure, or the quantity of real activity. It also depends upon the quality of expenditure, the pattern of real activity. No matter how many people are hired by the government, if the tractor factory is dismantled and sold for parts, there will be a problem should private demand for tractors stage a comeback. There might then come to be a problem about food.

    It’s true, but not sufficient, to say that public investment can substitute for private investment. One has to talk about the quality of the marginal investment, of the next investment a private actor would make vs the next investment Congress will make. Those are actually hard questions, because of course what Congress will do is not predetermined and susceptible to a compicated process that sometimes yields good and sometimes poor results. And there is no single marginal private investor — there are a whole lot of people indifferent between spending an additional dollar or holding it as money, and each of those people would do different things if they spent.

    There are two crucial thresholds though. The first is that, however the funds are spent, the quality of investment must be sufficient to at least sustain existing capacity, given how cranky people become if you try to deprive them en masse of the standard of living to which they have grown accustomed. Secondly and more subtly, the new investment, public or private, has to yield returns that at least cover the expected increase in money flows due to the growing stock of outstanding financial savings that might be mobilized. When thinking about all this, you have to be mindful of the fact that, if there is a shortfall in aggregate demand, it probably means people are unusually risk-averse, so there will be a predictable remobilization of savings if you succeed at persuading people that the future is noncatastrophic. Of course, you don’t have to get it perfectly right: you can always tax to fine-tune aggregate demand. But if you get it badly wrong, if you let the stock of money savings expand dramatically but let the productivity of real investment (relative to diverse human demands) stagnate or fall, you’ll find yourself in a situation where the government will have to tax substantially to regulate demand or else accommodate a drop in the purchasing power of savings.

    There are a lot of moving parts here. We don’t know to what degree the flow of remobilized money relates to financial asset savings (“wealth effects”), and how events like demographic changes or rumors of inflation might affect that. If flows from savings are insensitive to these things entirely, than burning and burying amount to the same thing. But I think it is silly to think that. The size of flows out of financial assets (demand to exchange financial assets for real goods and services) is likely to be strongly related to the stock of goods and services (except in the near complete absence credit restraint and debt aversion). So it matters how much we bury rather than burn (both for incentives ex ante and outcomes ex post), and it also matters how well we invest relative to the scale of those flows.

    I’m not dogmatic about where good investment comes from. I don’t think that it’s always true that the private sector invests better than the public sector, so I think that there are circumstanced when you are absolutely right that it’s a huge benefit, when people choose to hold government scrip rather than spend, because that enables productive public investment whose returns (in a real rather than financial sense) will more than cover the increase in quantity and volatility of flows provoked by a larger stock of financial savings. That might even be true now. It certainly would be true if our government were a little bit less stupid, and focused on carefully purchasing future public capacity (and offsetting depreciation of existing public capacity) rather than crassly “buying jobs”. By capacity, I’m not trying to be vague. I mean infrastucture, parks, hospitals, research, potentially energy infrastucture, etc. But it’s worth pointing out that the vagueness of the word calls attention to a real problem: there are lots of choices, potentially good, potentially awful. A government has to be competent to invest, to make those choices well. But a good government certainly might be. (Nothing necessarily makes private CEOs smarter about real investment than public officials. You can argue, reasonably, that private investors have their own money at risk, are subject to market discipline, and are less exposed to rent-seeking noise than public officials. That might be true of small businesses, but for large company CEOs, the short-term differences between public and private are more a matter of degree than of kind.)

    Alternatively (I think both you and I advocate this), the public sector could simply transfer funds to private sector agents it deems more likely than government officials to make good investment plans. That might take the form of a payroll tax suspension, or sectoral subsidies (perhaps the government is good at identifying broad needs but bad at managing specific projects), broad transfers (public dividends or tax cuts), or other clever means. But the constraints are the same. The government, when their is a shortfall of aggregate demand (and even when there’s not) must work to ensure that the productivity of aggregate investment is sufficient that the volatile flows made possible by the stock of financial savings doesn’t unduly put pressure on either the price level or the political system.

    And an increase in “not spending” REVEALS a preference — a preference for not-spending.

    It is also obvious that, in this environment, not-spending has made room for real investment, and that real investment is inventory. It’s not a “good” investment, but it is as real as a mountain of containers in Long Beach.

    This is all misguided. “Not spending” reveals a preference for not spending relative to a government supplied alternative of holding purchasing-power-guaranteed money. It supplies very little information to government or private investors about what should actually be done in the real world. Trying to increase satisfaction of that preference would involve governments strengthening the purchasing power guarantee, perhaps trying to offer high real returns on default risk-free investment. Retiree-voters might gain utility, for a little while, from such an exercise, but it would work to reduce and undermine incentives of potential entrepreneurs and investors to mobilize real assets rather than hold money. Accommodating savers’ preference for strong financial assets without attending to incentives to adroitly mobilize real resources is precisely the strategy that has lately brought us low. Prior to 2008, we had financial assets of all sorts offering high real returns, very comfortable savings, and our real economy devoted to building houses and shopping malls. Let’s not go back there.

    Inventory may be investment in an accounting sense, but it is only investment in a real sense if it is matched to future demand. Plus, the result of not-spending is not primarily stored inventory, but human attention that passes with time and is not recoverable if not efficiently used in each moment. Broadly speaking economic wealth is not storable: it is much more like electricity (that must be used in the moment it’s generated), than like furniture (which can be stored in warehouses). [Yes, electricity can be stored in batteries, at great cost in efficiency. I hope we solve that problem.] Funds not-spent doesn’t so much yield stuff not consumed, but human activity not performed. And the nonperformance of human activity decreases rather than increases capacity for future human activity, as humans are creatures of learning, habit, and practice.

    Savers with large balances create room for investing (which is good) and enable more Govt spending (which can be good) and lower taxes (which can be good) without triggering inflation (also a good thing). The less “not spending” you have, the less “investment” you can have. And you want investment, and you want that investment to be good!

    Also misguided. All real investment (other than cottage industry DIY) is spending on real goods and services. You cannot have a lot of investment if you have a whole lot of “not spending”. One person’s not spending can make room for another person’s spending, and if that other person spends more productively than the first person, that’s a good arrangement (and they can slit the surplus). This is the basis for all financial investing. But financial investing that’s not coupled to real spending by some party does nothing more or less than shift claims around. And financial investing that’s attached to nonproductive real spending helps no one other than the poorly supervised manager who may have enjoyed an ice cream cone.

    The difference between the government and other entrepreneurs is, that by fact of strong convention and its ability to tax in a pinch, it can issue securities for the real goods and services it wants without much altering the value of its securities. But it remains true that whether that will be good in aggregate depends upon how well it deploys those real goods and services that it harnesses.

  58. Social comments and analytics for this post…

    This post was mentioned on Twitter by MarkThoma: Insure depositors, not banks – interfluidity

  59. JKH writes:


    The current Fed requirement for collateral is probably a function of its current bureaucratic independence. It doesn’t want to go to the FDIC to get paid. Consolidating everything and eliminating that bureaucratic independence would accommodate unsecured lending because it all becomes part of the same bureaucratic risk continuum. And it also eliminates the Fed’s policy independence. So working it the other way, eliminating the Fed’s policy independence, and even eliminating any policy function at all as per Chartalism, is probably the prerequisite for eliminating collateralized lending.


    I see your point on wholesale. It does assume perfect operational efficiency and arbitrage, which is near enough to the truth I guess. One wonders how wholesale depositors make a rational decision on where to put their money, though, if there’s no competition.

  60. dave writes:

    I could be interpreting wrong, but I think in the last paragraph you reference from WS he’s referring to “investing” as a kind of spending, spending on capital rather then consumptive goods. Really that doesn’t challenge anything in the paragraph to follow, but I think you’ve assumed a conflict of ideas where perhaps it is merely a conflict of language.

    The difference between the government and other entrepreneurs is, that by fact of strong convention and its ability to tax in a pinch, it can issue securities for the real goods and services it wants without much altering the value of its securities. But it remains true that whether that will be good in aggregate depends upon how well it deploys those real goods and services that it harnesses.

    And that, in a pinch, is what worries me. Ultimately, a private actor that issues too much debt without showing some sign of using it productively has its credit downgraded and finds it harder to obtain new credit. This is the markets way of signaling both the the entrepreneur and those thinking of lending him capital that current spending decisions are not of good quality. When a government issues a lot of debt the buyers don’t care how its spent, they are buying the governments power to tax. So there is no market price signal to the government as to whether or not what they are doing with the resources obtained through sovereign debt issuance is of good quality.

  61. winterspeak writes:

    SRW: The equivelent to burning not-spent money is actually paying Federal taxes. All Federal taxation is money desctruction. All Federal spending is money creation. A lot of my income is burnt, I don’t have much say over that.

    You feel that the quantity of money stockpiled away increases inflation risk. It does not. Inflation is caused by transactions bidding up the price of a good, which can happen through credit expansion, or just velocity, with “base money” staying exactly the same. The famous Californian strawberry pickers who were buying $700,000 homes were certainly causing inflation, but not because their overfull bank accounts burst. I would ask you to think about Japan. According to your theory, it’s sure to start hyperinflating any time now. In truth, a constant stock of money can generate any quantity of inflation or deflation. The stock of money bears no connection to broad price changes.

    I never advocated that Government spending be used to makeup for the shortfall in AD. We agree that this is best done via the private sector, and at present tax rates, it could be done by simply confiscating and burning less of our income.

    The Govt tunes aggregate demand all the time through its fiscal policy. Monetary policy doesn’t really work. The Govt does not realize it has this tool, nor how it works, nor that it is using it. This does not fill me with hope.

    SRW: “Accommodating savers’ preference for strong financial assets without attending to incentives to adroitly mobilize real resources is precisely the strategy that has lately brought us low. Prior to 2008, we had financial assets of all sorts offering high real returns, very comfortable savings, and our real economy devoted to building houses and shopping malls. Let’s not go back there.” You and I see this period differently. The economy has been short savings (net financial assets) since the Rubin/Clinton surpluses of the 1990s. The problem wasn’t too much savings, it was bad credit decisions. But yes, let’s not go back there.

    Fundamentally Steve, you do not see the mechanics for how not-spending enables investment. Instead, you think it thwarts it, which is why you see not-spenders robbing the country of productive capacity. We’ve been over a bunch of times, and I have failed in explaining it to you. There was this great post on it on some random blog that I have now lost. If JKH has made it this far, maybe he can read my mind and dredge it up.

    So let me wrap up this way. If you change your mind and see how not-spending enables, but does not direct, investment, you will lay down your arms against not-spenders, seeing them as the benign actors that they are, and focus your fierce intelligence on improving the quality of investment. There is much work that needs to be done there, and I look forward to you devoting yourself to that task. I’ll be right by your side ; ) But until that moment you will remain stuck making “banks loan out deposit” based arguments, and running regressions to calculate the money multiplier.

    Let’s tally up the score and see how gloriously far apart we are!

    w: the stock of money has no impact on price levels
    srw: an increasing stock of money will increase price levels

    w: not-spending enables investment
    srw: not-spending kills investment

    w: Govt tunes AD through fiscal policy now
    srw: in a chartalist world, the Govt would have to tune AD through fiscal policy

    w: not-spending reveals a preference for saving. sate that preference
    srw: not-spending tells you nothing about what should be done in the real world. (sating that demand, apparently, not counting as a real world activity. Sorry — kidding, couldn’t help myself).

    w: housing bubble caused by bad credit decisions (enabled by bad regulatory environment)
    srw: housing bubble caused by savings glut (which forced/enabled bad credit decisions)

    both: private sector can likely do a better job of allocating resources than Govt

    both: some forms of investment (like unexpected inventory) are bad, even though they still count as investment

    Damn. Looking through the list I sound like a looney.

  62. winterspeak writes:

    JKH: Excellent point re: collateral. Because that’s basically the argument for doing it — you have the FDIC on the other side anyway that they’re part of the same Govt.

    Dave: No. The Govt does not borrow in order to spend. It borrows to control interest rates. It spends in order that we may fulfill our tax obligations, and save. Read the Mandatory Readings at Warren Mosler’s site — you can find it via Google. Govt debt and private debt are quite different.

  63. Steve Randy Waldman writes:

    dave — i think you’re right that some of the distance here is rhetorical. but some of it isn’t. winterspeak has helpfully catalogued some of the differences, but i think we’re actually closer than his tally suggests.

    you and i, and winterspeak too, share similar concerns, i think, about the quality of government investment, and the lack of mechanisms likely to ensure that it is of good quality. (i share those concerns about much private investment as well — i think many large firms, in the short term, behave very much like governments. like for governments, a crisis eventually forces those firms to put their affairs in order one way or another. eventually comes faster for large firms than for governments, but it can still take some time.)

    winterspeak — first of all sorry for the tone of the previous comment. i wrote very quickly, and as a matter of time-sanity i avoid proofing and editing comments (i know, it shows), but soon after i hit submit i regretted stuff. i don’t think any good purpose is served by calling my conversational sparring partners “misguided”. i do enjoy our sparrings.

    you are right, if you could persuade me that the mere act of non-spending enables investment, it would alter my views. i do see that not-spending can enable investment when the economy’s resources are pretty completely mobilized, and i do see how not-spending-but-delegating-spending to someone who is better capable of investing enables and can improve the quality of investing. but i do not see how not-spending when the economy is suffering from a lack of aggregate demand and market information is helped by the mere act of not savings. i do see that, with a government that “opened the umbrella” and sustained aggregate demand, the resources of the economy might always be completely mobilized, in which case not-spending does make room for investing. whether i think not-spending is a social good in this last context depends on the quality of the behavior of the umbrella opener: whatever it is the state does to zetz aggregate demand, is it genuinely productive and well matched to future wants?

    we agree that taxing-is-burning and saving-money-is-burying, i think. we do disagree on the ramifications of burning vs burying: riffing on an old insight of JKH’s, I view tax as akin to government equity and money as akin to government debt. if the government taxes away money, it can always choose to transfer it back if things work out well (in all the various ways governments transfer, explicitly or via tax cuts conditional on a good economy or via corrupt hidden subsidies). So it’s like equity: the government can issue dividends when the its tax base and the economy that supports it is healthy, but can retain past taxes (only in the sense of failing to return them, the government doesn’t meaningfully “own” money) in a bad state of the world. Savings, in the form of money or bonds, is risky because it creates a government commitment to deliver on its purchasing power commitment on a schedule the government cannot control, leaving it open to a potential hard choice between inflation and austerity. Taxes preserve government flexibility in both the rate, timing, and distribution of its transfers of purchasing power. (Of course that’s problematic because the state is untrustworthy. But using debt-financing in hopes that it will discipline the state and usher in an era of good government has been tried and found wanting.)

    I want to make clear that I absolutely don’t think high deficits, or a high cumulative stock of public debt, mean certain inflation. I think chartalists are usually right to call austerity hogs deficit terrorists, and my practical views are close to yours: the government should net-spend to fund durable public investments, and where politics or informational limitations make those hard to fund, it should delegate that funding to private sector agents who are likely to make good investments, or who at least consume in a way that is representative of broad consumption patterns and therefore inform potential private investors about the reshaping of consumer demand.

    I think the question of whether they’ll be inflation can’t be answered by looking at deficits or debt alone. One has to project income and assets as well. Deficit spending is entirely consistent with price stability, however scary big the numbers, if the spending is done in a way that will produce a sufficient surplus of real goods and services that people will remain willing to exchange that surplus of money balances at the conventional price level. Just as a firm can increase the value of its stock by issuing new shares, if will use the funds raised for a high-value project, a government can increase the value of its securities, by issuing more and using the real goods and services mobilized in exchange well. During an economic crisis, even deficit spending with relatively low absolute returns may be justified, because the returns on that spending are high relative to an alternative in which real economic resources would decay. (Of course those “resources” include human workers, and people are much more than economic inputs.) However, the likelihood that deficit spending will eventually lead to inflation depends in part on the absolute returns to public expenditure (in real, not financial terms). Actually, the likelihood that deficit spending will trouble price stability depends ultimately on the productivity of the whole economy, including both public expenditures and private production, as well as the relative weight people place on consumption vs insurance. (If people are “thrifty” and sacrifice consumption easily for insurance, there’s unlikely to be pressure on the price level. If they are confident and consumption-oriented there will be.) Once we “hit the barrier”, as I think Bill Mitchell sometimes puts it, then the stock of private financial savings might matter very much, as it becomes fodder for a bidding war, money for goods and services in short supply.

    Anyway, this is all my way of saying 1) it’s complicated, deficits don’t force inflation and aren’t necessarily bad; but 2) productivity matters, and if aggregate investment is insufficiently productive that increases the likelihood of an inflation, and once an inflation comes, money stocks matter. A government can always prevent (2) from biting via taxation, but there are political constraints too.

    I also think it’s important to think in terms of volatility, not just deterministic flows. To the degree that there are systematic factors that affect people’s preferences with respect to holding money or using real resources (risk aversion, demographic factors, famine or oil crisis, changing laws or norms about retirement support, etc), there will be endogenous variation in the desired net inflow of the private sector into financial assets. If that net inflow wiggles so much it turns negative (and the government doesn’t foresee the shortfall to use taxes to prevent it, or for political reasons fails to prevent the shortfall), there will be inflation. The volatility in flows will be proportionate to the net stock of financial obligations. So even when the expected path of the economy is clearly noninflationary (because the economy is generally productive and there is continual net demand for new money-savings), that expectation of price stability is more fragile, more vulnerable, when the stock of savings that sits beneath it is large.

    Anyway, I’m gonna riff on your summary:

    w: the stock of money has no impact on price levels
    srw: an increasing stock of money will increase price levels

    srw-revised: an increasing stock of money will not necessarily increase the price level, and may be indefinitely sustainable. but an increasing in the stock of money, especially if it is unmatched by increasing real aggregate return, will leave an economy more vulnerable to inflation if returns are subpar or there is a supply shock.

    w: not-spending enables investment
    srw: not-spending kills investment

    srw-revised: per Dave above, i’m not sure to what degree our disagreement is semantic here. when you say “not-spending enables investment”, do you mean some peoples’ not spending can enable other peoples’ spending on investment? if so than i agree, but think it’s important to pay attention to the quality of the spending (and its link to the nonspending). if this isn’t what you mean, then we don’t agree, and i’ll need more help to understand your position.

    w: Govt tunes AD through fiscal policy now
    srw: in a chartalist world, the Govt would have to tune AD through fiscal policy

    oh, i don’t think we disagree at all here. government definitely tunes AD through fiscal policy, it just does so less consciously and well than it might if it had a reasonable framework to make sense of what it’s doing. we probably disagree in that i think that monetary policy is not entirely impotent: i think that both the type and pattern of claims participates in the tuning of AD, not simply the net stock, so exchanges of short for long securities ad net-zero transfers can have very significant impacts on aggregate demand. (i think the direction of monetary policy, the conventional wisdom that high interest rates are contractionary when a chartalist could make the case both ways, is distributional — the holders of financial assets who receive interest payments generally have a low marginal propensity to consume, so the contractionary effect of an increased hurdle rate on investments is not offset by an increase in the net stock of money, because that money is saved, not mobilized.)

    w: not-spending reveals a preference for saving. sate that preference
    srw: not-spending tells you nothing about what should be done in the real world. (sating that demand, apparently, not counting as a real world activity. Sorry — kidding, couldn’t help myself).

    mostly i won’t contest this one.

    w: housing bubble caused by bad credit decisions (enabled by bad regulatory environment)
    srw: housing bubble caused by savings glut (which forced/enabled bad credit decisions)

    i don’t give primacy to the “savings glut” as the cause of the housing bubble, although i do think it was one of the factors that promoted it. i’d include the savings glut, bad credit decisions, historical factors that made housing-based lending particularly easy to securitize, an ideology and culture that turned homeownership into public virtue and a marker of seriousness, lots of public policy subsidizing and promoting housing, etc, etc as contributory to the housing bubble.

    both: private sector can likely do a better job of allocating resources than Govt

    i’d agree often, but not always or necessarily with this. i think that there are public goods that only the government can reasonably purchase that we are underpurchasing, and i’d be very excited to see the government make these public-sector buys. but, consistent with your thesis, politics usually prevents good choices and promoted dumb ones, and for non-public-goods and direct production, i’d much rather have private-sector allocation. so i mostly agree, but think we need to improve our capacity to select public goods and invest in those, because the private sector simply can’t substitute.

    both: some forms of investment (like unexpected inventory) are bad, even though they still count as investment

    yeah. i think accounting distinctions between consumption and investment are largely meaningless. lots of what’s billed as “consumption” is essential to maintaining and expanding activity, and much of what is called “investment” is waste or accounting entries unconnected to real-world activity.

    Damn. Looking through the list I sound like a looney.

    hey, i sound like an asshole, so don’t take it too hard…

  64. JKH writes:


    “If JKH has made it this far, maybe he can read my mind”

    I’ve only just reviewed the main event here between you and SRW, having been engrossed in my own unimportant piece.

    I can’t think of that reference you’re looking for. You guys are probably finished here, but I may try a comment regarding your discussion sometime in the next few days. It strikes me as something worthwhile doing from the perspective of my own understanding, even if you both move on to other things. There seems to be a fairly specific communication misunderstanding going on regarding “not-spending enables investment”, and I may be able to read your mind a bit on that.

  65. dave writes:


    I could once again be making assumptions, but I think your disagreement with srw relates to political constraints as well as private sector economic calculation. If a government taxes someone and “burns” the money the private actor doesn’t expect anything in the future. The money is simply gone. This has two important implications. The first is political. Money paid in taxes doesn’t represent a future claim to someone, so when that doesn’t materialize they aren’t angry. Someone who buys a government bond expects to be compensated with purchasing power in the future. And while it is true the government can go back on this promise by either taxing future income or devaluing this is extremely difficult politically. Moreover, I think spacing out taxation and inflation over time rather then having it all up front makes it difficult for the electorate to understand the implication of political proposals. For instance, we didn’t pay up front for the Iraq war in taxes. If the electorate had been forced to pay up front for the war in taxes it might have been less popular and not happened. However, since the cost is spread out over time it became politically possible to go to war. Using inflation and future taxation (the results of debt issuance) to obscure costs hurts our ability to govern.

    Secondly, I think when you provide someone with “savings” that you don’t back up with real productive capacity then you mess up their ability to understand outstanding productive capacity. Savings is not a carnal desire to be “sated” with paper promises, but a way in which people and societies try to plan for the future. Take as a simple example the demographic shift of the baby boomers. Soon our dependent to worker ration will be much higher. In order to maintain our standard of living each worker will need to be more productive. The only way that is possible is if they are more productive, because they have more capital. Saving is the way in which baby boomers plan for this, because they believe their savings represents and productive capital necessary for the younger generation to provide the resources they desire. If you “sate” their demand for savings with worthless paper that they only find out is worthless when they go to spend it in retirement then your asking for a real train wreck down the road.

    People worried about the deficit aren’t “terrorists” looking to destroy aggregate demand. They are people that understand you need to plan for the future by making smart choices today, and they don’t see wars in the middle east of billion dollar gift baskets to bankers as a sound retirement strategy for the nation.

  66. JKH writes:
  67. winterspeak writes:

    SRW: Yes, I think we are done here. The mechanism by which not-spending creates the space for investment is based on how all production in a period equals real income, and this output can either be consumed in that period, or not consumed (so it is available for the next period). Nominally, this is accounted for by spending your income, or not spending it. In real terms, this left over output shows up as capital stock or inventory. JKH may be able to do a better job of this.

    JKH: I hope you can do a better job of this.

    Dave: No. You, like SRW, haven’t fully embraced what true fiat currency means, or can enable. If you can figure out why SRW’s comment is silly, you’ll be a step closer though. “if the government taxes away money, it can always choose to transfer it back if things work out well (in all the various ways governments transfer, explicitly or via tax cuts conditional on a good economy or via corrupt hidden subsidies).” It’s like in the matrix, there is no spoon.

    Good luck!

  68. Steve Randy Waldman writes:

    Just a note in defense of my silly sentence.

    if the government taxes away money, it can always choose to transfer it back if things work out well (in all the various ways governments transfer, explicitly or via tax cuts conditional on a good economy or via corrupt hidden subsidies).

    In context here, I was not trying to suggest that if the government hadn’t have taxed, it would not have the money to transfer, that would be stupid, and out of context, that’s what it sounds like I’m suggesting. The chartalists are right that taxation does not enable future spending, money is a token.

    I was comparing the government’s situation after taxing to its situation after borrowing. If a government taxes to regulate aggregate demand or shape incentives, it can make transfers to if it pleases. If a government borrows to regulate aggregate demand, it commits to transfers on a timeframe and to parties it does not control or choose.

    I don’t think that difference is silly at all, I think it is quite important. Winterspeak, do you disagree that this distinction is real? If you agree that it is real, why do you think it is “silly”?

  69. Steve Randy Waldman writes:

    JKH — I read the Krugman post before yours, and naturally I thought of you.

    BTW, I’m a Canadaophile generally: I lived for several months in Toronto, had a long distance relationship for some time that kept me traveling there. It’s one of my favorite cities anywhere. I love to travel in Canada, and find the culture and politics of the country fascinating. I still retain a BMO bank account, absurdly, but as a connection.

    Reading through the FT piece, it seems to me the explanations of the banking system resiliency are threefold: 1) cultural — Canadians are really nice, principled, “still bankers”, and even in boardrooms consider whether a product would be appropriate for your mother-in-law; 2) Simple, old-fashioned prudential regulation, high ratios of vanilla Tier One capital, a regulator willing and able to say “no” when bankers get excited and start lobbying for deregulations and mergers; and 3) a nice, profitable, not-outrageously-competitive franchise that serves both keep insiders content with their principled, boring, prudential lot, and as a buffer that permits bad times to hit retained earning rather than settled equity.

    As we’ve been through before, I’ve expressed skepticism of using Canada as a model for bank regulations. I think perhaps I am overly unwilling to acknowledge is success in Canada, but perhaps that is because I fear its ability to work in the US, and find the idea of a misbehaving but well protected banking oligopoly frightening. The US can’t easily import Canada’s cultural assets, although in the medium term I think we will need to find ways of changing the culture of Wall Street or else nothing will really work. I don’t really know how to evaluate why Canadian regulators have been able to avoid capture when, as clearly sometimes happens, the bankers have got excited about the global next thing. Are there institutional checks we can import, that would be universal? is it a cultural thing? was it an accident of history that might not have lasted? Finally, protecting bank profitability as a buffer against both misbehavior and hard times makes a certain sense, but it’s a hard sell. Again the question is trust: if you create a structurally protected oligopoly, how much protected profit is enough? Will regulators be able to resist pressure to permit (or coordinate) a degree of collusion that is predatory or unfair (however slippery those terms)? Will the public get a fair deal in the bargain, trading an implicit tax for a stable banking system in which bankers take seriously their role as credit allocators and all the hazard for third parties that entails? I wouldn’t be willing to jump on that bargain here in the US, given our history, culture and politics. But I’d seriously consider it if I could be reasonably certain the bargain would be kept, rather than undermined and exploited as I’ve come to expect.

    A final note: Canada’s banking system certainly has weathered the storm well. It’s proven much less fragile than others. But there is still a question of valuations. A lot of people (Mish comes to mind, but he’s not alone) suggest that Canadian banks have lent into a housing bubble that hasn’t popped, that things like rent/income are outrageous, especially in places like Vancouver. Do you think that’s true? Quite apart from bank stability, do you think the Canadian banking system may have erred in letting a ratcheting upwards of valuation by comparables overwhelm more cash-flow oriented valuations in lending decisions? Entirely separately (I’m more interested in the first than the politically charged second question), do you think there are implications for banking stability going forward to an arguably still-overpriced housing stock?

  70. winterspeak writes:

    SRW: Just as Govts don’t need to tax in order to spend, Govts don’t need to borrow to spend. Debt is a govt liability, just like the money used to purchase it. So Govt borrowing changes the term structure of outstanding liabilities and thus interest rates, it does not change the quantity of outstanding liabilities. It is not like private borrowing at all. It has no impact on AD, or more precisely, the only impact it has is through whatever fiscal effect is generated. There’s a reason I said monetary policy was impotent, not just now, but as a rule. QE and ZIRP have obviously failed, and have been failing in Japan for 20 years. While this fact does not prove me right, it should certainly give Monetarists (I’m not accusing you of being one of them btw.) a good deal more pause that it has.

  71. Steve Randy Waldman writes:

    winterspeak — OK. Some things we can agree on: traditional monetary policy or government borrowing of money do not change the net stock of financial assets in the private sector. We both agree that the net stock of financial assets can have an impact on aggregate demand, to wit if the net stock is too small, individual agents may find financial savings to be scarce and counterproductively (in an aggregate sense) restrict economic activity in order to try to build stocks.

    Some things we don’t agree upon: That the net stock is materially the only input of government finance on aggregate demand. I think that term structure and especially distribution matter as well. For example, I think that the marginal value of financial savings is, beyond a certain point, decreasing to savers, so that the same net stock of financial savings can result in greater aggregate demand if distributed broadly rather than narrowly. (If distributed narrowly, many agents making the spend/save decision will be savings starved, and save. If distributed broadly, those same agents value the next dollar of savings somewhat less, increasing the likelihood of spending.)

    I also think that term structure matters: I think all agents value short-term assets more than long-term, and fear long term liabilities less than short-term, so that by increasing decreasing the maturity of the private sector’s net financial asset position and increasing the maturity of the private sector’s liability position one can increase aggregate demand. For the purpose of this analysis, I include the banking system within the public rather than the private sector. That is, I think the term structure of the nonbank net financial position matters. But it’s jointly determined: central banks cannot reduce the term structure of assets and increase the term structure of liabilities in the nonbank private sector, unless both private agents banks deem creditworthy seek to borrow short funds against long-term liabilities. It’s not the fact that term structure doesn’t matter that kills QE or ZIRP: it’s that in some states of the economy, the central bank is powerless to impact the term structure.

    We agree, I think, that term structure games won’t work now, when many agents are seeking to improve their net financial position. This is a distributional effect. Even though in aggregate, term structure changes don’t affect the net financial stock, for individual agents, accepting long liabilities for short assets is usually a prologue to spending the short assets, which harms that agent’s individual financial position. Since agents are jealous of their financial savings individually right now, they don’t wish to go down the maturity transformation road.

    Do you agree that there’s an important difference between when a government borrows to spend (not necessarily by issuing bonds — the government borrows simply by spending and crediting a private sector reserve account offset by a public sector overdraft) and when a government taxes simultaneously with the expenditure? You do, I know, in the present tense, because the spend-and-tax operation doesn’t affect the net stock of financial assets, while the spend-without-tax increases net private sector stocks. So you see that as stimulative.

    But where we differ is that I see offsetting costs and contractionary effects to spending-without-taxing: in the form of potentially larger and more volatile endogenous flows within the private sector that may put pressure in the price level (and therefore force taxation at a time when it is macroeconomically harmful) and diminished forward-looking investment by private sector agents who view the potential for taxation due to unpredictable flows as increasing the risk of their projects.

    This is the sense in which I claim borrowing is different than taxing (besides the effect in the net private sector financial position, about which we agree). Taxing and spending’s effect are all present tense. Borrowing and spending has stimulative present tense effects but can add to uncertainty about the future (rationally, not due to some irrational deficit-terrorist ideology). The magnitudes of these two effects, and how much they are worth caring about, is stuff one can reasonably argue about. The chartalist position, I think, is that the future is a series of presents, and the state can always find a way to make the present work out. I don’t think that’s right; I think there are cul-de-sacs. The “deficit terrorist” position is that all present spending is like ice cream, good for a moment but then gone and all you are left with is increased flow uncertainty. That’s wrong too, because if deficit spending is well managed, it can increase the economic surplus which diminishes the cost of the volatility of savings. I think the correct position is in the middle, questions of judgments about costs and benefits in which the quality of fiscal spending matters very much, and the size of the deficit is neither irrelevant nor an overwhelming concern.

    (BTW, I hate coming off as some kind of moderate. I am much more comfortable adopting some armchair quasirevolutionary schtick. But that’s actually what I think!)

  72. JKH writes:


    I don’t have my fingers that close to detailed Canadian housing data. Vancouver has been a problem in its own class for a long time and is more like a US city in terms of real estate history. It has the ocean and the Asian demand influence as well. Toronto has been a hot spot too. I’d describe the profile of housing prices generally over the past 18 months as resembling a dampened stock market profile – a noticeable plunge with a symmetric V recovery, except it’s a 100 per cent recovery. Somewhere recently I’ve seen a good comparative graph on inflation adjusted prices globally, and Canada was both comfortably dampened and lagged relative to the US, Europe, Australia, etc.

    David Rosenberg, who is now back in Toronto, has the same view as Mish, and thinks Canadian prices are overvalued by 30 per cent or so.

    I don’t have a strong view, but admit I’ve been numbed by the persistence of the strength since 9/11 and doubly shaking my head with the bounce back over the past year. It feels like capitulation might in terms of bubble denial. I really can’t see it in the bank numbers yet, though. Current default rates are still quite low as noted in the article. The employment situation is nothing like the US. What would account for such a lag, and why would it take so long to pop? It seems to me the only thing left is interest rates, and I can’t see rates moving high enough and quickly enough to cause major problems. I could be wrong on that, but I still see any interest rate adjustment taking hold over a fairly long time. I could definitely see prices flattening out for a multi-year period, but not a crash. Underwriting standards are pretty strong relative to cash flow testing. Again, I don’t know where a US comparable employment problem is going to come from to make it a huge Canadian banking issue. And I find it almost impossible to visualize a banking crisis due to residential real estate, just thinking about the headwinds and global correlation of some of the corporate disasters the Canadian system has endured over the years (including commercial real estate unfortunately.)

    Here are a couple of multi-part videos from FT you may want to check out briefly – the head regulator, Julie Dickson, and the Bank of Canada governor, Mark Carney. I’ve only looked at the first clip of 3 on Dickson. It gives some colour on culture. So far, Carney has not been strung up for his Goldman Sachs pedigree. Looking through the obvious swagger, though, I’ve been very impressed by his grasp of the public issues since he took over. He did spend a number of years in Canada’s finance department following Goldman.

    The first of the Dickson clips (Carney is on the right hand side scroll):

    Finally, here’s an interesting dividend history from Bank of Nova Scotia, which has paid them continuously since 1833. I found the depression era stage intriguing. BNS is a true global commercial bank – very rare – and is staffed at the top by long term commercial bank guys who really know what they’re doing, and critically have good cultural control over their investment banking arm – a very well run bank all things considered.,1608,CID7144_LIDen,00.html

  73. winterspeak writes:

    SRW: The non-Govt’s desire to net not-spend is literally the sum of millions of individual decisions. Many things can influence this, but I also believe that all the factors are small and unpredictable. I view this as a chaotic endogenous variable, and it is a truly animal spirit. Maybe The Animal Spirit.

    This is why I reject your Ricardian equivalence argument, and why I reject Scott Sumners expectation arguments. I think that Nick Rowe is heading in the right direction when he muses if the monetary transmission mechanism is purely psychological. Nick just doesn’t appreciate what a bloomin’ buzzin’ confusion is already in that head.

    When you figure out how the mechanism between savings and investment actually works (JKH is on it!) you may get a different understanding of the Chartalist position. We, like our future and past selves, get to consume what we have (real). We cannot send goods and services backwards in time. This is what you are interpreting as a “series of presents”.

  74. Mr. E writes:

    ” I think that the marginal value of financial savings is, beyond a certain point, decreasing to savers, so that the same net stock of financial savings can result in greater aggregate demand if distributed broadly rather than narrowly. (If distributed narrowly, many agents making the spend/save decision will be savings starved, and save. If distributed broadly, those same agents value the next dollar of savings somewhat less, increasing the likelihood of spending.)”

    I see this distinction as being crucial to the development of any economic regime, and to the increase in true wealth in our society. This is a simple market segmentation approach – different consumers will have different supply demand curves for essentially everything, and this includes money.

    If you just use a simple model where 10% of the people have 90% of the savings, and the other 90% has 10% of the savings, you will get this kind of economy where bad things happen.

    1. The demand for savings in the lower 90% causes deflation in the consumer prices of most non-necessary consumer goods. This will squeeze most businesses profit margins and decrease the demand for investment in primary markets like IPO’s and new debt issuance.

    2. The surplus of savings causes financial assets and real estate to bubble in price as the supply of savings outstrips the demand by the 10%. So secondary markets will be over subscribed as the supply of attractive new investment falls relative to the supply of funds due to 1.

    The aggregate level is hugely important, but the distribution is crucial. As a reducto ad absurdum, imagine there was the same level of savings in the world, but only held by 1 person. This clearly would result in very little economic growth.

  75. Steve Randy Waldman writes:

    JKH — Thanks for the thoughts and the pointers to the Dickson/Carney clips. I’ve begun to go through them, though it will take me a while to get through them all. The stylistic difference between Canada and the US are stunning right from the start — Julie Dickson is a very different kind of person, who takes a very different kind of attitude, than a Geithner, Bernanke, or even a Sheila Bair.

    As I watch, I guess there are two questions going through my mind. The first is, as we’ve discussed before, to what degree was Canada’s resilience a matter of kinetics, that the Minskian corrosion that affected most other developed banking systems was occurring, but more slowly, and to what degree are there differences that might exempt Canada from that dynamic. The second question is to what degree are solutions transportable, and to what degree do they depend on cultural differences that may be hard to reproduce in a very different business and regulatory culture. Watching Dickson speak makes me a bit jealous: she is not boasting, grandstanding, or accusing. She sounds like someone who takes an important job seriously and is able to resist headwinds and stay focused. The question of capture is on my mind: it seems as though she is very close to the main banks in Canada, but she does not seem to have adopted the financial-utopian perspective that even now sits just behind the pronouncements of Bernanke and Geithner (although not Bair). (She acknowledges but sounds not particularly moved by banks pleading of competitiveness concerns as an argument for deregulation, whereas our regulators constantly try to influence the political process with exactly that argument. She seems as though she values soundness more than activity, whereas our regulators are desperately concerned with anything that will curtail activity.)

    Anyway, I’ve seen 1.5 out of 8 videos so far, so maybe more later. First impression is that it’s unclear to what degree ’twas a difference in kinetic vs a difference in kind, but regulators’ willingness to take a go-slow, wait-and-see attitude in the face of bankers’ pleading and urgency (re leverage ratio) suggests a mix of both: that they were able to resist at all is a difference in kind, that the resistance was couched as wait-and-see suggests a longer period of stability might have proven toxic. It is the cultural differences that seem most astonishing.

    I wonder about the institutional structure of Canadian regulations (maybe further videos will inform). It’s clear that Ms. Dickson is cognitively more distant from aggressive bankers than US regulators. (Perhaps Canadian bank leaders are cognitively more distant from aggressive bankers than US executives, but perhaps something in the institutional structure and accountability lends resistance.) How is the superintendency of financial institutions constituted in Canada? I suspect there are not multiple regulators that large financials can choose among, and I suspect that the leadership is not a board the majority of whose members are nominated by the banks that will be regulated. (Is insurance regulation unified with bank regulation in Canada? Is all financial regulation federal, or is there a provincial component?)

    Anyway, I will watch the rest, with my two core questions close at hand. Thanks for the pointer, and the thoughts on the housing market (to which I have little to add, except that it’s worth watching).

  76. Steve Randy Waldman writes:

    Winterspeak — My view is very, very far from Ricardian equivalence. Ricardian equivalence is the claim that a government deficit cannot affect current demand because agents’ expectation of a future tax burden will exactly offset the effect of greater present wealth or income. I think that position is bullshit, empirically, and that the models upon which it is based are deeply flawed. I think it’s quite clear that a current deficit can dramatically boost current demand and alter individual behavior. But that doesn’t mean that the existence of a growing notional debt has no effect, quantitatively or qualitatively, on behavior (which I think is your position, but feel free to correct me). My view is that public debt, particularly short-term public debt, renders current and future behavior less predictable and at the margin reduces incentives to undertake the sort of investment for which predictability and stability are important. I think there are costs to increasing public debt, but those costs can easily be justifiable if the spending that creates the debt is productive: The volatility of flows only matters relative to a surplus of economic capacity; if the sustainable surplus of economic capacity is expanded by public spending more than the volatility of potential flows is increased, the public spending is net beneficial both in present and future terms.

    I don’t believe that the traditional long-term government solvency constraint (that the expected present value of future tax receipts be greater than the debt) binds, or is even meaningful in practical terms. Governments are infinite horizon enterprises, until they are unexpectedly not, and future appetite for government savings as well as future taxing power affect solvency, appetite for government savings is not at all predictable based on “fundamental analysis” of government solvency, both savings flows and growth-related flows are best understood as random variables (whose distribution we can influence), and reasonable strategies trade of maximizing growth with minimizing volatility of flows in order to avoid a government solvency crisis. (A government solvency crisis is defined, obviously, not by an inability to meet financial obligations in its own currency, but an ability to do so without adopting policies that either violate price stability commitments or else enforce austerity harmful to the real economy.)

    We, like our future and past selves, get to consume what we have (real). We cannot send goods and services backwards in time. This is what you are interpreting as a “series of presents”.

    I agree very much with you, that we only get to consume, or invest, what we have in any present. However, I do think chartalists often underplay the degree to which choices in one present affect the range of choices we have available in some future present, either because investment was unwise in the first present, leaving us with fewer or worse real resources to work with, or because contractual or financial arrangements in the first present imply a choice between poor distribution of resources or breaking promises going forward.

  77. Steve Randy Waldman writes:

    Mr. E — Well said, and I very much agree. (I’ve little to add, forgive my terse enthusiasm!)

  78. SRW, JKH, Winterspeak . . . wish I’d gotten in on this. Great thread. Thanks for the link, Winterspeak. By the way, is this the link on loans create deposits you were looking for?

  79. winterspeak writes:

    SRW: “My view is that public debt, particularly short-term public debt, renders current and future behavior less predictable and at the margin reduces incentives to undertake the sort of investment for which predictability and stability are important.”

    I think you actually have this exactly backwards. Can you figure out why?

    Mr E. & SRW “The aggregate level is hugely important, but the distribution is crucial. As a reducto ad absurdum, imagine there was the same level of savings in the world, but only held by 1 person. This clearly would result in very little economic growth.”

    Not sure what you mean by “economic growth”, but the quantity of investment would be exactly the same. It will be good for you to figure out why that’s the case.

    Scott: You rock!!! That’s what I was looking for : )

  80. Ramanan writes:

    Here is how a simple production looks like with a bit of abstraction:

    1. The act of bank lending creates deposits: Assume the banking sector has lent $100. This means

    Change in production firms’ assets (increase in deposits) = $100
    Change in production firms’ liabilities (increase in loans) = $100
    Change in banks’ assets (increase in loans) = $100
    Change in banks’ liabilities (increase in deposits) = $100

    2. Now comes the production part. Let us say that the firms produce inventories worth $100 using the borrowing. The deposits move around between firms and the production leads to an increase in real assets but at the same time, firms also foot the wage bill. Typically firms keep very little cash balances at banks and wages are paid in advance. The deposits move from the production firms sector to the household sector. Hence we have

    Change in households’ assets = $100
    Change in firms’ assets = $100 (real assets) – $100 (loss of deposits)=$0

    The result of 1 & 2 is that the act of investment leads to an increase in households’ savings. At the end of stage 1 & 2 combined, this is what the changes in balance sheets look like

    Increase in firms’ assets (inventories) = $100
    Increase in firms’ liabilities (loans) = $100
    Increase in banks’ assets (loans) = $100
    Increase in banks’ liabilities (deposits) =$100
    Increase in households’ assets (deposits) = $100

    So previous savings didn’t create an investment, rather investment created saving.

  81. Ramanan writes:

    I got into this blog through Winterspeak’s latest post. Seems like a good discussion. I haven’t looked at the whole thing – going to take some time to read through. Good points raised about bankers chasing us for deposits which seems to contradict the fact that “loans create deposits” – I guess this is the first time I am seeing such a discussion in the blogosphere. Here is my take on it:

    Banks chase us for deposits because it replaces an existing liability in the money market with a cheaper one – deposits. This is independent of the risk of bank runs and even if the FDIC guarantees all deposits for free, banks would look to get deposits. In fact losing deposits is good from a bank run risk perspective!

    I wish there was a book called Money Markets and ALM from a Modern Money perspective or so. Have so many comments to catch up! Great discussion such as the Zero rate limit of all this etc…

  82. Steve Randy Waldman writes:

    Scott — Welcome. I’ve enjoyed your articles at Economic Perspectives from Kansas City, in Billy Blog comments, and of course on Skype Twitter.

    Ramanan — Thanks for the very nice balance sheet walkthrough. I like that you start from a zero net financial position, which of course is how all things do start, and emphasizes the post-Keynesian tradition too neglected in the blogosphere, circuitism.

    In the example, the core problem that has to be solved is that labor belongs to households but can only usefully transformed by firms. The financial system — banks — arise in this story to overcome a time and trust lag. A preexisting real resource (potential labor) could be transformed to a more valuable resource, but the labor has to trust the firm to deliver something of value after the labor is delivered. The bank, which is a priori assumed trusted, can overcome that, by giving to the firm a token that households accept to be valuable, in exchange for a debt of the firm. The bank trusts the firm (or at least that it has the power to recover from the firm), and households trust bank notes, so a chain of trust can be established to overcome the initial deadlock. Note also that, although there appears to be a de novo creation throughout the process, that is not what in fact has occurred. A real resource (labor) was transformed into another real resource (inventory). Left off-balance-sheet is that the household sector held the real resource at time zero, although arguably it would have squandarded that resource by time one. The net financial change of $100 might be a token representing real production (if labor would have been squandared), or might represent nothing but the replacement of an unaccounted resource with an accounted for claim (if the labor would have been used to create something of equal value absent the firm and bank). The reality is probably somewhere in between.

    It’s easy to agree that loans create deposits. That’s pretty obviously true. I find the question, “does investment produce savings” or “does savings produce investment” to be a bit silly. What occurs in the real world is that resources are transformed to patterns that human beings value more than old patterns. There are always preexisting resources, and then there are new resources. One might describe the preexisting resources as “savings” and the process by which those investments are mobilized as “investment”, in which case you’d say that savings is prerequisite investment. But then the whole exercise is fairly meaningless, because there is always “savings”, and if you tried to attach a scalar value to existing resources, you’d either value it at somewhere between its untransformed value and its best transformed value net of costs of transformation. (Which value you’d attach would depend upon your role and market power: if you were a monopsony employer hiring labor, you’d value prexisting resources near the low value. If you are a monopoly supplier of inputs, you’d value those inputs at near the transformed value.) In real terms, I can’t really wrap my head around the distinction between savings and investment, other than as a kind of stock vs process thing, or maybe as a way of categoring resources by their intended use (“investment” vs “consumption”).

    What is absolutely clear is that while real resources must precede real investment, financial savings need not precede real or financial investment. As your balance sheets nicely demonstrate, from no financial position at all, an investment project can be undertaken, financial tokens can be defined and advanced, and that process can facilitate the transformation of real resources into more valuable real resources. Financial savings needn’t precede real investment. Financial resources can be created de novo to overcome coordination problems in order to maximize the value of available real resources. And financial resources always net to zero.

  83. JKH writes:

    SRW / Winterspeak / Scott,

    I was planning on saying something along the lines of what Harless said. The correct definition of saving is essentially “not spending”. So I would restate Winterspeak’s “not-spending enables investment” along the lines that not-spending (saving) at the macro level is the necessary result of investment. And I like the clarity of Ramanan’s example, although I would add the following:

    The production of investment goods in the current accounting period results in matching income for the factors of production. As a result, considering the macro economy as a whole, an equivalent amount of income cannot be used to purchase consumption goods and services. By construction, this amount of consumption goods and services hasn’t been produced to accommodate such expenditure. Therefore, this level of income must be saved at the macro level. This allows for the specific micro example where the household in question (i.e. the factor receiving income from investment production) may choose to spend rather than save in the same accounting period. If so, somebody else can’t do the same thing. One way or another, saving must equal investment. As Harless says, “the decision to invest can force people, collectively, to save.” The fallacy of composition and proof by contradiction are important in resolving macro with micro.

    I disagree slightly with Harless where he says in a comment, “Regarding physical output vs. financial representations, I don’t see how saving can be seen in terms of physical output.” I think such visualization can be useful. The primary question is not whether saving can be seen in terms of physical output, but whether income can be. In a hypothetical economy without (modern) money or finance, but with some strange mechanism for allocating output as income, the real output that is not consumed must be output that is saved. Thus, investment equals saving in real terms.

    These things elaborate somewhat on one of the issues raised by Winterspeak, but don’t go that far in addressing a number of Steve’s concerns. That depends on some more complex finance derivations flowing from “investment creates saving”…

  84. Steve Randy Waldman writes:

    Winterspeak —

    SRW: “My view is that public debt, particularly short-term public debt, renders current and future behavior less predictable and at the margin reduces incentives to undertake the sort of investment for which predictability and stability are important.”

    I think you actually have this exactly backwards. Can you figure out why?

    Well, I’d prefer not to play guessing games, and let you make your case. If I have to guess, I’d say that you can make the argument that if a state predictable manages aggregate demand, that helps investors plan for the future, as they needn’t worry about demand shortfalls. The chink in that reasoning is that if the state manages aggregate demand, and a large stock of cumulative deficit permits old stocks of money to be unpredictably mobilized, in order to stabilize aggregate demand under a price stability constraint, the state may have to tax, and that potential and uncertain taxation (if income rather than wealth taxation) will reduce the potential entrepreneur’s risk-adjusted expected return relative to the alternative of holding safe and not-subject-to-further-taxation risk-free assets. (If the price stability constraint is weak, volatile flows might mitigate towards holding real resources instead of nominal safe resources. If the price stability constraint is uncertain, agents would want a mixed portfolio. But in any case, projects that require a stable planning environment would be harmed.)

    As in most things, I think there’s a trade-off: aggressive demand management via fiscal policy is both encouraging and discouraging of high quality real investing. A state’s ability to sustain stable and growing aggregate demand is encouraging of good investment. But if its ability to do so indefinitely is not credible (e.g. for political reasons) or if its strategy for doing so creates a financial savings stock that grows far faster than the real surplus (enabling volatile flows from dissaving), it implies the potential for uncertain taxation or inflation. So I think strategy is very important: states should deficit spend (among other strategies to help support aggregate demand), but do so in a way that maximizes the productivity of investment (to maximize the scale of the real surplus relative to the stock of savings and diminish the likelihood that attempted dissaving will put pressure on the price level).

    But you probably have a different case. Don’t keep us hanging. Tell your story.

    Mr E. & SRW “The aggregate level is hugely important, but the distribution is crucial. As a reducto ad absurdum, imagine there was the same level of savings in the world, but only held by 1 person. This clearly would result in very little economic growth.”

    Not sure what you mean by “economic growth”, but the quantity of investment would be exactly the same. It will be good for you to figure out why that’s the case.

    Do you mean investment in an accounting sense, or a real sense? What are you presuming of a state and/or banking system?

    If you presume a state that will with certainty generate a GDP-maximizing level of aggregate demand indefinitely and an optimal banking system, I agree. The one zillionaire puts all of her zillions in a bank, where it does nothing directly. But the state spends to ensure that, at all times, people have sufficient income to sustain demand, which means entrepreneurs have projects that would be valuable if funded, which means that banks will lend into projects that exceed in risk-adjusted terms the central bank’s policy rate. If banks invest on behalf of the state, and they are optimal investors, and if the state sustains some optimal level and distribution of demand via fiscal policy, then it doesn’t matter who saves what. As long as real productivity grows monotonically and past savers in aggregate do not unexpectedly dissave or switch to alternative savings vehicles, this sort of arrangement is theoretically stable indefinitely.

    But that’s a lot of optimization. If the state does not sufficiently manage the pattern and quantity of aggregate demand, if the state/banking-system combo invests more poorly than less-intermediated private investors, if the economy is subject to adverse shocks that might trigger the mobilization of past savings, then it matters very much how income is distributed, how much is saved as money vs how much is privately invested, and how large a stock of money/government obligations accumulates. How perfect a world do we need to imagine?

  85. Ramanan writes:


    Thanks for your response. Yes, my example was at some level of abstraction. I also agree with you that the question is actually silly :) But the reason it is important is that it also points out to the fact that “saving create investment” is actually sillier. The real world is more complicated and borrowing your words, The reality is probably somewhere in between. For example, the Saving = Investment identity also includes production firms’ retained earnings and that saving creates investment. In the real world, corporate borrowing instead of going directly to banks also create investment, but the important thing in the end is that the truth is so different from what a mainstream textbook tells you.

  86. JKH writes:


    “It’s easy to agree that loans create deposits. That’s pretty obviously true. I find the question, “does investment produce savings” or “does savings produce investment” to be a bit silly.”

    Steve, I think there is a problem here.

    One needs to step back and ponder an overarching question about mathematics, measurement, and economics:

    Is double entry book keeping an impediment or a catalyst toward clear thinking in economic analysis?

    Is it an inconvenient and unnecessary weight that should be jettisoned by creative thinkers in order to find the truth, or is it a ground of mathematical and logical truth against which conceptual thinking should eventually be tested?

    I know where I stand. I would develop economic theory using the logic of double entry book keeping as an important and steady anchor for truth testing.

    As a practical matter, it’s worth noting that the entire financial system is a construction of double-entry book keeping. Chartalism and MMT are important riffs on this theme. In my view, there is no self-constructed algebra of economics that can have integrity and at the same time reject double entry book keeping as a pristine logical component and reference point.

    Numerical details are often unimportant, as are the debates about things that should be capitalized and things that shouldn’t be, or marked to market versus accrued, or whatever. The logic of consistent interconnectedness is what is important; for example, as in the conceptual relationship among various components in the system of national accounts. Gradations along the continuum of consumption and investment may be debatable, for example, but such debates do not nullify the importance of relative conceptual distinctions among consumption, investment, income, and saving. This is the required framework for the differentiability of causalities.

    Acknowledging the value of the discipline has nothing to do with the profession per se. The intellectual kernel in question is a bit beyond a comparison of the personalities of accountants and actuaries. But in defence of objectivity, in full disclosure, and in paraphrasing the elephant man, it is important for me to note – I am not an accountant.

  87. JKH writes:

    SRW # 84,

    Incremental conversation tempts incremental observation. I think Winterspeak is still arguing fundamental causality here (i.e. in the first half), while you are arguing risk, volatility, and strategy with respect to government debt accumulation and potential mobilization of non government net financial assets at some point in the future. Whatever one may think of the importance of clearly delineated causality, the two issues are somewhat separable, and that may be part of the communication problem. Nevertheless, on causality, you are also correct in that it is very easy for the non government sector to use cumulative net financial asset saving as the immediate means of payment to factors of production for either consumption or investment at some point in the future, generating future income and perhaps even saving of potentially dubious value. That dynamic is simply an exchange of a past saving financial asset for present output, at that future point. Past saving at that point is being used in financial form as the medium of exchange to jump start the process of output and income and saving during that accounting period, rather than as the unleashing of past saving as the “cause” of subsequent spending or investment. The past saving has occurred and is done with. But its value as a financial asset can be used at any point. That is the temptation and the source of risk and potential volatility that you are talking about for the future, I think. It is representative of a somewhat classic existing money supply (or near money supply) argument, as opposed to the incremental new money example put forth by Ramanan. It is a risky event whose risk Chartalists maintain is a function of “real economy constraints” which should be greeted with fiscal discipline at that future point of risk, a potential event that I suspect you maintain should instead be discounted for its attached probability in today’s strategy, as we have discussed previously.

  88. JKH writes:


    Krugman op-edited his blog on Canada:

  89. winterspeak writes:

    SRW: Ack! I posted a link on my site because I thought the conversation was interesting. I didn’t mean to flood the zone with the PK Posse. In future, I’ll abstain until the conversation is over.

    To my 2 points, and then I’ll sign off.

    1. The reason why a larger public debt, all else equal, is a stabilizing force is because it means that the private sector is less levered, and thus less likely to fall into debt deflation. Feel free to look through historic deflationary and inflationary episodes and see what makes most sense to you.

    2. On the distribution of savings being immaterial to the quantity of investment, see JKH and the Harless post. The equality of those two (technically “not spending” and “investment”) is from correct double entry book keeping, and like all accounting systems it is meticulous about quantity but not the right place to for for quality.

  90. JKH writes:


    Here is some general information on OSFI that you may find helpful. The mandate is fairly comprehensive. I’ll just add that securities commissions fall into a separate regulatory category in Canada. And they’re balkanized – one per province – which has been an outstanding issue for some time. The feds have been trying to consolidate them for a decade at least, but run into resistance, particularly from Quebec. The Ontario Securities Commission (OSC) has come under much pressure over the years for failing to prosecute insider trading cases and fraud aggressively enough, but have been getting tougher in recent years.

    Office of the Superintendent of Financial Institutions Canada


    The Office of the Superintendent of Financial Institutions (OSFI) was established in 1987 by the Office of the Superintendent of Financial Institutions Act. OSFI is divided into a Corporate Services Sector, a Regulation Sector, a Supervision Sector, and the Office of the Chief Actuary.


    OSFI is responsible for regulating and supervising all banks in Canada and all federally incorporated or registered trust and loan companies, cooperative credit associations, life insurance companies, fraternal benefit societies, and property and casualty insurance companies. It also regulates and supervises federally regulated private (employer-sponsored) pension plans that are subject to the Pension Benefits Standards Act, 1985. Through the Office of the Chief Actuary, OSFI provides a range of actuarial services, under legislation, to the Canada Pension Plan (CPP) and some federal government departments, including the provision of expert and timely advice in the form of reports tabled in Parliament

    Bank Act
    Cooperative Credit Associations Act
    Green Shield Canada Act
    Insurance Companies Act
    Office of the Superintendent of Financial Institutions Act
    Pension Benefits Standards Act, 1985
    Trust and Loan Companies Act


    Corporate Services Sector

    The Corporate Services Sector provides the ressources and infrastrucure necessary to support supervisory and regulatory activities. It leads the strategic planning process, and ensures sound administration of the organization. This Sector comprises four divisions: Communications and Public Affairs; Finance and Corporate Planning; Project Management and IT Services; and Human Resources and Administration.

    Regulation Sector

    This Sector has seven divisions that develop regulatory policy and guidelines, prepare recommendations with respect to applications for regulatory consents under the legislation, develop rulings that set precedents, and draft technical aspects of financial sector legislation. The Sector also contributes to the risk assessment and intervention process by providing specialized expertise to the Supervision Sector. The Sector coordinates OSFI’s relations with other regulators and supervisors and works with a number of domestic and international organizations that deal with financial stability issues. The Sector also provides legal support to OSFI and supervises all federally regulated pension plans.

    Actuarial Division

    This Division ensures that appropriate actuarial knowledge, advice, and standards are applied to OSFI’s regulatory and supervisory functions. The Division assists in the formation of policy and the preparation of guidance on actuarial and insurance issues. It works closely with the Capital, Accounting and Research Division with respect to insurance industry matters. It also works closely with the Actuarial Standards Board and the Canadian Institute of Actuaries to ensure that appropriate actuarial standards are in place. The Division also provides support to the Supervision Sector in the supervision and examination of federally regulated insurers. It also specifies requirements for and reviews the content of several actuarial reports prepared by insurers’ appointed actuaries.

    Capital, Accounting and Research Division (CARD)

    The Division develops capital standards for domestic financial institutions, contributes to the development of sound international capital standards, approves capital precedents, provides specialist support on capital issues to the supervisory and regulatory functions, and promotes understanding of capital requirements by OSFI staff and external stakeholders. In addition, the Accounting Policy unit within CARD acts as the centre of expertise in accounting and auditing standards for FRFIs; as well as international accounting and auditing policy and practices. APD develops accounting guidance and regulatory standards that apply to FRFIs. It contributes to international accounting guidance and supports the Supervision Sector by interpreting and clarifying accounting practices. The Division also performs research and analyzes industry-wide, sectoral and company-specific issues.

    Compliance Division

    The Division evaluates the effectiveness of the compliance functions of financial institutions supervised by OSFI, with due regard to emerging risks. The Compliance Division conducts assessments of financial institutions’ anti-money laundering and anti-terrorism financing procedures, and works closely with the Supervision Sector and the Financial Transactions and Reports Analysis Centre of Canada (FINTRAC), Canada’s financial intelligence unit. The Division also acts as OSFI’s liaison with other agencies on financial crime-related matters.

    International Advisory Group (IAG)

    The International Advisory Group provides technical assistance to countries seeking to strengthen their supervisory and regulatory practices. This Group works primarily with supervisors in the Caribbean, Africa, Asia and Latin America. IAG’s responsibilities are to develop, implement and administer technical assistance programs, including hands-on advice and training, seminars and workshops.

    Legal Services Division

    This Division provides or obtains all legal advice and legal services required by OSFI, including interpretation of statutes, regulations, and other legal instruments and preparation of draft legislation and regulations, contracts and legal documents. It also provides advice on policy development and lends support to litigators acting on behalf of OSFI.

    Legislation and Approvals Division

    This Division has three principal areas of responsibility. It contributes to the development of legislation and regulations that maintain or strengthen public confidence while recognizing the need for financial institutions to compete. It prepares recommendations related to applications made by financial institutions for regulatory consents under the legislation and provides rulings and interpretations pertaining to legislation, regulations and OSFI guidelines. It works with supervisory agencies in other jurisdictions to enhance ongoing cooperation and appropriate information exchange.

    Private Pension Plans Division

    This Division is responsible for supervising private pension plans that fall under federal jurisdiction. These include pension plans for some federal Crown corporations, banks, companies involved in interprovincial and/or international transportation, and communications companies, among others. The mandate of the Division is to protect the rights and interests of plan members by determining whether pension plans are in sound financial condition and in compliance with the requirements of legislation and OSFI policies and guidelines. The Division relies on plan administrators for good governance and prudent management of the pension plans for which they are responsible. In addition, the Division intervenes, when necessary, to ensure that appropriate corrective actions are taken.

    Supervision Sector

    This Sector supervises all federally regulated financial institutions. The Supervision Sector is responsible for overseeing the overall safety and soundness of the institutions by: analysing risks and identifying trends and practices within the financial services industry; developing a supervisory plan for each federally regulated financial institution that includes examining and monitoring the financial condition and effectiveness of oversight and control of those entities; and making recommendations and following up on corrective action, consistent with OSFI’s early intervention mandate.

    Capital Markets Risk Assessment Services

    This Division evaluates the risk management practices used by financial institutions to measure and monitor capital market-related risks, conducts ongoing research into capital market-related risk issues and promotes understanding of capital markets’ operations by supervisory and regulatory staff.

    Credit Risk Department

    This Division carries out monitoring, on-site and early intervention activities at financial institutions, with respect to credit risk and risk management practices.

    Financial Conglomerates Group

    This Group supervises federally regulated conglomerate deposit-taking institutions and conglomerate insurance companies, including all the federally regulated entities in each conglomerate. The responsibility for each conglomerate is assigned to a Relationship Manager, who carries out the supervisory work with a team of supervisors.
    Financial Institutions Group (FIG)

    This Group supervises non-conglomerate deposit-taking institutions and insurance companies. It is divided into two sections. One focuses on deposit-taking institutions (Schedule II Banks and trust and loan companies) and one supervises companies in the insurance industry.
    Monitoring and Analytics Support Division

    This Division is responsible for providing key metrics, including peer benchmarking, to support Relationship Managers in their monitoring activities. In addition, the Division is responsible for monitoring the health of the deposit-taking, life and property and casualty industries to identify emerging issues, trends and risks in these industries.

    Regulatory and Supervisory Practices

    This Division develops and recommends effective regulatory and supervisory practices and assists in the implementation of approved practices. The Division is committed to ensuring that OSFI practices are up-to-date, effective and adjusted on a timely basis to reflect developments in Canada and internationally.

    Risk Measurement and Analysis Assessments Services

    This Division provides expertise to assess federally regulated financial institutions’ (FRFIs) risk measurement capabilities. In addition, the division prepares recommendations related to applications made by financial institutions for regulatory consent to use models in the calculation of regulatory capital.

    Supervision Support Group

    The Supervision Support Group provides technical support to supervisors in the areas of capital markets, credit and operational risk assessment. The group is divided into five divisions.

    Audit and Consulting Services

    This Division reports directly to the Superintendent and provides internal audit and consulting services to the Office. It provides independent objective assessments, information, and solution-oriented advice to help OSFI divisions achieve their business objectives efficiently and effectively in order to fulfill OSFI’s mandate.

    Office of the Chief Actuary

    The Chief Actuary and his staff provide a range of actuarial services and advice to the Government of Canada regarding the Canada Pension Plan; the Old Age Security program; and federal public sector pension and insurance programs. In accordance with the Public Pensions Reporting Act and the Canada Pension Plan, this office prepares statutory actuarial reports on these programs for tabling in Parliament. The Office of the Chief Actuary (OCA) also provides actuarial information to government departments to assist in designing, funding and administering these programs. OCA became responsible for undertaking the annual actuarial review of the Canada Student Loans Program in 2001.

  91. Steve Randy Waldman writes:

    JKH — I don’t mean to attack or diminish double-entry bookkeeping or the usefulness of a balance sheet perspective. I do think questions of what’s laced on- or off-balance sheet are fairly crucial; that choices of what to include and at what values relative to other resources can completely reshape ones perspective of financial position, and what sort of choices seem reasonable. To take an obvious example, alternative accounting choices that would have included more contingent liabilities at higher (negative) valuations would have increased the apparent leverage at many banks and forced earlier retrenchment.

    In the savings/investment context, I think the problems get to be definitional. Do you treat the savings as the liability side of the balance sheet and investment as the asset side? I’m fine with that, but then of course they are an identity. One does not cause the other, although the existence of financial savings, and the careful accounting it enables (attaching claims to assets, allocating them between multiple claimants) might catalyze the process that transforms real assets. Similarly, the existence of real assets is prerequisite to their transformation.

    Growth, in both savings or investment, is in an accounting sense a bringing on-balance-sheet of unaccounted items, or a revaluation of already accounted items. (Or mirror-image occurrences with respect to liability.) When a firm sells inventory for profit, it realizes a value for assets above its earlier book value. It’s book value, usually based on historical cost, was erroneous, and a conventional and prescribed way. A more accurate balance sheet would value inventory at the expected value of the realized sales (which immediately calls attention to the problem of using point estimates when we’re talking about distributions), and would show profits immediately upon production that might be adjusted upwards or downwards depending on how realized sales perform relative to expectations. But then we can step a level back: the inputs of production then shouldn’t be valued at cost, but at the expected realization value of the transformed product (net of production costs). If you take this reasoning to its logical conclusion, you get a balance sheet that is what a very theoretical accounting would prescribe, a balance sheet whose asset side represents the capitalized expected net cash flows from the entire future operation of a firm, and whose liabilities reflect all expected financial claims, and whose book equity is (under am efficient markets hypothesis) precisely the market value of equity.

    Of course we don’t do that, because despite its theoretical elegance, it’s informationally vacuous (our ability to form expectations is too imprecise) and ruins our incentives (if we already own the happy future of our firm, why must we work so hard to ratify that future?). In fact, the expected performance of a firm is conditional upon our behavior, and our behavior is conditional on how we construct our balance sheet, so we construct our balance sheets “conservatively”, in a manner that creates on-average positive income statements, so we can feel like we have to work to get what we want (rather than feeling like we already have what we want but have to work to avoid losing it). The whole exercise is a lot like setting a clock forwards by five minutes to avoid trying to be late, although fortunately the uncertainty of the magnitude of our systematic misvaluation of enterprises means we can’t just easily and automatically compensate, we get trapped by our helpful delusion.

    Accounting in the end is just that: it’s about creating helpful delusions. No set of accounts can be accurate. That’s impossible. So we define conventions that are inaccurate in helpful ways: that define clear allocations of claims in ways that are likely to inspire production (and that are true because we ratify them ex post), that understate values to encourage us to realize what might in some sense be expected, but what is only potential, and also (crucially) in ways that prevent us from overcounting chickens before they hatch.

    I’m all for it, I think accounting is crucial (and malleable, and worth thinking about old ways and potential alternatives that might be more helpful). Ultimately I’ve forgotten where I’m going with this. But I still don’t get the point of arguing about causility between savings and investment. I’m all for pointing out that it’s bullshit to say that financial savings is prerequisite to investment, to imagine that we are constrained in aggregate by a “loanable funds” market. Financial savings certainly does not cause, and is not prerequisite to investment, although it may encourage investment by making agents more risk-tolerant. Beyond that, real production increases both savings and investment as we usually account for them and cumulative deficits in one sector create the “net financial savings” of a sector defined by its compliment, in a process that need bear no relation at all to the transformations we call production (but that we try to arrange to be helpful to that production).

  92. Steve Randy Waldman writes:

    Winterspeak — The deleveraging point is a good one, but also a subtle one. Consider two (related) implications of an entity being less leveraged: 1) generally speaking, a less leveraged entity is more risk-tolerant, as it can bear more volatility before facing distress costs; 2) when adverse events do occur, an unleveraged entity can draw from equity to cover losses or claims without surrendering control or cutting back on the core activities from which its value is derived.

    For agents individually in an economy, cumulative deficits do on average deleverage in both senses above. I think you are right to suggest that there is a pro-investment side to stock deficits because equity positions increase agents’ willingness to take risks, and even though agents in aggregate can have positive equity without government debt (real savings), as long as government money is valued like real assets, the accumulated deficit will add to agents’ perceived equity position, reducing leverage as you say, and encouraging risktaking. Further, for individual agents, claims on government can be used to cover losses and pay down debts. I agree that deficits can have a real positive effect in this way, essentially by helping to pool the risk of real investment by offering insurance in the form of equity cushions.

    However, it’s important to note that this equity, while it has aggregate behavioral effects (implication 1) can not be drawn down in aggregate without forcing agents to bear real costs. That is, if an individual firm needs to cover a loss, it can spend its cash to do so. But if the economy suffers aggregate losses in real terms, no pattern of spending apparent equity will conjure the lost resources, people unexpectedly will have to deal with the real shortage (unexpectedly,because each of them may have thought they had an equity cushion to protect them).

    The risk-tolerance enhancement effect is a function of distribution as well as quantity, and for any given agent is almost certainly diminishing in total equity. (If I have a turnip worth two dollars, and I owe the bank a dollar for it, giving me one dollar dramatically reduced my leverage and I care about it. If I have a turnip worth $2, no debt and $100 in the bank, giving me $1 more probably doesn’t make me more or less risk averse in my behavior.)

  93. winterspeak writes:

    SRW: “Winterspeak — The deleveraging point is a good one, but also a subtle one.”


  94. hi

  95. JKH writes:


    “I’m all for pointing out that it’s bullshit to say that financial savings is prerequisite to investment, to imagine that we are constrained in aggregate by a “loanable funds” market. Financial savings certainly does not cause, and is not prerequisite to investment, although it may encourage investment by making agents more risk-tolerant. Beyond that, real production increases both savings and investment as we usually account for them and cumulative deficits in one sector create the “net financial savings” of a sector defined by its compliment, in a process that need bear no relation at all to the transformations we call production (but that we try to arrange to be helpful to that production).”

    That sounds pretty good as far as causality is concerned.

    Accounting is malleable; the design of an accounting system is like the design of a financial system. It’s important to have measurement or operational consistency for the whole system.

    I think accounting, disclosure, and risk management are sometimes misaligned. It’s not always necessary to bring the worst marked to market case directly into accounting in order to have full disclosure and informed risk and capital management. How much MTM to bring into accounting and capital measurement is not an easy call.

  96. najdorf writes:

    I’m still confused by the original proposal in the main post – money market funds are an adequate substitute for insured deposits? I assume that Rajan is not talking about Treasury MMFs because he calls the substitute funds “near riskless” rather than “a more convenient way to hold risk-free assets”. Every standard money market fund I’ve observed holds both CDs in banks backed by FDIC insurance (even if they’re not insured to the amount of the CD, the FDIC increases the stability of the institution the MMF is exposed to) and risky commercial paper assets payable by non-banks (such as the Lehman exposure that brought down the Reserve Fund). So the structure of a MMF is in fact significantly risky and significantly dependent on the existence of insured bank deposits. The only way to get around this problem would be to declare government insurance on money-market funds, which of course is what the government did (temporarily), but which also makes no sense since there are already ways to get government savings insurance in a fashion deemed socially useful and it’s not entirely clear why AIG or GE’s equity-holders should be allowed to fund their holding company by selling CP to investors who only want it because the government is taking the downside.

  97. Hi how are you i really liked this.

  98. JKH writes:


    I think I’ll revisit this investment/saving thing down the road a bit. Looking at Harless’s explanation again, I don’t think it’s adequate. Arguably, it’s even wrong. (But not because saving funds investment.)

  99. Philo writes:

    “Ordinary households”? “Salesclerks and schoolteachers”? Yet you want to insure *one million dollars* for each depositor (two million per couple)? Are you anticipating hyperinflation?

  100. vimothy writes:

    JKH (or anyone, really),

    Are you still about–can we open this debate up again? I don’t think that Harless is right. There look like there are some (instructive) logical fallacies in his account.

  101. JKH writes:


    I’m about (have to be careful on how to pronounce that word).

    I see you’ve done a post. Any additional thoughts? I’ll be back sometime tomorrow.

  102. vimothy writes:

    The post on my blog? Are you reading my mind!

    I’m thinking the loanable funds angle might be confusing the issue. Not really got time time to properly read Andy’s post properly, but if it’s consistent his title, I think he has the causality and / or the sectoral flows mixed up.

  103. vimothy writes:

    “…properly read Andy’s post properly…”


  104. vimothy writes:


    While we’re on the same subject, I was struck by Steve’s GDP identity with decomposed government spending. Actually, I think we can make things significantly clearer if we aggregate private sector investment and consumption spending, giving,

    P = C + I

    Y = P + G + NX

    Where P is total private sector spending and C and I are private sector consumption and investment spending, as usual. Now each component represents a discrete sectoral balance. We tend to think that investment is a sector and consumption is a different sector. I think that this tendency is central to the problems we have understanding the savings-investment paradox.

  105. vimothy writes:

    Basically, the aggregate supply of paper assets in the private sector is not altered by any change in spending patterns, be it C or I, up or down, left or right, even if you assume the loanable supply of funds hypothesis. Savings does not cause investment, investment does not cause savings.

  106. vimothy writes:

    aggregate = net, but I’m sure you can see what I mean.

    Not-spending can never produce additional funds for investment.

  107. vimothy writes:

    Admin to thread!

  108. vimothy writes:


    Apologies for the piecemeal nature of my comments. I’m just adding thoughts whenever I get a spare moment. When we both have the time we can work out something more systematic.

    Reading Andy’s post, he starts with a very amiguous statement and then goes right off track. I am not sure why Warren didn’t pull him up on this.

    Bottom of the second para: “Thus “output equals consumption plus investment.” Savings are defined as unconsumed income. Thus “savings equal income minus consumption.” You do the algebra.”

    Indeed, let us do the algebra and untangle this mess. I think Andy has introduced a logical inconsistency into his model; thus it melts into air.

    In any discrete sector, output equals consumption spending plus investment spending.

    Total income equals total expenditure for that sector.

    And in any discrete sector, the flow of savings equals the flow of total income minus the flow of total expenditure.

    I wonder if we could ask Andy to join this debate? I’m sure it would help to properly nail this conceptually…

  109. Steve Randy Waldman writes:

    vimothy — Definitions can get so slippery.

    In a traditional world, in which the financial sector is transparent, S = I. Investment expenditure is savings, period.

    The chartalists divide the world into a public and private sector, and introduce the concept of net private financial savings. So now we have two forms of savings: 1) traditional savings, which is investment expenditure, and 2) financial savings, which represents a surrender of real goods or services to the state in exchange for money or government debt. (Thus “government spending creates financial savings” — when the government exchanges money for your real goods and services, it is spending.)

    However, this wrinkle does not eliminate traditional, real savings, which is precisely aggregate investment expenditure. It is an additional, odd form of savings, an accounting entry that may or may not represent some real resource, depending on how government employed the goods and services it purchased.

    Anyway, I think it’s not accurate to say “the flow of savings equals the flow of total income minus the flow of total expenditure”. For the chartalists, the flow of private savings equals traditional investment expenditure plus government net spending, or income minus consumption expenditure. The flow of real savings is investment expenditure, the flow of financial savings is net government spending, and the only possible place for any other income to flow would be consumption.

    The chartalists (to their discredit, in my opinion) de-emphasize the direct real investment component of savings. I think they sometimes imagine a world where all real investment is made by banks, privately directed but with capital provided by the state. Under a simplification of the world, all private parties save net financial assets by failing to use real goods and services but surrendering them to the state. The state then puts all unused real resources to use as investment via the banking system. Under this model, there is by assumption no private real investment, all private net financial savings is precisely matched by public (bank-mediated) real savings, so net government spending equals net real investment, and savings flows are the difference between income and all private expenditures, as you suggest.

    There’s an elegance to that view of the world, but I don’t think it’s a good framework for reasoning about real problems. Direct investment of private capital, private expenditure that is savings, is a very important component of real investment, while the efficiency with which private accumulation of money balances translates via the state or banks into productive real investment is suspect, in my opinion. So I don’t think it’s a good idea to define private savings as income minus all expenditures. Private savings includes private real investment.

  110. mtmiwuqlbq, Cialis, SnKfhot, Cialis, FUkSXpC, Generic Levitra, yKPbSWs, Buy Cialis, MsELFFp, Propecia, bwrJXcO, Tramadol, jCDWuNc, Phentermine, qeCbTSJ, Viagra, CfckJIB.

  111. vimothy writes:


    I hope that this has nothing to do with Chartalism per se. All I am aiming for is logical consistency.

    And we will get nowhere if we try to integrate S = I into this problem from the outset. Be very careful here!

    Let me post my toy economy and see if you can spot any internal inconsistency. Then if we agree perhaps we can try to integrate the S = I identity.

    A common fallacy is to assume that if spending on consumption falls, then investment spending will increase because there are more savings available for investment, so that there is no change to output.

    If (and only if) C falls while Y remains constant, then of course the additional spending must have come from one of the other components in the identity. If G and NX do not rise then it must be from I—this is simply circular reasoning.

    However, you cannot hold output constant: it is the dependent variable, the determinate of the other variables in the identity. Aggregate savings do not increase as C falls. This paradox is hard to grasp, but let me give it a shot.

    Consider a closed economy at T0 with potential output of $100 and a stock of money equal to the same. If at T1 AD is $100 the economy is at potential. But what happens when the economy as a whole saves 5% of its income? At T2 total spending equals $95 and so total income equals $95. GDP is now $95—output has fallen below potential and the economy is in recession. The stock of paper wealth did not increase so there are no additional funds available to boost investment spending—all we have is the same $100 we started with.

    There can never be an aggregate increase in paper wealth from within this closed system. It has to come from somewhere else—a cross-sectoral deficit flow.

    Do you agree?

  112. vimothy writes:

    Feel free to mess with my toy economy, but please don’t feed them salt because they die.

    For example, adding a banking system changes nothing, be it 100% reserve, money multiplier, neo-Wicksellian endogenous gredit growth. Everything still nets to zero for now additional net financial assets to fund investment and maintain output at potential.

  113. vimothy writes:

    To make things simpler, I’ve just led a tea party over-throw of the government and salted the rest of the world, whilst chanting Jonah Godberg’s name backwards three times. The monsters are off the map. Now for the economy as a whole,

    Y = C + I

  114. Steve Randy Waldman writes:

    vimothy — I certainly don’t mean to impose S = I as a definition. That’s at least as misleading as imagining that all private savings is public spending. The core Keynesian/Chartalist insight that in a monetary economy what people perceive as savings is decoupled from real investment is critical. So it’s better to define something like:

    private sector savings = private real investment expenditure + govt deficit
    total real investment = private real investment + govt/banking-sector real investment

    It’s certainly a fallacy to imagine less consumption implies more investment, since as you say, Y is a bitch and refuses to hold constant (or grow at a nice steady pace). If C falls and Y remains constant, then either the private or public sector are spending on goods we are classifying as investments. And if C falls, can just fall too, it’s as simple as that. The Chartalist view (which I think has a lot going for it, I’m more a sympathizer than a critic) is that if C falls, and private investment falls, better to have the state make up the shortfall in spending (call it C or I, however you wanna split it), and let the financial consequence — more net money or govt paper — happen. In a monetary economy, as long as private agents continue to accept money, the government can sustains C+I=AD in real terms, and the fact that people have more money makes them feel better too.

    In your closed economy, it’s pretty clear that we end up with $100 of private financial wealth. Assuming that in order to maintain economic capacity, all real resources must be mobilized, there are two possibilities for the real economy. 1) The price level fell, so that we mobilize all real resources with $95 in spending, and real capacity is unharmed; or 2) the price level was constant, and real resources went to waste, in which case we can expected inflation eventually as the constant stock of financial wealth chases less capacity.

    But we have to keep our accounting of the financial and the real distinct. That paper wealth is constant at $100 is really not what’s interesting. It’s what happened to the real economy (as a consequence of choices that may have depended on our attitude towards paper) that matters. At T2, the net stock of financial wealth will still be $100. Will that mean we are well off, in real terms, or poor? The money tells us next to nothing.

    There can never be an sectoral increase in financial wealth unless there is a flow from some other sector. That’s trivially true. But that tells us little about savings as a real economic phenomenon. And it doesn’t imply a necessary role for the state, if the household sector wants to hold financial assets rather than real resources. The business sector is in the habit of keeping large net-negative stocks of financial assets. Expanding business balance sheets translate to expanding household sector net financial wealth in the absence of any government / monetary changes.

    [I’m good re presuming a closed economy. I only brought in banking as the only way I know to rationalize a sometimes-Chartalist contention that net private financial savings — government spending — matches real investment. As I said, I think that’s mostly a bad story. Otherwise, we can leave banking out of this.]

  115. Rakeback writes:

    quucaolwyj, Poker, IOWaPjf.

  116. Live Sex writes:

    xzfucgajuv, Free poker games online, yzXTXjl, Online Casino, txtDqdh, mortgage calculator, qChZwTa, Live Sex, YQTuzQU, Homemade Porn, jrenUOr, Texas Holdem, EnUzHIa.

  117. vimothy writes:

    Steve–thanks for that. Your responses do seem to be a lot more thoughtful and substantial than mine. In my defence, I am just knocking these out while I’m at work doing other things. When I clock off for the day I will try to put something less measly together.

    You have done something to the petri dish and it looks a lot more complicated!

    Let’s step back a sec–do you agree that without a cross-sectoral net spending flow output will fall? Time is not continuous in my toy economy and total income is total expenditure so that if any positive net savings rate equals a reduction in output for that period for no extra net financial assets to increase investment spending. Yeah?

    Remember that we can still include a banking sector in the toy economy. As far as I can tell (we’re at the early beta stages yet!), it makes no odds to the model’s logic. Choose money multiplier, endogenous credit growth or 100% reserves. However, we can also ignore it–since that makes no difference either to the final outcome–but just remember that the reason we’re ignoring it is because it doesn’t get us out of the paradox of thrift.

  118. JKH writes:

    I hadn’t been planning on getting back into this subject so soon, but here is my conception of the issue, roughly:


    Saving is equal to income not spent on consumption goods and services. It is defined by what it is not, rather than by what it is. I agree with Harless there. And it is defined as a component of income. Therefore, in economic measurement terms, it is a flow, and not a stock.

    The very important missing piece in Harless’ exposition has to do with the relationship of saving to subsequent stock measurements in economics. And this is the piece that’s missed by virtually everybody, and it is the piece that comes up most often in my discussions with winterspeak and Steve Waldman on this issue.

    That is the issue of balance sheet equity. Balance sheet equity is defined as assets minus liabilities. It is a measure of net worth. And it applies explicitly and/or by construction to all agent balance sheets in the economy. Households in particular have a balance sheet equity position. Importantly, balance sheet equity is not the same as an equity financial claim, as is the case with common stock, which is a claim on the balance sheet equity of a corporation. Balance sheet equity is not the same as a financial claim on it. And this logic is clear from the existence of household balance sheet equity without the existence of a corresponding financial claim.

    Why is this important to the issue of saving?

    It is important because it frames the definition of saving consistently in both an income context and a balance sheet context.

    Saving is what is not spent from income in a defined income accounting period.

    Cumulative saving (sometimes called savings) is what is not spent from income over prior income accounting periods.

    Cumulative saving shows up as an addition to balance sheet equity.

    Importantly, there is an analogous caveat with respect to the definition of what cumulative saving IS NOT. It is not a balance sheet asset. Therefore, cumulative saving, or saving for that matter, IS NOT a financial asset such as money, and it IS NOT a real asset such as plant and equipment. Cumulative saving is an addition to balance sheet equity.

    Working with a continuous time model where saving as a flow instantaneously meets cumulative saving as a stock, saving from income is added continuously to balance sheet equity.


    In a closed economy, saving equals investment. This is independent of any financial arrangement.

    Importantly, it is a macro relationship, true of the entire economy, in aggregation.

    It is not necessarily true otherwise – i.e. in disaggregation. In fact, it is almost inevitably not true.

    The logical reason it is not true in disaggregation is because of the existence of micro dissaving. Relative to any defined closed economic space, macro dissaving does not exist, because micro saving and micro dissaving net to an amount equivalent to investment.

    Micro dissaving exists with sectorsin and across sectors.

    The best illustration of micro dissaving occurs in that of a current account deficit in the context of the world as the closed economy. The US current account deficit represents dissaving at the micro to world level, because the US is purchasing mostly consumer goods and services in excess of its own income. Because it is “not spending” a negative amount, its micro to world saving is negative, in respect of that deficit. It is dissaving. Conversely, the rest of the world is saving an amount of income equivalent to the value of its net exports to the US, because by definition that income is not available to purchase it own output. Net saving between exporters and importers is zero, and no investment has taken place relative to those (presumed consumer goods) flows.

    An example of national closed economy dissaving is that of borrowing from the bank to purchase consumer goods and services. By the same logic as above, that represents dissaving, and an amount of income equal to that earned by the factors of production relative to the same output must be saved. But net saving between the buyers and the factors is zero, and no investment has taken place.

    That brings us to the very interesting case that is the trophy of Chartalism – the saving relationship between government and non government. Through deficits, government is a dissaver and non government is a saver. The net position is zero. Real investment is not mentioned. Enough said on that.

    All of these examples are consistent with the existence of sector or agent dissaving in a world in which aggregate saving net of dissaving must equate to real investment. Saving is what is not spent from income, and cumulative saving is what is accumulated over time as balance sheet equity.

    There is a practical problem of proper classification of investment goods versus consumption goods, of course. But this classification issue must be seen for what it is and dealt with accordingly. It should not be seen as some nihilistic confirmation of futility in making such a conceptual distinction.

    My disagreement with Harless is mostly in the area where he begins to describe financial assets as saving. They are not. Saving is income not spent; cumulative saving is balance sheet equity.

    In fact, the correct accounting relationship between saving and financial assets is found in the sources and application of funds accounting statement (at the macro level, flow of funds), which is a separate measure from either the income statement or the balance sheet. The balance sheet is only a partial derivative of the income statement. It is a total derivative of the flow of funds statement.

    E.g. saving may initially be reflected as an increase in balance sheet equity and an associated increase in bank deposits. The income statement would show the source of the balance sheet equity increase as profit (or earnings less expenses in the case of a household). The flow of funds (sources and applications) statement would show equity increase as the consequence of the income statement result, and the bank deposit result as the “application” of that equity increase. The source is saving; the equity position is cumulative saving; the bank deposit is the deployment of saving.

    The fact that saving or balance equity can be deployed in either a real asset or a financial asset demonstrates that neither real investment nor financial asset acquisition can be a consistent or logical definer of saving (micro or macro).


    In summary:
    – Saving is not investment
    – Saving is not acquisition of or holdings in financial assets
    – Saving is income not spent
    – Cumulative saving is balance sheet equity
    – Micro saving plus micro dissaving equals macro saving equals macro investment
    – The Chartalist model is a micro construct relative to a world of government and non government; it does not contradict the fact that macro saving equals macro real investment

    SRW, I agree there is no need for Chartalists to de-emphasize real investment. The Chartalist model is primarily a “net financial asset” bolt-on feature. They could do with more integration. After all, the sector financial balances model (from which the bolt-on is derived) is itself a derivation of the national accounts model, which features the equivalence of real investment and saving at the macro level.

    I submit that all of the above is necessary in order to have a consistent and logical framework for what saving is, and what it isn’t.

  119. Steve Randy Waldman writes:

    vimothy — by assumption, intra-sectoral assets are only willing to spend $95, output as measured in financial terms will certainly fall, unless there is some cross-sectoral expenditure.

    what that means in real terms is less certain though. if the $95 spent mobilized all real resources, there has been a 5% deflation, but otherwise the real economy functions just fine. if the $95 has not mobilized the real resources, than there has been real waste. if there had been a cross-sectoral flow, we might have maintained price stability (neither inflation nor deflation) while mobilizing all resources, and accommodated the intrasectoral desire to hold $5 under mattresses.

    introducing a financial sector broadly complicates this story, because deflation becomes less palatable: it becomes less possible to imagine that people’s desire to hold cash doesn’t impinge upon the use of real resources because the price level falls. with a financial economy and nominal liabilities, a fall in the price level leads to default and uncertainty, and people strongly resist that, so absent cross-sectoral flows we’d expect wasted real resources and (carrying the model further in time) reduced real income (potentially eventually leading to stagflation, where the price level rises because resources have grown scarce, but debt is not easy to service because that scarcity implies real poverty).

    It’s the terribleness of deflation in a financialized economy (combined perhaps with psychological resistence to nominal wage and price cuts) that makes cross-sectoral flows so attractive: the economy won’t smoothly deflate to use all real resources, so if people want to hold money, best to inject some more…

  120. vimothy writes:

    – Saving is income not spent

    This is the key.

  121. vimothy writes:

    “Saving is equal to income not spent on consumption goods and services.”

    I think we disagree. Within a sector, the flow of saving is income minus expenditure. Expenditure includes both investment spending and consumption spending.

  122. Steve Randy Waldman writes:

    JKH & vimothy — I’m not sure where exactly we’d disagree, because broadly I’m with you. But I won’t acceede to “Saving is income not spent”. Saving is income spent on investment goods in aggregate (putting aside the complexity of the C/I classification.) For any given accounting entity, saving is an increase in or flow to book equity. I’m good with that. But an accounting entity can expand its book entity in lots of ways: it can receive financial income and not spend it, expanding its asset base with a line called “cash” and also expanding its equity. An accounting equity can also expand its book equity by purchasing an asset: I purchase a loom for $100, deduct cash from my balance sheet but replace it with loom. My book equity has still expanded, and my inclusion of loom on the balance sheet (as opposed to movie tickets I may have purchased) is justifiable because the future income I will earn from weaving is worth more than the $100 I have parted with. I have spend and I have saved at the same time, and my perfectly orthodox balance sheet has properly expanded.

    I think this kind of action is really the essence of saving. It is (overconservatively) accounted for on balance sheets, makes use of income flows for future benefit rather than present consumption. For an individual accounting entity, not having spent at all would have also been saving (and would have been accounted for equivalently, despite real differences). But in aggregate, savings looks much more like loom-buying than holding of cash. In aggregate, the holding of cash is irrelevant to real savings, except to the degree that it spurs and helps to make possible risky real investment, by helping people to manage risk.

  123. Steve Randy Waldman writes:

    vimothy — we definitely do disagree. both individually and in aggregate, and within any sector, i claim expenditures on investment goods are real savings.

  124. vimothy writes:

    Hahaha–if the identities are not nailed down, they float away!

  125. JKH writes:

    “Income not spent” where used is short form for “income not spent on consumer goods and services” – I thought that was self-evident

    The amounts of saving and real investment are equal in aggregate; but saving in substance is not equal to investment in substance, in my view. Again, should be evident from my comment above.

    Financial income is income according to standard national accounting. No different than wages. Both forms can be saved. No controversy there.

    You don’t change your book equity by exchanging cash for a different asset (e.g. loom). Neither affects your saving or cumulative saving position.

  126. Live Sex writes:

    cheezgooat, Live Sex, rpXawpq, Dating, IsgoVNV, Life Insurance, MGoIOWJ, Cigarettes, SqtzKjf, Free Porn, SfclrJH, Penis enlargement stretching, OflAjyk.

  127. dturner writes:

    very insightful discussion. unfortunately I think that the govt’s policies are going to cause considerable pain for the banking system in the long run because they have failed to deal with the fundamental problem of too much leverage and insufficient capital. and on a somewhat related topic, I recently came across this page, which has a really good collection of articles from many well known economics blogs and finance sites around the web:

    One piece I especially enjoyed was Marc Faber’s interview where he claims that if the US were a corporation, its credit rating would be junk

  128. Absinth writes:

    dudehisfwb, insanity, gjHqLyj, Casino Paypal, EIpEWzN, Buy Phentermine, KiHBQaz, Phentermine, EtSGiUj, Ambien, gaDnZky, Buy absinthe, OdLoDgj, Signs of vicodin addiction, EfLSfCr, Withdrawal vicodin, FurlpSB.

  129. vimothy writes:

    The important thing to remember is that within my amoebic economy, for any given period,

    The flow of savings can never equal the flow of investment spending. Not spending destroys total income for no net gain to loanable funds. As far as I can see this has to be true for the model to be consistent.

    Otherwise there is no paradox of thrift, output doesn’t fall when AD declines and there is no business cycle. Otherwise, any fall in C raises I and round and round we go.

  130. Steve Randy Waldman writes:

    JKH —

    I think that we are agreeing:

    “You don’t change your book equity by exchanging cash for a different asset (e.g. loom). Neither affects your saving or cumulative saving position.”

    Exactly. It’s isn’t not-spending that distinguishes saving from not saving. My exchanging cash for a loom — spending — doesn’t matter either way, from an accounting perspective. What matters is not-wasting-or-consuming, not burning the loom or spending on resources one couldn’t put on a balance sheet because they confer no ongoing benefits.

    As long as…

    “Income not spent” where used is short form for “income not spent on consumer goods and services” – I thought that was self-evident

    …then I’m pretty happy with calling the long form savings (as long as we are presuming minimal waste or destruction).

  131. Steve Randy Waldman writes:

    vimothy — I think in your amoebic economy the flow of savings could equal the flow of investment spending, but it needn’t necessarily: people can choose to hold cash, or hoard pre-existing resources rather than organizing new production. As long as people don’t have to invest savings, there can be a paradox of thrift. Sometimes people do invest savings, and there is no paradox: during a boom, that people start businesses or build houses or whatever doesn’t harm aggregate demand. During a bust, it is the character, not the quantity of savings that changes, and can lead to a paradox of thrift. Individuals, instead of saving-by-investment, seek to save-by-holding-money. That may keep everybody’s balance sheet safe and whole, but leaves the real economy struggling and real resources decaying. Over time, it reduces future income in both real and financial terms, as savers refuse to exchange to create financial income or mobilize productive activity.

    A paradox of thrift is never a necessary consequence of saving. It is a consequence of some kinds of saving, that may be individually rational but that in aggregate are inconsistent with the maintenance and expansion of real activity.

    I had a post about this once:

  132. Free Sex writes:

    zzomibwjtk, Free Sex, gtVQxUe.

  133. vimothy writes:

    Ah, Steve–I feel that we are very close to nailing this now! There are some important insights in your post, and hopefully if we work through the implications you will understand my position.

    Just got another hour or so and then work finishes…

  134. JKH writes:


    “Exactly. It’s isn’t not-spending that distinguishes saving from not saving. My exchanging cash for a loom — spending — doesn’t matter either way, from an accounting perspective. What matters is not-wasting-or-consuming, not burning the loom or spending on resources one couldn’t put on a balance sheet because they confer no ongoing benefits.”

    Not sure of your point here. If your loom is classified as an investment, which was my assumption, then you haven’t changed your equity position. That reflects not spending on consumption goods and services. So not spending (on consumption goods and services) means saving. Not spending, in the sense intended, matters. Other than that, you’ve swapped a financial asset for a real asset.

  135. […] flags Charles Calomiris making the argument, I wrote about David Henderson making the argument, interfluidity doesn’t get Rajan’s editorial on the subject and there are many more out […]

  136. Great post. It is clear You have a great deal of unused capacity, which you have not turned to your advantage.

    The way you write shows you have a need for other people to like and admire you, and yet you tend to be critical of yourself.

    It seems to me that while While you have some personal weaknesses you are generally able to compensate for them.

  137. Steve Randy Waldman writes:

    JKH — We’re not disagreeing in any substantive way. You are emphasizing that “not spending, in the sense intended” is saving, while I want to emphasize that spending is perfectly consistent with saving, if the expenditures are for investment goods. There’s no conflict, because the sense you intend is spending on consumption goods.

    I think your emphasis is intended to focus on balance sheet accounting. Not spending income is sufficient to increase equity. From a static balance sheet perspective, the composition of ones portfolio is epiphenomenal, so there’s no point in focusing on anything but the simplest way of letting ones balance sheet grow.

    My emphasis is intended to counter what I see as a widespread confusion that suggests that saving and investment spending are in conflict. I want to talk up the fact that you can go to a store and buy something, and still be saving the wealth you “spent”.

    On an idealized balance sheet, the issues are perfectly clear: As you say, it’s saving if it increases equity, whether you hold cash or swap assets. As long as our hypothetical accountant doesn’t try to capitalize expenses that don’t yield future returns, we can distinguish between saving and not saving (or dissaving) by watching that equity line.

    But I always want to remind that investing is saving too, because one of the many chips on my shoulder is that actively participating in the continual reshaping of the real economy is important, that practices that seem financially sensible but are distant from real activity may in a dynamic sense lead to poor aggregate outcomes (i dislike everything from holding cash to passive indexing). Again, this is all shading. I think our views are perfectly consistent.

  138. JKH writes:


    We’re understanding each other then.

    And you are being consistent:

    “I always want to remind that investing is saving too, because one of the many chips on my shoulder is that actively participating in the continual reshaping of the real economy is important.”

    That’s a rational basis for blending the two meanings. I keep them separate, but connected.

  139. vimothy writes:

    Guys (?)

    Thanks for this. You’ve given me an immense amount to think about. I have so may questions and so little time. Perhaps this is surprising, but I don’t think anything that you’ve said is inconsistent with my substantive point (insofar as it is substantive and mine), that the logic of the paradox of thrift is unassailable.

    Please me patient with me. I’ve only been trying to mentally “get” this issue for the last three years. I have no formal training, only some limited spare time, unlimited access to academic papers and a few second hand textbooks. If there are any putative disagreements, it may well be that am not expressing myself with due care and diffidence to accepted norms and mutually agreed on definitions.

    Despite what I take to be the fundamental compatibility of our positions, I few issues remain. In general, they are mine. I’ve been watching my amoebic single-sector economy and it doesn’t look well. Volatility is off the scale and it is mutating at pace. I’ve been experimenting with its structure and composition. Trying to integrate everything at once is proving very complicated and I think it is going to take a lot of work to get it up to spec. Once I get this sorted out hopefully
    I can explain my problem in terms of your individual framework and you can tell me if it has consistency or if I should go back and try again. Steve hopefully if I can do that you will see why I think you are slightly mis-conceptualising the paradox and thus misinterpreting a particular diagnosis of what I take to be, in the final analysis, as alternator trouble, as an ideological–or at least idiosyncratic–position (i.e. defining a Chartalist position).

    I did prepare a massive post with lots of iterations of the sequence with changed parameters and new variables, but that became needlessly convoluted and in any case I obviously need to sync scripts because at the moment I’m not sure we’re really talking about the same thing. Can I first just ask a some ad hoc questions about your last few posts?


  140. vimothy writes:

    Aaargh–apologies for typos galore in that post.

  141. vimothy writes:

    Think the only non-trivial one is,

    “(i.e. defining a Chartalist position)”

    Should be

    “(i.e. defining it as a Chartalist position)”.

  142. vimothy writes:

    Steve, I may have the question: why must one sector divide into two: private and public (or (C + I) and G)?

  143. vimothy writes:

    First thought:

    1, To fix the inherent intra-sectoral AD problem at an ultimate level;

    2, To enable its solution.

  144. vimothy writes:


    I submit that all of the above is necessary in order to have a consistent and logical framework for what saving is, and what it isn’t.

    Agreed. This framework appears to be consistent with the paradox of thrift as I understand it and so consequently suggestive of no immediate solution. I very much want to be able to frame the problem in terms your schema though and will try to do so. Do tell me if I’m coming loose.

    The following statements still confuse me:

    Saving is what is not spent from income in a defined income accounting period…. saving from income is added continuously to balance sheet equity.

    Why do you underline this–I think this is implicit in what I wrote–what is the disagreement here?

    “In a closed economy, saving equals investment.”

    Given stated axioms (“Saving is not investment”, e.g.), how am I to understand this? How does the change in spending on capital goods equal the change in balance sheet equity?

    “The best illustration of micro dissaving occurs in that of a current account deficit in the context of the world as the closed economy. The US current account deficit represents dissaving at the micro to world level, because the US is purchasing mostly consumer goods and services in excess of its own income. Because it is “not spending” a negative amount, its micro to world saving is negative, in respect of that deficit. It is dissaving. Conversely, the rest of the world is saving an amount of income equivalent to the value of its net exports to the US, because by definition that income is not available to purchase it own output. Net saving between exporters and importers is zero, and no investment has taken place relative to those (presumed consumer goods) flows.”

    This looks like a near perfect summary of the problem. It just needs to be reformulated in terms of a Keynesian AD problem where “America” equals the “public sector” and “China” equals the “private sector”.

    I conceive of an ultimate macro-level limiting sector (global macro) that can always disaggregate into two or more composite sectors that together form the aggregate macro of the initial state, and on into infinity if need be, since it is only monads all the way down.

    This cross-sectoral deficit dependency (US-China; public-private) can at any time hold for any two constituent sectors in a larger macro. One sector is a net addition to AD in the other sector, boosting total output in one an destroying total output in the other.

    The Other Sector:

    1, It *must* be present or AD remains pro-cyclical;

    2, Only the government can solve the coordination problem.

  145. vimothy writes:

    Also: Every individual sector can itself be thought of as a macro and disaggregated into composite sectors.