I. Motivating Question
How Do Macroeconomists Account for Episodes of High Inflation?
Historically, hyperinflations have always followed the same pattern: governments have budget deficits that they are unable or unwilling to finance without printing money, and the resulting high rate of money growth leads to very high inflation. The central role of money growth in creating hyperinflation is consistent with the macroeconomic model developed in previous chapters.
II. Why the Answer Matters
Hyperinflation tends to erode the basic economic mechanisms at work in market economies and can have political as well as economic implications. Two important episodes of hyperinflation occurred after the First and Second World Wars. In addition, a number of economies have suffered bouts of hyperinflation over the past thirty years, and Russia has struggled with high inflation since the liberalization of its economy, so the issue remains relevant today.
III. Key Tools, Concepts, and Assumptions
1. Tools and Concepts
i. Hyperinflation is very high inflation (inflation over 50% per month).
ii. Seignorage is revenue available to the government from the printing of money. The inflation tax is the loss in value of real money balances because of inflation. The inflation tax equals the inflation rate times real money balances.
iii. Dollarization refers to the use of U.S. dollars in the domestic transactions of another country.
iii. Incomes policies refers to some combination of wage controls or guidelines and price controls or guidelines. Heterodox programs to end a hyperinflation include incomes policies; orthodox programs do not.
IV. Summary of the Material
Hyperinflations are episodes of very high inflation. Typically, such episodes last less than two years. Historically, hyperinflations have always been accompanied by monetary expansions of an order of magnitude comparable to the inflation rate. A government budget deficit financed by money creation has been at the root of all hyperinflations.
1. Budget Deficits and Money Creation
The start of a hyperinflation is typically characterized by two circumstances: a social or economic upheaval that increases the budget deficit and an increasing unwillingness or inability of the government to finance its deficit by borrowing from the private sector. When these two circumstances coincide, the government effectively borrows from the central bank, which prints money to buy the bonds issued by the government. This process is called debt monetization. Under this scenario, money creation is given by
∆M/P = Deficit, (23.1)
where Deficit is the government deficit measured in real terms. The LHS of equation (23.1) is called seignorage. It represents the revenue available to the government from printing money. It is convenient to express seignorage in terms of money growth and real money balances:
seignorage=(∆M/P) = (∆M/M)(M/P). (23.2)
2. Inflation and Real Money Balances
From the money-market equilibrium condition, equation (23.2) becomes
∆M/P = (∆M/M)YL(r + πe). (23.3)
During a hyperinflation, changes in actual and expected inflation are enormous relative to changes in output and the real interest rate. Increases in actual and expected inflation reduce demand for real money balances. People make more use of barter—the exchange of goods for other goods, rather than for money—and demand more frequent wage payments. Moreover, sometimes they began to use foreign currencies rather than the home currency as a monetary instrument. A common phenomenon is dollarization, the use of U.S. dollars in another country's domestic transactions.
3. Deficits, Seignorage, and Inflation
If money growth is constant, Chapter 9 showed that actual and expected inflation eventually equal the rate of money growth. In this case, equation (23.3) can be written as
∆M/P = (∆M/M)YL(r + ∆M/M). (23.4)
Money growth creates inflation, which reduces the real value of money balances. Thus, money growth is said to impose an inflation tax: the tax rate is the rate of inflation (equal to ∆M/M), and the tax base is the stock of real money balances (equal to YL(r + ∆M/M)). The inflation tax generates seignorage for the government, because people need to increase their nominal money stocks in order to sustain their real money balances. Money growth has two effects on seignorage. On the one hand, an increase in money growth increases the tax rate, but on the other hand it reduces the tax base, as higher nominal interest rates cause people to shift away from holding money. Empirically, the relationship between inflation and seignorage is hump shaped. At low inflation rates, seignorage increases with inflation; at higher rates, seignorage declines with inflation. Thus, there is some rate of money growth that maximizes seignorage.
In hyperinflations, money growth typically far exceeds the rate that maximizes seignorage. Why does this happen? In the long run, it is impossible for the government to raise more seignorage than the maximum possible under constant money growth. In the short run, however, the government can exploit the time it takes for real money demand to adjust to faster money growth and increase seignorage temporarily by increasing the rate of money growth. If the government implements this strategy, however, eventually real money balances will decrease. As a result, more money growth will be required to generate a given level of seignorage. Thus, an ongoing attempt to generate seignorage in excess of the maximum under steady inflation requires accelerating money growth and rising inflation. Moreover, real tax revenues tend to decline with inflation (the so-called Tanzi-Olivera effect), because nominal tax collections occur with a lag. As a result, the real deficit (and thus the need for seignorage) increases over time, contributing to the monetary acceleration.
With respect to output, an increase in money growth initially increases output by increasing expected inflation and reducing the real interest rate. As inflation becomes very high, however, the monetary exchange system becomes very inefficient, prices carry less information because they change so often, and borrowing and lending is curtailed because of uncertainty about inflation (and hence about ex post real interest rates).
4. How Do Hyperinflations End?
Hyperinflations do not end naturally. Ending them requires the implementation of a stabilization program, which includes, at a minimum, credible fiscal reform (reduced spending plus replacement of the inflation tax by ordinary taxes) and a commitment by the central bank not to monetize government debt. Orthodox stabilization programs rely solely on these two elements. Heterodox stabilization programs add incomes policies (guidelines for wage and price setting) designed to coordinate the expectations of wage and price setters. Disinflation implemented during hyperinflation may actually be less costly (i.e., have a lower sacrifice ratio) than disinflation implemented during a more moderate inflation, because high levels of inflation tend to eliminate inertia arising from the staggering of wage and price setting. When inflation is high, institutions change to allow wages and prices to adjust much more quickly. Still, inertia may persist if the stabilization program is not credible, and in this case, the associated output costs may make it difficult to sustain the stabilization effort.
The Core describes an adjustment mechanism whereby output returns to its natural level in the medium run. Moreover, the Core discusses how monetary and fiscal policy can be used to speed adjustment. Chapters 22 and 23 describe how these basic features of the economy can break down. An economy in a slump or depression can end up in a deflation, and the deflation can move the economy further away from its natural level. In such a situation, monetary policy may be constrained by a liquidity trap, and fiscal policy by increasing levels of government debt. Moreover, faced with a major adverse shock, a government may run large and increasing budget deficits, and have no financing option other than money creation. Such a situation can lead to high inflation or hyperinflation.
Instructors might wish to clarify two points. First, as noted in the text, there is a distinction between seignorage and the inflation tax revenue. Seignorage equals money growth times the real money supply. Inflation tax revenue equals the inflation rate times the real money supply. Under constant money growth, however, inflation will eventually equal money growth, so seignorage will equal inflation tax revenue.
Second, the chapter makes it clear that hyperinflation has its roots in fiscal policy. So, as a qualification to Friedman’s adage that “inflation is always and everywhere a monetary phenomenon,” fiscal policy can sometimes drive monetary policy. Why does the central bank lend to a government that requires deficit finance? It may have no choice if it is politically subordinate to the Treasury (or Finance Ministry). Or the central bank may monetize the deficits because it fears the consequences of a government default more than high inflation.
1. Government Solvency
Why would a government be unable to borrow to finance a budget deficit? Potential creditors do not have to lend to the government. They can earn a market rate of return by lending to the private sector or governments of other countries. They will only lend to their government if the present value of expected repayment, calculated using the market rate of return, is at least as great as the amount lent. (The expected repayment must, of course, be adjusted for default risk.)
The total repayments that the government makes to its creditors consist of interest payments plus amortization (repayment of principal) on the amounts borrowed. Where does the government get the resources to make these payments? The resources available to the government to pay its creditors consist of the primary surplus in the government’s budget (the government’s fiscal surplus exclusive of interest payments) plus the seignorage that the government can raise by printing money. Thus, for the government to be perceived as a good credit risk, the present value of its primary future surpluses plus seignorage must be expected to be at least as large as the debt that it has to repay. When this condition is met, the government is said to be solvent. Otherwise, it is insolvent.
Recall that seignorage is not unlimited. It reaches a maximum value even if governments are willing to tolerate very high rates of inflation. Thus, creditors will be willing to lend to the government only if they believe that the government will produce sufficiently large primary surpluses in the future to repay its debt.
2. Wage and Price Controls
Why do governments sometimes use wage and price controls in inflation stabilization programs? In the context of high inflation, the economy can be described by a long-run equilibrium in which the AD and AS curves are both shifting up over time at the rate π. The drift of the AD curve reflects monetary expansion; the drift of the AS curve reflects expectations-driven demand for higher nominal wages. As an extreme case of stabilization, consider a fiscal and monetary reform that eliminates monetary expansion. If such a program were implemented, the upward drift of the AD curve would stop immediately, but, assuming that expected inflation changes slowly, the AS curve would continue to shift vertically upward, reducing output. A freeze of wages and prices would eliminate the upward shift of the AS curve and the concomitant loss of output. The elimination of the output loss would reduce the temptation of the government to abandon the stabilization program. Moreover, the immediate stabilization of inflation through the combination of monetary-fiscal discipline and wage and price controls might make the program more credible.
Why would stabilization programs not include wage and price controls? One reason is that it is difficult to implement them without freezing relative prices, which introduces costly microeconomic distortions. A second reason is that governments often think of wage and price controls as substitutes for the fiscal and monetary measures that are the prerequisites for successful stabilization. Finally, a more subtle reason is that controls may themselves undermine the credibility of the stabilization program. If wage and price setters attribute the success of the program to the controls themselves, or interpret the imposition of controls as a signal that the government has a low tolerance for recession, they may question the government’s commitment to the stabilization effort.