The Fiscal Theory of the Price Level has been percolating among monetary theorists for over three decades: Eric Leeper being the first to offer a formalization of the idea, with Chris Sims and Michael Woodford soon contributed to its further development. But the underlying idea that the taxation power of the state is essential for the acceptability of fiat money was advanced by Adam Smith in the Wealth of Nations to explain how fiat money could be worth more than its minimal cost of production. The Smith connection suggests a somewhat surprising and non-trivial intellectual kinship between the Fiscal Theory and Modern Monetary Theory that proponents of neither theory are pleased to acknowledge.
While the Fiscal Theory has important insights, it seems to promise more than it delivers. Presuming to offer a more robust explanation of price-level or inflation fluctuations than the simple quantity theory (not that high a bar), it shares with its counterpart an incomplete account of the demand for money, paying insufficient attention to the reasons for, and the responses to, fluctuations in that demand.
In this post and perhaps one or two more to follow, I use a 2022 article by John Cochrane showing how the Fiscal Theory accounts for both recent and earlier inflationary and disinflationary episodes more persuasively than do other theories of the price level, whether Monetarist or Keynesian regardless of specific orientation. Those interested in a fuller exposition of the Fiscal Theory will want to read Cochrane’s recent volume on the subject.
Let’s start with Cochrane’s brief description of the Fiscal Theory (p. 126):
The fiscal theory states that inflation adjusts so that the real value of government debt equals the present value of primary surpluses.
Most simply, money is valuable because we need money to pay taxes. If, on average, people have more money than they need to pay taxes, they try to buy things, driving up prices. In the words of Adam Smith (1776 [1930], Book II, chap. II): “A prince, who should enact that a certain proportion of his taxes be paid in a paper money of a certain kind, might thereby give a certain value to this paper money . . .” Taxes are a percentage of income. Thus, as prices and wages rise, your dollar income rises, and the amount of money you must pay in taxes rises. A higher price level soaks up excess money with tax payments. Equivalently, the real value of money, the amount of goods and services a dollar buys, declines as the price level rises. But the real value of taxes does not change (much), so a higher price level lowers the real value of money until it equals the real value of tax payments.
It’s useful to quote Adam Smith about how to account for the value of intrinsically worthless pieces of paper, but Smith was explaining the source of the value of fiat money, not necessarily the actual value of any given fiat money at any particular time or the causes of fluctuations in the value of fiat money over time. Precious metals were originally used as media of exchange only because they had a value independent of their being used as media of exchange. But once they are so used, their value in exchange rises above the value those metals would have had if they had not been used as media of exchange.
For a century or more before the mid-1870s, when both gold and silver were widely used as media of exchange, an ounce of gold had been worth between 15 and 16 times more than an ounce of silver. Many countries, including the US before the Civil War, operated on a bimetallic standard in which the legal or mint value of gold in terms of the local currency was set between 15 to 16 times the mint value of silver in terms of the local currency. As long as the relative market values of gold and silver remained close to the legal ratio, bimetallic systems could operate with both gold and silver coins circulating. But when the market value of one of the metals appreciated relative to the other, legally overvalued coins would disappear from circulation being replaced by the legally undervalued coins. Gresham’s Law (“bad” money drives out “good” money) in action. But, inasmuch as increased monetary demand for the overvalued metal tended to raise the market value of that metal relative to that of the other, bimetallic systems had a modest stabilizing property.
After the North prevailed in 1865 over the South in the Civil War and the unification of Germany in 1871, both the US and Germany opted for a legal gold standard rather than a bimetallic standard. And by 1874, the increased demand for gold had raised the value of gold sufficiently to breach the historical 16 to 1 upper bound on the value of gold relative to silver. The countries remaining on a legal or de facto (bimetallic) silver standard experienced inflation. To avoid importing inflation by way of Gresham’s Law, countries on the silver standard began refusing silver for coinage, thereby accelerating the depreciation of silver relative to gold, and promoting the international transition to the gold standard, which, by 1880, was more or less complete.
So, once there is a monetary demand to hold fiat money, the simple fiscal theory of the value of money cannot provide a full account of the value of money any more than a theory of the value of gold based on the non-monetary demand for gold could account for the price level under the gold standard.
The limitations implicit in the Fiscal Theory are implicit in Cochrane’s summary of the Fiscal Theory: inflation adjusts so that the real value of government debt equals the present value of primary surpluses. In other words, the Fiscal Theory treats both bonds and money issued by the government or the monetary authority (i.e., the monetary base or outside money) as government debt. But that’s true only if the monetary base and government bonds are perfect, or at least very close, substitutes. Cochrane argues that the monetary base is, if not perfectly, at least easily, substitutable for bonds, so that the real value of government debt is, at least to a first approximation, independent of the ratio of government bonds held by the public to the monetary base held by the public.
However, if the demand for the monetary base, apart from its use in discharging tax liabilities, is distinct from the demand for government bonds, the monetary base constitutes net wealth not merely a liability. The basic proposition of the Fiscal Theory must then be revised as follows: inflation adjusts so that the real value of government debt does not exceed the present value of primary surpluses. The corollary of the amended proposition is that if the monetary base constitutes net wealth, inflation need not be affected by the real value of government debt.
If the fiscal constraint isn’t binding, so that the primary budget surplus exceeds government debt (exclusive of the monetary base), the monetary authority can control inflation by conducting open market operations (exchanging outside money for government debt or vice versa). By creating outside money to purchase government debt, the monetary authority decreases the real debt liability of the government correspondingly. However, the extent to which outside money constitutes net wealth depends on the real demand of the public to hold outside money rather than government debt or inside money. If the real demand to hold outside money declines, the wealth represented by the stock of outside money is diminished correspondingly. Unless outside money is retired by way of a government surplus or by the sale of government debt by the monetary authority, the price level will rise.
Explaining why outside money and government debt sold to the public are equivalent, Cochrane argues:
In the monetarist story, assets such as checking accounts, created by banks, satisfy money demand, and so are just as inflationary as government-provided cash. Thus, the government must control checking accounts and other “inside” liquid assets. In the basic fiscal theory, only government money, cash and bank reserves, matter for inflation. Your checking account is an asset to you but a liability to the bank, so more checking accounts do not make the private sector as a whole feel wealthier and desire to spend more. The government need not control the quantity of checking accounts and other liquid assets. However, in the basic fiscal theory, government debt, which promises money, is just as inflationary as money itself. Reserves and cash are just overnight government debt.
Cochrane is correct, as James Tobin explained over 60 years ago, that inside money supplied by banks is not inherently inflationary. But what is true of bank liabilities, which are redeemable on demand for government issued outside money, is not necessarily true of government outside money. In a footnote at end of the quoted passage, Cochrane acknowledges that difference.
Reserves are accounts banks hold at the Federal Reserve. Banks may freely convert reserves to cash and back. The Fed issues cash and reserves, and invests in Treasury debt, just like a giant money-market fund. Because the interest the Fed pays on reserves comes from the interest it gets from Treasury securities, and since it remits any profits to the Treasury, we really can unite Fed and Treasury balance sheets and consider cash and reserves as very short-term and liquid forms of government debt, at least to first order.
Since banks began receiving interest on reserves held at the Fed, the distinction between Treasury liabilities held by the public and Treasury liabilities held by the Fed was—to first order—nullified, as was the operational distinction between the Treasury and the Fed. But the conceptual distinction between money and debt is not inherently a nullity, and, insofar as the operational distinction has been nullified, it’s because, in 2008, the Fed began paying competitive interest on bank reserves held at the Fed. So, insofar as the Fiscal Theory relies on the equivalence of government debt and government fiat money, it relies either on a zero nominal interest rate or a policy of paying competitive interest on reserves held at the Fed. I shall return to this point below.
Keynes, in Chapter 17 of the General Theory, despite erroneously explaining interest as merely a reward for foregoing the exercise not of time–but of liquidity–preference, argued correctly that the expected return on alternative assets held over time would be equalized in equilibrium. Expected returns from holding assets, net of holding costs, can accrue as pecuniary payments e.g., interest, as flows of valuable in-kind services, or as appreciation. Keynes’s insight was to identify the liquidity provided by money as an in-kind service flow for which holders forego the interest payments or expected appreciation that they could have gained from holding non-monetary assets.
The predictions of the Fiscal Theory therefore seem contingent on blurring the distinction between inside and outside money. Outside money is created either by the government or the central bank. Instruments convertible into outside money, such as commercial bank deposits and Treasury debt, are alternatives to outside money, and may therefore affect the demand to hold outside money. So, even if Treasury debt is classified money, it is properly classified as inside, not outside, money. As long as the demand of the public to hold high-powered money is distinct from its demand to hold other assets, the monetary authority has sufficient leverage over the price level to conduct monetary policy.
If there’s no distinct demand for outside money (AKA the monetary base), then differences, during a given time period, between the quantity of outside money demanded by the public and the stock created by the monetary authority have no macroeconomic (price-level) consequences. But if there is a distinct demand, the stock of outside money, contrary to the presumption of the Fiscal Theory, isn’t a net liability of the monetary authority or the government; it’s an asset constituting part of the net wealth of the community.
Nevertheless, Cochrane is right that financial innovation over time has steadily increased the importance of inside money compared to outside money, a process that nineteenth century monetary economists (notably the Currency and Banking Schools) were already trying understand as bank deposits began displacing banknotes as the primary monetary instrument used to mediate exchange and to store liquidity. Continuing financial innovation and the rapid evolution of electronic payments technology, especially in this century have again transformed how commercial and financial transactions are executed and how households make purchases and store liquidity.
The Fiscal Theory described by Cochrane therefore provides insight into our evolving and increasingly electronic monetary system. While Cochrane emphasizes the payment of interest on reserves held by banks at rates equal to, or greater than, the yields on short-term Treasury debt, an alternative arrangement in which the Fed paid little or no interest on bank reserves could also operate efficiently by means of an overnight interbank lending market. The amount of reserves held by banks would fall drastically as the banking system adjusted to operating with minimal reserves sufficient to meet the liquidity needs of the banking system, with the Fed discount window available as a backstop.
Thus, in our modern monetary system, the Fed can either operate with a large balance sheet of Treasury and other highly liquid debt while paying competitive interest on the abundant reserves held by banks, or with a small balance sheet while Treasuries and other highly liquid debt are held by banks holding only minimal reserves. The size of the Fed balance sheet per se is relatively insignificant as a matter of economic control. What matters is that by paying competitive interest on bank reserves held at the Fed, the Fed has rendered itself, as Cochrane correctly argues, incapable of conducting an effective monetary policy. Awash in reserves, banks have become unresponsive to changes in the Fed’s policy rate.
By significantly reducing or eliminating interest on bank reserves, the Fed would not only shrink its balance sheet, it would increase, if only to a limited extent, the effectiveness of monetary policy by making banks more responsive to changes in its policy rate. However, given that most banks can operate effectively with reserves that are a small fraction of their deposit liabilities, Cochrane may be right that the Fed’s monetary policy in the modern system would still be limited, because changes in the Fed’s interest-rate target would induce only small adjustments in banks’ lending practices and policies.
While it’s true that the huge stock of currency now in the hands of the public (likely held mostly abroad not in the US) would continue to provide a buffer against inflationary or deflationary fiscal shocks, the demand for currency is likely not very responsive to changes in interest rates, so that Fed policy changes would have little or no macroeconomic effect on the demand for US currency. Indeed, any effect would likely be in the wrong direction, an increase in interest rates, for example, tending to reduce the amount of currency demanded thereby reducing the dollar exchange rate, and raising, not reducing, inflation.
Almost 40 years ago, in my book Free Banking and Monetary Reform, written in the wake of 1970s inflation and the brutal Volcker disinflation, I argued for a radical monetary reform. After discussing the early manifestations of the financial innovation then just starting to transform the monetary system, I proposed a free-banking regime in which competitive banks would pay interest on demand deposits (which was then prohibited). An important impetus for financial innovation was then to avoid the implicit taxation of bank deposits imposed by legal reserve requirements. The erosion of the tax base by financial innovation caused reductions in, and eventual elimination of, those reserve requirements. As I pointed out (p. 169):
As long as there is a demand for high-powered money, the Fed can conduct monetary policy by controlling [either directly or, by using an interest-rate target as its policy instrument, indirectly] the quantity of high-powered money. Since there is a demand for high-powered money apart from the demand to hold required reserves, reserve requirements are not logically necessary for conducting monetary policy. Nor is control over the overall quantity of money necessary for the Fed to operate a monetary policy. All it needs, as noted, is to control the quantity of high-powered money. And it would have that control even if required reserves were zero.
But as we just saw, the stability of the demand for high-powered money is also important. If the demand for required reserves is more stable than the demand for other components of high-powered money, reducing demand for required reserves makes the overall demand for high-powered money less stable. And as I pointed out earlier, the less stable the demand for high-powered money is, the greater the risk of error in the conduct of monetary policy will be.
So, although the Fed could, even with a greatly reduced stock of bank reserves as a basis for conducting monetary policy, still control inflation, the risk of destabilizing policy errors might well increase. One response to such risks would be to reimpose at least a modest reserve requirement, thereby increasing the stock of bank reserves on which to conduct monetary policy. The effectiveness of reimposing legal reserve requirements in the current environment is itself questionable. But in my book, I proposed, adopting Earl Thompson’s idea (inspired by Irving Fisher’s compensated dollar plan) for a labor standard stabilizing a wage index using the price of gold as a vehicle for a system of indirect convertibility. (See chapter 11 of my book for details). An alternative for achieving more or less the same result might to adapt Thompson’s proposal to stabilizing nominal GDP, as Scott Sumner and others have been advocating since the 2008 financial crisis.
So, despite my theoretical reservations about the Fiscal Theory of the Price Level, it seems to me that, in practice, we have a lot in common.