MMT stabilization policy — some comments & critiques
— Abba Lerner (1943)
First, I want to make clear that the critiques I’ll offer below are not intended to discredit or dismiss MMT. As I’ve said before, I think MMT offers a coherent and important perspective on fiscal and monetary issues that ought to be understood, on its own terms rather than in dismissive caricature, by anyone serious about macroeconomics. MMT is not “true”, but then no theory is “true”. We ultimately judge theories by how useful they are, both in making sense of “the data” we already know and in offering guidance for policy going forward. In my opinion MMT is one of the most useful perspectives in thinking about fiscal and monetary questions.
However, it is still just a perspective. Enthusiasts sometimes present MMT in a manner that’s too complete and hermetically sealed. While some MMT theorizing is based on “double entry accounting” or “obvious, unarguable facts”, when MMT adherents offer non-trivial conclusions, they rely upon assumptions about human behavior that are in fact contestable. I continue to place non-zero weight on theories of government insolvency that MMT-ers have persuaded me are, in a sense, incoherent. Life is complicated, and even absurd prophesies can prove self-fulfilling.
This will be a long post. I’ll discuss each of the seven points I outlined in my summary of MMT stabilization policy. Then I’ll offer some general comments. Before you continue, you should understand the point of view being examined. Please read my previous post first. Or much better yet, read Chapter 1 (Tymoigne and Wray) and Chapter 5 (Tcherneva) of A handbook of alternative monetary economics (ed. Arestis & Sawyer). These essays offer a polished, concise introduction to the MMT perspective. Then spend some time with the “mandatory” or “101” readings on Warren Mosler and Bill Mitchell’s websites.
The summary points from my previous post are repeated below in bold. New comments then follow. I am critiquing my own distillation of MMT stabilization policy, so there is the danger I have set up straw men. If I have, I apologize and look forward to being set straight in the comments. As usual, almost nothing I say will be original. Many of the points I’ll make have been made better by others, for example, in the comments to the previous post, which are extraordinarily good. At a Kauffman Foundation blogger convention last week, I discussed MMT informally but at some length with David Beckworth, Megan McArdle, Mish, and Mark Thoma. My comments will undoubtedly be informed by those conversations.
The central macroeconomic policy instrument available to governments is regulating the flow of “net financial assets” to and from the private sector. The government creates private sector assets by issuing money or bonds in exchange for current goods or services, or else for nothing at all via simple transfers. Governments destroy private sector financial assets via taxation. MMT-ers tend to view financial asset swaps, whereunder the government issues money or debt to buy financial assets already held by the private sector (“conventional monetary policy”) as second order and less effective, although they might acknowledge some impact.
While the flow of net private sector financial assets does strike me as an important and powerful tool for macroeconomic policy, it is not a uniquely effective tool. Changes in the relative price of financial assets (the object of conventional monetary policy) and in the distribution of financial assets can also powerfully affect behavior, and there are costs and benefits associated with each lever. What is the justification for focusing almost exclusively on managing the level of “net financial assets”?
A government that borrows in its own currency cannot be insolvent in the same way as private businesses. That is, such a government will never face a sharp threshold where it cannot meet promised payments, leading to a socially unanimous or even legal declaration of insolvency and an almost certain run on its liabilities.
It is unassailably true that a government cannot be forced into insolvency for want of capacity to pay in its own currency. But a government might find itself politically or institutionally unable to meet an obligation despite access to the printing press, and there might be a sharp run on government obligations even without the focal point of formal insolvency that usually occasions private sector runs. It strikes me as an open question the degree to which protection from formal insolvency protects government obligations from disruptive races to redeem. Point #7 below strikes me as stronger protection.
However, the value of money and government claims in real terms is absolutely variable. Governments do and properly should manage their flow of obligations with an eye to supporting that value, among other competing objectives (such as, especially, full employment).
I think almost no one would argue with this point.
The real value of money and government debt is not reliably related to any theory of government balance sheets. In particular, the stock of outstanding government obligations is largely irrelevant. The value of government obligations is a function of financial flows.Government claims will retain their value so long as the private and foreign sectors wish to expand their holdings of those claims at the current price level, that is so long as agents are willing to sacrifice real goods and services today to reduce their indebtedness, improve their financial position, or stimulate their export sectors. The value of government claims will come under pressure when agents, on net, seek to increase indebtedness or redeem existing claims for real goods and services.
This is a place where MMT-ers, quite rightly, call out conventional economists on adherence to dogma ill-supported by the data. Empirically, the relationship between government balance sheet quantities and either the price level or private/foreign willingness to absorb government claims is weak. Conventional economists intone seriously about our growing debt-to-GDP, and discuss solvency criteria that no one believes as though they were real. (I don’t think any sensible person believes indebted governments will ever run surpluses of present value greater than the accumulated stock of public debt. Yet that is the party line solvency criterion. [Update: See below]) Theories of the public balance sheet that have proven unreliable continue to be endorsed to avoid inconsistency in the edifice of neoclassical finance. It is true, in extreme cases, that governments that experience hyperinflations go through periods of high indebtedness relative to GDP, but what is cause and what is effect there is murky.
Macroeconomic theory is often stupid about debt. Common models impose a “no Ponzi condition” that is absurd not only for governments, but also for private firms. All firms and governments eventually end, and when they do, they usually leave substantial claims unsatisfied. Agents lend to corporations and governments not because they believe the debt will be paid down, but because they believe the almost certain eventual default or debasement of claims is unlikely to happen within their investment time horizon. In the real world, governments and corporations balance actual gains from the transfer component of increased borrowing against increased hazard that the end will come quickly and potential “distress costs”. Typically, governments and firms find these costs easy to manage as long as indebtedness grows no faster that “size” (whether measured in terms of revenue or asset values). While it is risky to “lever up” — to increase debt faster than size — many firms and governments do so successfully. We have no reliable criteria of maximum leverage even for firms, let alone for governments. Governments are special. Their core asset is their taxing power. Their liabilities, whether notionally bonds or money, are best understood as preferred equity rather than debt. They face very diffuse liquidity constraints.
All of that said, I think MMT-ers sometimes err in the opposite direction. They are right that ultimately it is flows (actual or desired) between private agents in aggregate and governments that determine the value of government obligations. But the whole purpose of balance sheet analysis, in the private sector and the public sector, is to predict future flows. That conventional theories of public balance sheets are foundationally stupid and overstate the hazards associated with large stocks of outstanding debt doesn’t invalidate the intuition that flow volatility is likely to be proportional to the outstanding stock of government claims. Suppose, because of a sunspot, private holders of government claims get nervous and try to redeem them for current goods. If the net stock of claims in private sector hands is small, it takes very little taxation to offset that flow. If the net stock of government claims is large, than the desired flows might be massive, and governments might be faced with unappetizing choices between taxation or accommodating inflation. There is little evidence that increasing the stock of government obligations, by itself, increases the likelihood that the private sector will seek (impossibly but disruptively) to divest itself of those claims. But there are undoubtedly fluctuations in the private sector’s enthusiasm for holding money and government debt, and it strikes me as implausible that the difficulty of managing those fluctuations is entirely unrelated to outstanding stock of those claims.
Also, although MMT-ers are typically regarded as “left” economists, I think they underplay the distributional costs that attend expanding the stock of government obligations. Government obligations, like all financial assets, are disproportionately held by the wealthy. If the government did not accommodate the private sector’s demand for net financial assets, preferring different policy levers to stabilize the economy, wealthy people might be forced to store wealth in the form of claims on real resources, and would have to oversee the organization of those resources into value-sustaining projects. A large stock of “risk-free” financial assets allows people wishing to carry wealth forward to shirk their duty to steward resources carefully and bear the consequences of investment failure. Thus, the availability of government obligations simultaneously degrades the quality of real investment (by disincentivizing supervision) and magnifies the distributional injustice that attends failures of aggregate investment by shifting the burden of those shocks onto risk investors and workers. In theory, governments can mitigate this injustice by careful transfers and expenditures ex post, and that might be the right policy, but in practice those who disproportionately hold existing government obligations disproportionately hold political power, and resist the issue of new obligations which might put dilute the value of existing claims. In practice, a large stock of government claims serves as the lifeboat through which prior wealth inequality carries itself into the future. Absent an accommodative stock of government obligations, recessions would be crucibles that separate the deserving from the undeserving rich, and would thin the ranks of the rich generally. Recessions should be periods that decrease inequality, but the availability of default-risk free claims whose purchasing power is politically protected inverts the dynamic.
MMT-ers are right, I think, to argue that, for fiat-money issuers who borrow in their own currency, conventional government solvency criteria are false. They are right to argue that such governments have a great deal more latitude to issue money and debt than conventional theories suggest. But that shouldn’t be taken as license to defend carelessness in the distribution of new claims, or to treat expansions of money or debt as entirely cost-free. To be fair, this is a bit of a straw man. Serious MMT-ers think about distributional issues and quality of expenditure, and don’t claim that deficits should be “carelessly” expanded. But in the heat of current policy debates, rhetoric about “deficit terrorists” and money being nothing more than spreadsheet entries unhelpfully obscures that. At its best, a deep point of MMT is that the absence of short-term fiscal constraints creates space for government to craft policy that focuses on the productivity of the real economy. If the mobilization of real resources is wise, fiscal maneuvers will be rendered sustainable ex post. If the real economy will be mismanaged or let to languish and decay, no amount of “fiscal discipline” will save us. The version of MMT that I like best is, oddly, wedded to an almost Austrian sensibility about real investment.
The “solvency” of a government is best understood as its capacity over time to manage the economy in a manner that avoids net outflows. “Net outflows” here means attempts by nongovernment actors in aggregate to redeem government paper for current goods and services.
I agree entirely. I think this is the best definition of government solvency.
The MMT-sympathetic Traders Crucible objects, however, to my use of the word “solvency” here, even with the scare quotes. After all, what currency issuing governments must concern themselves with is not insolvency per its dictionary definition (an inability to pay debts), but something quite different, a decay in the value of its claims in terms of real goods and services. Here’s TC:
[T]he impossibility of insolvency does not mean the fiat currency will have value. A government might be fully solvent even with a worthless currency… This distinction between insolvency and debasement is at the heart of MMT…
Why is the Traders Crucible going nuts…about the difference between insolvency and debasement?
Well, we can directly observe the debasement of a currency in an economy through the inflation rate. We can directly observe the process of debasement and loss of value of the currency through inflation. We cannot directly observe the risk of insolvency — it must be inferred from bond price action… the resulting process is one of guesswork, misstatements, boneheaded plans, wild specualtion, and dumbassery, because there is no way to observe the risk of insolvency directly even though it is one of the ideas that govern our spending. …[B]y removing the fear of insolvency, we can more directly observe the risk of debasement… [W]e don’t need to rely on the bond market to “give us signals” about the potential loss of access to their club to determine if we need to lower spending, or raise spending. We can just witness inflation and unemployment and make decisions on these two variables, instead of the three variables of unemployment, inflation, and insolvency… This is a much simpler task, and is perhaps the core strength of the MMT paradigm.
This is an important point, but contestable. We know with some confidence that the threat of traditional insolvency can lead to powerful and unpredictable runs and a lot of turbulence in the value of private claims. I’m glad to concede that, at the margin, absence of a sharp solvency threshold reduces the likelihood of such events. But does the lack of a sharp solvency threshold eliminate the possibility of sudden stops, Wile E. Coyote moments, etc? Can we be confident that, absent the danger of outright default, any debasement of fiat claims would take the form of an observable spiral, which would start slowly and thereby offer time to apply a policy antidote? Would we in fact observe and recognize the signs, and would they be different than, for example, a 500% increase in the price of gold in the span of a few years and recurring bouts of commodity inflation? Are employment pressure and labor costs the sole true and perfectly reliable indicators of debasement hazard?
One can make a strong case that increases in labor costs are in fact the sine qua non of uncontrollable inflation, that absent labor income to “ratify” price rises, inflation in inherently self-limiting. But you can make other cases too. Perhaps transfers and deficit spending can substitute for wage power, bidding up commodity prices and the capital share of income even while wages are held back by the reserve army of the unemployed. I’m not sure about any of these stories. But my experience as a trader in capital markets makes me wary of accounts that suggest sharp swoons in the price of any asset cannot happen, or would definitely be preceded by warning signs that would permit one to get out early.
So, I’ll to acknowledge TC’s objection as important and potentially valid, but defend my positing of an MMT “solvency” constraint, at least with scare quotes in place. I don’t think it’s reasonable for MMT-ers or anybody else to write off the possibility of sharp and unexpected changes in the value of a fiat currency. The possibility is dangerous enough that it should focus the mind in a precautionary way. If MMT policy advice is to be taken seriously, it must offer a some assurance of safety against that scenario. The absence of formal default hazard provides some assurance, but without Point #7 as a backstop, not enough.
Avoiding net outflows is easy in times like the present, when i) low quality and difficult to service debt in the private sector leaves many agents eager to reduce indebtedness and increase their holdings of financial assets; ii) there has been little inflation or devaluation in the recent past; and iii) resource utilization is slack, as evidenced especially by high unemployment. Avoiding net outflows is more difficult when private sector agents’ balance sheets are healthy, or when agents come to expect inflation or devaluation, or when real resources (especially humans) are fully employed.
I think this point is unobjectionable.
However, a sovereign government can always create demand for its money and debt via its coercive ability to tax. That is, if optimistic agents with strong balance sheets start up a spending spree, or if gold bugs fearful of devaluation ditch government paper for commodities, a government can reverse those flows by forcing private agents to surrender real goods and services for the money they will owe in taxes.
On the one hand, I consider this point is one of MMT’s deepest insights, and its secret weapon. So long as a government’s taxing power is strong, so long as it is capable of persuading individuals to surrender highly valued real goods and services for the ability to escape liabilities imposed by fiat, exercise of that taxing power creates a floor beneath which the value of a currency, in real goods and services, cannot fall.
However, relying too overtly on taxation to give value to a currency strikes me as dangerous and potentially counterproductive. A government’s taxing power is limited and socially costly. Governments must maintain a patina of legitimacy so that people pay taxes “voluntarily” or else they must intrusively or even brutally force compliance. In a decent society, it’s perfectly possible that governments will find it politically impossible to tax at the level consistent with price stability goals. A wise, MMT-savvy government would try very hard to regulate the issue of government obligations over time in a way that avoids the need for sharp fiscal shifts in order to stabilize the price level. Avoiding the need for sharp contractions later on might imply slower issue of obligations than would be short-term optimal during recessions. But once you acknowledge this kind of forward-looking dynamic, MMT starts to sound very conventional. We start having to trade off the short-term benefits of fiscal demand stimulation with long-term “exit costs”.
Two other points are worth making here:
Even though, in principal, taxation could be used to regulate economic activity and put value underneath the currency, the institutions that would be necessary to do this successfully are simply not in place in existing democratic polities.
Within the MMT community, smart people have given a great deal of thought to institutional forms under which which fiscal policy might be used to regulate activity. As far as I know, they have mostly converged upon the institution of a “job guarantee (JG)” or an “employer of last resort (ELR)”, whereunder the size and wage of a “buffer stock” of public labor would become the economic instrument of macro stabilization. This is an ambitious idea, both politically and technically. Not only must one develop appropriate policies for stabilizing the economy on the fiscal side (i.e. the equivalent of a Taylor Rule for ELR wage levels), but one must also plan and implement real-world projects for a variable-sized pool of (hopefully) transient workers. These projects should usefully employ and develop the productive capacity of ELR participants, while remaining distinct from and and not interfering with the ordinary private and public sector workforces. (As I understand the proposal, ELR employees would be distinct from other public employees, in that they’d be paid a standard, low but livable, package of wages and benefits. ELR employment would always be viewed as a backstop that individuals would be encouraged to transition out of, rather than as permanent employment.)
I’m interested in and sympathetic to the project of designing a government-guaranteed full employment policy that would be complementary to a vibrant private sector and that would anchor rather than disrupt macroeconomic stability goals. But however richly MMT-ers have outlined such an institution in theory, we are very far from implementing such a thing in practice. MMT-ers participate actively in current fiscal policy debates, arguing that “sovereign” governments have sufficient space to let fiscal concerns be secondary to resource utilization goals given their power to tax. Yet the power to tax is next to worthless if we do not have well understood and broadly legitimate means of exercising it in a timely manner.
Taking a page from status quo macro management — that is, from the world of central banks — the least costly way to meet macro stabilization goals is to maintain credible expectations among the general public that tax policy will in fact be managed with sufficient dexterity and force that those stabilization goals are rarely tested. Existing fiscal institutions are mostly quasidemocratic legislatures that act in sporadic and highly politicized bursts. Their policy ventures typically mix interventions on the liability side of the public balance sheet with ad hoc changes to programs on the asset side that are often difficult to reverse. These institutions seem poorly suited to the task of credibly managing expectations and ensuring, in high-frequency real time, an appropriate fiscal stance. Promoting fiscal license in actual policy while the institutions that would render such license sustainable do not exist strikes me as reckless. When participating in practical debates about fiscal policy, it would be better if MMT-ers would bundle their support for “fiscally loose” stabilization policy with advocacy of institutional changes that could be plausibly implemented in time to matter and that could ensure support of the value of government claims, should that become necessary.
Some MMT-ers (Warren Mosler and Winterspeak come to mind) have proposed less ambitious institutions than an employer of last resort program, specifically using the level of existing payroll taxes as the instrument of discretionary macro policy. A government can stimulate by reducing the level of payroll taxes (and thereby increasing the flow of net financial assets to the public sector in a manner that directly encourages job formation), and could fight inflation by raising payroll taxes, rather directly reducing wages and putting pressure on employment. Macro policy by unemployment is detestable, despite its long, proud tradition at the Federal Reserve. If it can be made practical, I’d much rather we work out an effective ELR program. But ELR is not an achievable option in the time frame of the current business cycle. Delegating management of the level of the payroll tax to a “technocratic, independent” institution, whether the existing central bank or some new entity, is practically achievable on a short time frame (although the politics would be rough). Perhaps there are better easy-to-implement means of conducting credible, high-frequency macro policy. I’ve no special attachment to payroll taxes as an instrument. (I’d prefer that we use transfers as an instrument.) Whatever the specifics, relying on ad hoc interventions by Congress to thread the needle between inflation and underemployment strikes me as unlikely to work out.
This is a technical point that would usually apply mostly to small, open economies, but that arguably applies to the United States today. Taxation can support the value of government claims, when priced in domestically produced goods and services. Taxation cannot support the foreign exchange value of a fiat currency, except to the degree that foreigners desire to purchase domestically produced goods and require expensive domestic currency to do so. A country that runs a large current account deficit owing to decisions by foreign governments to accumulate its currency and that faces competitive export markets cannot rely on taxation to support its currency, should foreign governments revise their policy of accumulation. For a country like the United States which is structurally “short” tradables, one may view the possibility of a difficult-to-counter fall in the value of the currency as a good thing or a bad thing. People like Dean Baker and Paul Krugman argue that a weaker dollar is exactly what the US needs to eliminate the structural gap in tradables production and spur domestic demand. People like Warren Mosler argue that a very weak dollar would be a bad thing, an adverse terms-of-trade shock and a loss of opportunity to trade cheap nominal claims for valuable real resources. Regardless of how you view the event, the taxing power of the government will not be able to undo it.
Therefore, a government’s “solvency constraint” is not a function of any accounting relationship or theories about the present value of future surpluses. A government’s solvency constraint ultimately lies in its political capacity to levy and and enforce the payment of taxes.
I think this is true, a deep and powerful way to think about public finance. Note that a government’s “political capacity to levy and and enforce payment of taxes” depends first and foremost on the quality of the real economy it superintends. The value that a government is capable of taxing if necessary to sustain the value of its obligations increases with the value produced overall. A government that wishes to be solvent should first and foremost interact with the polity in a manner that promotes productivity. Secondly, the political capacity to levy taxes depends upon either the legitimacy of or the coercive power of the state. A government that wishes to sustain the value of its obligations must either gain the consent of those it would tax or maintain an infrastructure of compulsion. In theory, either choice could be effective, although along with the libertarians, I like to imagine excessively coercive regimes are inconsistent with overall productivity, so that legitimacy is a government’s best bet. The two strategies are not mutually exclusive — a government could be sufficiently legitimate as to be capable of taxing some fraction of the population without resistance and sufficiently coercive as to force the other fraction to pay up. That probably describes the best we can hope for in real governments.
I’ll end with a few miscellaneous comments:
I’d like to see more attention paid to quality-of-expenditure concerns. That is, if a stable economy requires continuing government deficits to accommodate growing private sector’s demand for financial surplus, then the government must actually make choices about how to spend or transfer money into the economy. These choices will undoubtedly shape the evolution of the real economy, for better or for worse. Should we rely on legislators to make direct public investment choices? Should we put funds in the hands of individuals and then allow consumer preferences and private capital markets to shape the economy? How? Via tax cuts? A job guarantee? Direct transfers? Perhaps the government should delegate management of public funds to financial intermediaries, and rely upon banking professionals to find high value investments? While MMT focuses mostly on the liability side of the public balance sheet, many critics fear that ever increasing public outlays imply increased centralization of economic decisionmaking that will lead to low quality choices. Whether that is true depends entirely on institutional and political choices. These concerns can be and should be specifically addressed.
MMT-ers sometimes blur the distinction between “private sector net savings”, which is necessarily backed by public sector deficits or an external surplus, and household savings, which need not be. In doing so, MMT-ers rhetorically attach the positive normative valence associated with “saving” to deficit spending by government. This is dirty pool, and counterproductive. The vast majority of household savings is and ought to be backed by claims on real investment, mediated by the liabilities (debt and equity) of firms. There is no need whatsoever for governments to run deficits to support household saving. When household savings increases, an offsetting negative financial position among firms represents increase in the amount or value of invested assets, and is usually a good thing. Household savings is mostly a proxy for real investment, while “private sector net financial assets” refers to a mutual insurance program arranged by the state. It is a category error to confuse the two. Yet in online debates, the confusion is frequent. Saving backed by new investment requires no accommodation by the state. It discredits MMT when enthusiasts claim otherwise, sometimes quite aggressively and inevitably punctuated by the phrase “to the penny”.
In general, the MMT community would be well served by adopting a more civil and patient tone when communicating its ideas. I’ve had several conversations with people who have proved quite open to the substance, but who cringe at the name MMT, having been attacked and ridiculed by MMT proponents after making some ordinary and conventional point. Much of what is great about MMT is that it persuasively challenges a lot of ordinary and conventional views. But people who cling to those views, even famous economists who perhaps “ought to” know better, are mostly smart people who simply have not yet been persuaded. Neither ridicule nor patronizing lectures are likely to help.
My complaint is a bit unfair. The MMT community has been sinned against far more than it has sinned, especially within the economics profession. Whether you ultimately agree with them or not, the MMT-ers have developed a compelling perspective and have done a lot of quality work that has pretty much been ignored by the high-prestige mainstream. But a sense of grievance may be legitimate and still be counterproductive.
The internet is a fractious place. Many MMT-ers are civil and patient, and devote enormous energy to carefully and respectfully explaining their views. There’s no way to police other peoples’ manners. Still, even by the standards of the blogosphere, MMT-ers have a reputation as an unusually prickly bunch. That might not be helpful in terms of gaining broader acceptance of the ideas.
Anyway, that was a lot. I hope that it’s not entirely useless. Despite my complaints and critiques, learning about MMT has added enormously to my thinking about economics. In practical terms, I think that MMT offers the most promising toolkit for crafting a desperately needed replacement of status quo central banking.
With that, I’ll shut up. Feel free to be as nasty as you wanna be in the comments.
(I don’t think any sensible person believes indebted governments will ever run surpluses of present value greater than the accumulated stock of public debt. Yet that is the party line solvency criterion.)
No, that’s not the party line. In fact, the party line would say that is impossible. The party line says that the *expected* present value of *primary* surpluses (plus seigniorage, if that’s not included) is *equal to* the existing debt. That party line is perfectly consistent, in a growing economy, or in an economy with positive inflation, with perpetual deficits, as conventionally measured (i.e. non-primary, to include interest on debt). Basically, if Nominal GDP is growing at rate n%, then a government can run a conventional deficit of n% times the outstanding debt forever. (Because that means the debt will grow at the same rate as NGDP, so the debt/NGDP ratio stays constant over time.)
Rowe is right to call me out — my wording was sloppy. It was especially unforgivable that, in characterizing the conventional intertemporal goverment budget constraint, I omitted the modifier “primary” before surpluses.
But I was only sloppy, not mistaken. Rowe suggests that, when accurately characterized, the conventional intertemporal government budget constraint is something that sensible people actually believe. I cede no ground at all on this point. Rowe himself does not believe it. He gives himself a partial way out just in the part quoted above, when he writes “plus seigniorage, if that’s not included”. In most macro models, it is not included. There are two ways that you can incorporate seigniorage:
You can treat seigniorage as a cost expected by borrowers and holders of money, in which case it is not disruptive to conventional models. It is equivalent to reducing the interest rate and taxing the value of real money balances, if those are in your model.
You can treat seigniorage as a form of sporadic default. That is, you can claim that at some point the government will simply write down the real value of accumulated nominal debt, in practice by allowing debt or money growth without sufficient yield to prevent an increase in the price level.
The second view is disruptive of conventional models. Rowe may argue that we account for everything by the use of expectations rather than certain values in conventional models. That could be true, but broadly, it’s not. There are some models of default inspired by third-world debt crises, but outside of that, explosive growth in the price level and/or default are modeled are presumed or constrained not to occur. Monetary policy models are a partial exception, in that they derive conditions under which the constraint would hold, which then come to guide central bank policy. But in the limit, under standard equilibria in a decent country, it is assumed or derived that real primary surpluses will (in expectation) be generated in sufficient quantity to offset the real existing stock of debt. I view that as a very unlikely characterization for many existing governments.
Suppose that we do include the possibility of default. What happens? Rowe and I agree furiously on this:
Suppose there’s a 1% chance every year that a firm of government will disappear and default totally on its debt. The probability of default approaches 100% in the limit, going forward. But a risk-neutral investor will happily hold the debt with a 1% risk premium on the yield.
But what does this do to the intemporal government budget constraint? There is no constraint whatever in this characterization. Suppose that the government offers a 2% premium over investors’ cost of borrowing, and the probability of default is exogenously 1%. Then investors borrow and invest without limit! Obviously, we need to “close the model” somehow to make things realistic, but there are lots of ways to do so that violate any ordinary interpretation of the conventional intemporal government budget constraint.
I claim that realistic models, which incorporate consumers who face liquidity constraints and idiosyncratic risk in an economy subject to systematic risks of production shortfalls, do not conform to an intertemporal government budget constraint of remotely the conventional form. I’ve already described half of such a model here. I should add the other half, and write the math down in a way that economists can understand. The key insight is one that both quasimonetarists like Rowe and MMT-ers accept but rarely state explicitly — much of our motivation in “lending to the government” is not to capture growth, bit to self-insure against idiosyncratic risk. There is nothing novel about this — conventional treatments of the permanent income hypothesis characterize the conditions under which individuals will engage in precautionary saving. Redemption of precautionary savings in the form of money or government debt usually works not by government provision of goods and services when an individual faces shortfalls, but by virtue of transfers of real goods and services to those who face shortfalls from those experiencing surpluses. In other words, money and government debt are the medium by which we conduct a mutual insurance program. The stock of government debt then grows as a function of determinants of precautionary savings, which include income, but also risk preferences, the idiosyncratic risk that agents face, and the degree to which borrowing constraints bind. In all periods where the government does not default, participating in this insurance program is straightforwardly beneficial. The risk of rare government defaults, due to systematic shocks, may be insufficient to offset the benefits of participation in the mutual insurance program, and government debt need only be the least risky available medium, conditional on its use for insurance purposes, to rationally attract insurance-motivated lending. Under some circumstances (satiable preferences, steep reduction of incomes post-government-default combined with rationing based on prior savings, little relationship between the scale of insurance borrowing and the likelihood of default), I claim, it is reasonable to expect government debt to grow without bound. I consider these circumstances to be pretty realistic.