Interpreting Modern Monetary Theory
By Jeffrey Rogers Hummel
Modern Monetary Theory (MMT), a non-mainstream economic doctrine, has recently emerged from popular and academic obscurity to become a hot topic. Enthusiastically embraced by assorted progressive politicians, MMT allegedly demonstrates that such expansive government programs as the Green New Deal will not impose significant financial burdens on government. Economists critical of the theory range widely across the ideological spectrum, from George Selgin of the Cato Institute and Scott Sumner of the Mercatus Center all the way to New Keynesian Paul Krugman, and Post-Keynesian Thomas Palley, both of whom otherwise sympathize with progressive programs. Yet many have found the expositions and defenses of MMT to be unclear, obscure, and even evasive.
If we focus solely on MMT’s essential claims about money, distinct from any associated policy proposals, it is neither new nor modern. It simply justifies funding government expenditures by issuing fiat money, which, of course, all economists have long been aware is possible. MMT then attempts to downplay the potential inflationary impact of such financing with manipulations of the government and central-bank balance sheets. But it merely puts the standard analysis into different boxes. Indeed, sophisticated advocates of MMT recognize that inflation can result from substantial increases in government expenditures unless one of three conditions holds: (1) there is significant unemployment in the economy; (2) government uses its taxing power to control inflation; or (3) the banking system somehow counteracts the government’s monetary expansion. Yet as we will see, advocates of MMT are overly optimistic about the efficacy of these three potential constraints on inflation, either overlooking their limitations, being unclear about how exactly they would be implemented, or grounding them in dubious economic doctrines.
With respect to the first constraint about unemployment, Selgin has pointed out that “reasonable people can disagree about whether we still have some way to go before achieving full employment,” but “it’s 2019, not 1933; and the (labor) unemployment rate now hovers around 4 percent, rather than above 20 percent.”1 This explains why MMT is usually coupled with advocacy of job training and guarantee programs in which the government becomes the employer of last resort. Supposedly the resulting increase in output from increasing the labor-force participation rate would dampen any price increase. Beyond this simplistic knife-edge, on-off view of inflation, with an almost stationary Phillips Curve trade-off with unemployment, advocates of this scheme have been cavalier about such details as how many who are currently not in the labor force actually want to take government jobs and how such an extensive government program might disrupt private labor markets.2
Using Taxation to Limit Inflation
The most unusual feature of MMT is the second possible constraint: the claim that government taxation can control inflation. As Jim Kavanagh, defending MMT in Counterpunch, asserts, “taxes are that portion of the money… which the government withdraws for reasons that have nothing to do with needing to collect money to spend. Economically, they are for controlling inflation.” One of the most prominent of MMT theoreticians, L. Randall Wray of the Levy Institute of Economic Research at Bard College, admits that he is “skeptical of use of discretionary tax hikes to fight inflation,” but then concedes that “if there were a prolonged stretch of inflation we would—of course—recommend pro-actively raising taxes and/or reducing spending.” 3
Yes, government can always use taxes as an alternative to issuing money for financing expenditures. But this does not seem to be what most MMT proponents have in mind. Instead, they appear to claim that after newly issued money has paid for a government program, the government can use its taxing power to pull the money out of the economy. But what MMT proponents are never entirely explicit about is that the government can keep the newly spent money out of the economy only if it doesn’t simply turn around and spend the money it collects. To understand how this might be possible, we need to the look at the relationship between the Federal Reserve System and the Treasury Department. Both are government agencies that not only receive and disburse money but also create it. The Treasury issues coins, whereas the Fed issues Federal Reserve notes and its near equivalent, bank reserves held on deposit at the Fed. These are the forms of government fiat money, constituting what economists call the monetary base.
These two government agencies also hold balances of each other’s money. When the Treasury collects money through taxes or fees, it deposits that money at an account it holds at the Fed.4 If the money then just sits there, reissued neither by the Treasury nor by the Fed, it is indeed out of circulation, effectively burned. After all, the Treasury and the Fed are just two agencies of the same government, so how much they hold of each other’s money is just an accounting convention. But in fact, the money rarely sits for long in the Treasury balances at the Fed. The Treasury manages its Fed bank account roughly the same way the general public handles its bank accounts. Money is constantly flowing in and out of Treasury balances at the Fed, disbursed though checks the Treasury writes against the Fed. The total size of this account usually fluctuates within some desired range, and currently is about $350 billion.
Only once has the Treasury let its balances at the Fed pile up without making expenditures. That was during the financial crisis of 2008, when the Treasury helped the Fed engage in quantitative easing by issuing securities to borrow money from the general public. Yet the money still went right back into circulation, because the Fed used the money from the growing Treasury deposit to purchase financial assets or make loans to financial institutions. In fact, the Fed, like private banks, normally reissues most of the money that has been deposited with it, including the amount regularly in Treasury balances. Thus if the Treasury were to allow increasing deposits at the Fed to lie entirely idle, making no additional expenditures itself, the Fed would likely still put the bulk of that money back into the economy.
Although the Fed makes loans to banks and other financial institutions, and during the crisis, bought mortgage-backed securities, it normally issues money by purchasing Treasury securities. But legally it can make those purchases only from private parties, not directly from the Treasury, so the money is still being injected into the economy. Even if the Fed gained the authority to purchase Treasury securities from the Treasury, the money would only end up in Treasury deposits at the Fed, to again be re-spent by either the Treasury or the Fed. Even taxes that generate government surpluses rarely decrease the monetary base. The last time this happened in the United States was after the Civil War, when surpluses were used to retire Greenback paper currency. On those few occasions in recent times that the U.S. Treasury has run a surplus, the resulting revenue was used to make the final payment on maturing Treasury securities or to buy back outstanding Treasury securities. Either way, the money went back into the hands of the public.
In short, the U.S. government for over a century has never employed taxes to reduce or regulate the size of the monetary base. Of course, the government could do so. If the Treasury increased tax revenues without increasing expenditures, the Fed could allow Treasury deposits to simply accumulate. This would permanently take base money out of the economy. But this increasing Fed liability would have to be offset somewhere on the Fed’s balance sheet. One way to offset is just to let any increase in the Fed’s Treasury liabilities reduce the Fed’s capital account, in which case the total size of balance sheet would remain constant. But reducing the size of the Fed’s capital account, currently only $40 billion, has limits, unless the Fed adopts the accounting fiction of a negative capital account.
An alternative is for the Fed to offset an increase in its liabilities with a newly invented asset, equal in amount to the increase in Treasury deposits, that it could label Treasury currency or Reserves. (In fact, the Fed already does this with an asset labeled “Coin” for the small amount of Treasury money that regularly circulates through its vaults.) Then the Fed’s balance sheet would increase on both sides by the amount of accumulating tax revenue. Either of these balance sheet expedients could permanently pull money out of circulation, reducing the monetary base through taxation. But proponents of MMT tend to dismiss these accounting intricacies as the government’s “self-imposed constraints” that could easily be swept away. They remain vague about what new institutional arrangements they have in mind.
MMT stresses that the private sector would have no fiat money to spend if government had not issued sufficient quantities in the first place. This is obviously true, but irrelevant: those injections occurred mostly in the past. Indeed, the clearest cases are early fiat moneys, such as the Continental currency of the American Revolution, the Confederacy’s paper money during the Civil War, and the Greenbacks issued by the Union during that conflict. These are cases in which governments printed large quantities of money and directly spent it, generating severe inflation. Most modern fiat money, in contrast, emerged through a more subtle evolution, in which the government’s central bank issued banknotes redeemable at a fixed rate for some commodity, typically gold, with both items serving as money until the government broke the link with the commodity. This method sometimes entailed a more restrained and drawn out growth of the monetary base, with any severe inflation offset by economic growth.
Devotees of MMT, who belabor the distinction between stocks and flows, should be especially conscious that the large existing stock of what they call “sovereign currency” has little relevance to the inflationary impact of a dramatic increase in its flow. Admittedly, the U.S. government has enlarged the monetary base over time and will probably continue to do so in the future. And as long as it does so gradually, inflation will be mild. Higher levels of taxation could even slightly increase the amount of money sitting inert in the Fed’s and Treasury’s balance sheets. But that factor alone will hardly neutralize the inflationary impact of the monetary expansions required to finance the massive government expenditures that some progressives are pushing. Only pulling the new money back out of economy will do the job, unless of course new taxes are substituting for money creation.
Another possible interpretation of the claim that taxes can control inflation might focus on money’s velocity rather than its quantity. If higher taxes somehow reduce the velocity of circulation, they would dampen inflation. One controversial but mainstream approach in which taxes affect velocity is the Fiscal Theory of the Price Level (FTPL), whose most prolific champion is John Cochrane of the Hoover Institution. This theory even shares MMT’s insistence on the idea that taxes anchor the purchasing power of fiat money. But the FTPL’s policy implications are polar opposites. It does have taxation reducing money’s velocity through anticipated future surpluses. Yet advocates of MMT are not fans of surpluses. Nor are their tax proposals likely to inadvertently generate surpluses. Moreover, for reasons explored below, their theory has no room for expectations about future taxes increasing the public’s cash balances and thus bringing down money’s velocity.
Taxation is not the only possibly way to take money out of circulation. Government also borrows money from the public through the issue of Treasury securities. But as we’ve seen, Treasury borrowing likewise ends up deposited at the Fed, requiring the same manipulations explained above to permanently withdraw the money from the economy. Moreover, by issuing securities, the government has committed itself to paying interest in the future. It conceivably might get around that “self-imposed constraint” by repudiating its debt, but so far no advocate of MMT wants to take that route for either controlling inflation or eliminating the national debt. Indeed they don’t think that a growing national debt is likely to be a problem.
Some MMT proponents, including Kavanagh and Bill Mitchell, have suggested eliminating the government’s debt by having the Fed buy up all Treasury securities.5 But this is probably just a rhetorical flourish on their part. The asset side of the Fed’s balance sheet currently has a little over $2 trillion of Treasury securities. If the Fed buys the remaining $14 trillion outstanding (not counting the government trust funds, which is money the government does owe itself), proponents are well aware that doing so merely converts the form of the debt from interest earning Treasury securities to interest earning bank reserves. The Treasury would then pay interest on the bonds to the Fed, most of which the Fed would in turn pay out as interest to banks, except for a small residual the Fed remits to the Treasury.
More important, converting the entire national debt into bank reserves would cause an enormous increase in the monetary base, which, even at the peak of quantitative easing, reached only slightly over $4 trillion. Such a near quadrupling of the monetary base hardly seems a recipe for curbing inflation, unless the Treasury simultaneously pulled all the new money out of circulation with taxes confined exclusively to flooding its deposits at the Fed. And the surplus revenue such taxes would have to generate would have to almost equal the entire $14 trillion increase in the base. But a final proposition in the MMT canon is that increases in the monetary base, even without interest on reserves, do not necessarily generate inflation at all. This conclusion is based on peculiar views about how the banking system works; that brings us to their third posited constraint on inflation.
Using the Banking System to Limit Inflation
The claim that banks can constrain inflation is the most complex and convoluted of MMT’s arguments. What gives this view some credence is what happened when the Fed generated an unprecedented increase in the monetary base through its quantitative easing during the financial crisis. Inflation did not take off, as many predicted, but it was because the Fed simultaneously started paying interest on the reserves that banks held on deposit at the Fed. Advocates of MMT, to their credit, recognized that paying interest on reserves was a game changer.6 The interest rate on reserves, if high enough, can induce banks to hold reserves rather than make loans that would otherwise increase deposits. This indeed permits the Fed to expand the monetary base while limiting the impact on the total money supply. As a result, still today the aggregate reserve ratio for checking accounts is well over 100 percent, compared with around 10 percent prior to the crisis. Although many mainstream monetary economists are likewise aware of the profound implications of paying interest on reserves, that realization has not fully penetrated through to politicians, journalists, and the general public.
MMT, however, goes further. It contends that, whether the Fed pays interest on reserves or not, the government’s management of the monetary base does not translate into control over broader monetary measures. MMT derives this claim from Post-Keynesian economics, a related and older heterodox school of thought. More interventionist than mainstream New Keynesians, Post-Keynesians are quite diverse in their views, and not all proponents of MMT accept Post-Keynesianism fully. But many accept one of the core Post-Keynesian tenets: that the total quantity of money circulating among the general public is almost completely endogenous. By “endogenous” they mean beyond the discretionary control of government, and the money supply they have in mind consists of currency in circulation and deposits held at private banks. Their conclusion hinges on the Post-Keynesian approach to interest rates.
Post-Keynesians reject the standard analysis of what determines interest rates. They, along with advocates of MMT, deny that a monetary economy has a natural real rate of interest whose equilibrium level is determined by the supply of savings and demand for investment in the loanable funds market. Stephanie Kelton, an economics professor at Stony Brook University who embraces MMT, writes: “evidence suggests that interest rates don’t matter much at all when it comes to private investment.” Instead, investment decisions are “forward-looking, heavily influenced by ‘animal spirits,’ and overwhelmingly dependent on the state of profit expectations.”7 Post-Keynesians hold that saving, on the other hand, is primarily a function of people’s income, and not much affected by interest rates either. In the jargon of economists, they claim that saving is interest-inelastic.
As a result, Post-Keynesians and MMT advocates conclude that the loan market utterly fails to equilibrate planned saving and planned investment. Notice how this diverges from even the views of New Keynesians, who hold that that the interest rate affects how much private saving will be invested versus how much will be simply hoarded in cash balances or bank reserves. MMT attaches little importance to the effect of interest rate changes on investment or spending more generally. As Wray puts it, “when liquidity preference is high, there may be no rate of interest that will induce investment in illiquid capital.”8 Using MMT terminology, the economy exhibits no “intertemporal coordination.” Current saving cannot be depended upon to bring about an increase in capital goods that will increase future income. Thus, if too little spending causes a recession, only government borrowing can extract money from the idle cash balances of the general public and thereby increase spending.
Moreover, MMT emphasizes that government deficits increase the net financial assets of the private sector. Financial assets that arise between parties within the private sector, by contrast, are exactly offset by financial liabilities and have no net effect on the public’s financial assets. Without deficits providing the public with Treasury securities, the only government liability held by the private sector would be fiat money. Thus deficits increase the public’s total financial wealth, which allegedly increases income and stimulates spending. Even if deficits crowd out financial claims between private parties (which MMT finds unlikely), this does not diminish the deficits’ contribution to total wealth, which in turn spurs spending and leads to greater real investment. This is how MMT arrives at its counterintuitive conclusion that deficits can actually lower interest rates, a conclusion that has baffled critics like Sumner and Krugman.9 In essence, MMT treats government debt and government money as very close substitutes, more or less interchangeable. Therefore expanding either of them can have roughly similar impacts on the economy.
Because exponents of MMT consider increased money and increased debt as complementary ways to increase aggregate demand, they see little need to worry about the size of the government debt. Given their insistence that the government has no budget constraint, they are oblivious to the possibility that expectations about increased future taxes can affect perceptions of wealth, either as proposed in the FTPL or through even a partial working of Ricardian Equivalence.10 “Bond illusion,” in which government debt is considered net wealth, reigns supreme. The only expectations that have a major role in MMT’s macroeconomic theory are expectations about future profits, which are erratic due to the fundamental uncertainty that plagues the economy. MMT even challenges the empirical significance of the Fisher effect, in which expectations about future inflation cause a divergence between nominal and real interest rates. Government policy, in their view, should therefore focus exclusively on nominal interest rates.
How do all these assumptions lead to an endogenous money supply? Well, if private investment is a slave to animal spirits, so is bank lending. Advocates of MMT argue that deposits are not the source of bank lending but the other way around: bank lending creates deposits. Thus, the supply of bank loans is entirely demand driven, and that demand determines the amount of money banks create in the form of deposits. The defenders of this view realize that banks do wish to hold some desired level of reserves for clearing purposes and to satisfy customer demand for cash. And this desired level can change. But reserves, they claim, place no limit on bank landing. Nor do increases in total reserves create any impetus for bank lending.11
Changes in total reserves will, therefore, not affect how much money banks create through deposits. If the central bank should create more reserves than banks wish to hold, the banks will attempt to lend the excess reserves to other banks in the overnight lending market, which, for the United States, is the Federal funds market. This drives the Federal funds rate below the Fed’s target. To keep that rate on target, the Fed will have to reduce reserves with open market sales of Treasury securities, which automatically reduces the monetary base. Similarly, if banks wish to hold more reserves, they will try to borrow them from other banks, raising the Federal funds rate and forcing the Fed to increase the monetary base. This assumed endogeneity of bank lending thus makes both the total money supply and the monetary base endogenous. As Wray puts it, “causation must run from loans to deposits and then to reserves.”12 If the government were to impose a 100-percent reserve requirement, essentially converting private bank deposits into indirect central bank deposits, that would merely remove one link in this inverted causation.
Every step in this reasoning relies on a belief that the underlying risk-free nominal interest rate is not a market phenomenon but is ultimately arbitrary. In a reply to Krugman, Kelton asserts that “the Fed can pursue any rate policy it desires.” Wray has even proposed that central banks should “set the overnight rate at zero, and keep it there.”13 This will either eliminate any interest on short-term government debt or make it negligible, bringing close to consummation the union of base money and government debt into near perfect substitutes.14 The distinction between monetary policy and fiscal policy will miraculously vanish. Although longer-term and less-liquid financial assets and also real assets will still enjoy positive returns, profits, expectations of future profits, and liquidity will solely determine the spreads between various financial and real assets. These “interest rates (and thus asset prices) adjust to ensure,” in Wray’s words, consistency with the public’s “portfolio preferences.”15
To be sure, MMT concedes that the ability of an endogenous money supply to constrain inflation has limits. Government spending can still run up against the scarcity of real resources. Although some progressives invoking MMT seem unaware of this, Wray readily acknowledges that “just because the government can afford to spend does not mean government ought to spend more.” Government “must weigh the consequences in terms of withdrawing resources from other (perhaps more desirable) uses, as well as possible impacts on prices and exchange rates.”16 This returns us full circle to the first MMT constraint on inflation: the size of the “buffer stock” (Wray’s term) of unemployed labor that can be put to work.
The Topsy-Turvy World of MMT
There you have the topsy-turvy world of MMT. With accounting games, advocates of MMT attempt to reverse the roles of the government treasury and the central bank. They believe that the Treasury should control inflation and the Fed should finance government expenses. One of the most emphatic assertions of MMT, to quote Wray, is “taxes are not needed to ‘pay for’ government spending.”17 Taxes are needed only to make sure people accept fiat money and, if necessary, to keep inflation in check. And because both the treasury and central bank are government institutions, there is some truth to the idea that both institutions have dual roles. But as many others have pointed out, MMT theorists have yet to address or even consider the enormous public-choice problems that could hinder how their desired role reversal might function in practice.18
Equally important, critical parts of MMT’s edifice are built on Post-Keynesian foundations. As Kelton and Wray, along with Scott Fullwiler proclaim: “We have never tried to separate our ‘MMT’ approach from the heterodox tradition we share with Post Keynesians, Institutionalists and others. We have tried to extend that tradition.”19 A comprehensive and extensive critique of the Post-Keynesian paradigm is beyond the scope of this article. But if you strip away Post-Keynesian precepts, much of MMT’s edifice collapses, taking down many of its policy proposals with it.
 Jim Kavanagh, “Behind the Money Curtain: A Left Take on Taxes, Spending and Modern Monetary Theory,” Counterpunch (January 22, 2018); L. Randall Wray, “Response to Doug Henwood’s Trolling in Jacobin,” New Economic Perspectives ( February 25, 2018).
 I’ve omitted two complicating details that have little effect: (a) The Treasury holds a small amount of the revenue it collects in the form of actual Federal Reserve notes (currently about $300 million) rather than as deposits at the Fed. (b) Tax revenues are temporarily lodged in Treasury deposits at private banks, in what are called “Tax and Loan Accounts,” before being transferred to the “Treasury General Account” at the Fed. This adds only a rest stop, creating a short lag as tax revenues flow toward expenditures. For more on this aspect of government finance, see George Selgin, “On Empty Purses and MMT Rhetoric,” Alt-M (March 5, 2019).
 Kavanagh, “Behind the Money Curtain”; and Bill Mitchell, “The US Government Can Buy as Much of Its Own Debt as It Chooses,” Bill Mitchell—Modern Monetary Theory (August 22, 2013).
 Stephanie Kelton, “Paul Krugman Asked Me About Modern Monetary Theory. Here Are 4 Answers,” Bloomberg Opinion (March 1, 2019).
 Ricardian Equivalence is the idea that when the government increases its debt, the country’s residents will anticipate higher future taxes to pay the interest on the debt and the principal, and, therefore, will increase their saving.
 One of the most persistent critics of the MMT view of bank behavior is Julien Noizet, “A Response to Scott Fullwiler on MMT banking theory,” Spontaneous Finance (January 14, 2014). See also the reprint of one his other posts and links to many more at Noizet, “The Problems with MMT-Derived Banking Theory,” Alt-M (March 15, 2019).
 Wray, “The Endogenous Money Approach.”
 Kelton, “Paul Krugman Asked Me About Modern Monetary Theory. Here Are 4 Answers”; Wray, “A Post Keynesian View of Central Bank Independence, Policy Targets, and the Rules Versus Discretion Debate,” Journal of Post Keynesian Economics, 30 (2007): 138. I have challenged the common belief that the Fed determines interest rates: Jeffrey Rogers Hummel, “The Myth of Federal Reserve Control Over Interest Rates,” Library of Economics and Liberty (October 7, 2913).
 It is interesting to note that at one time Milton Friedman, in “A Monetary and Fiscal Framework for Economic Stability,” American Economic Review, 38 (June 1948): 245-64, proposed something similar—eliminating all government borrowing and relying entirely on taxation or issuing money to finance government expenditures.
 L. Randall Wray, Modern Money Theory: A Primer on Macroeconomics for Sovereign Monetary Systems, 2nd ed. (Basingstoke, UK: Palgrave Macmillan, 2015), p. 112.
 Ibid., pp. 193, 195.
 See Stan Veuger, “Modern Monetary Theory and Policy,” AEI (January 8, 2019); Scott Sumner and Patrick Horan, “How Reliable Is Modern Monetary Theory as a Guide to Policy?” Mercatus Center Policy Briefs (March 11, 2019), PDF file.