Cato Institute
1000 Massachusetts Ave, NW
Washington, DC 20001-5403

Phone (202) 842 0200
Fax (202) 842 3490
Contact Us
Support Cato

Cato Policy Report, May/June 1999
Vol. 21, No. 3

A Plea for (Mild) Deflation

by George Selgin

Weaver, Sperling, Tanner Carolyn Weaver (left) of the American Enterprise Institute and Michael Tanner (right) of the Cato Institute teamed up on Gene Sperling, director of President Clinton’s National Economic Council, at Cato’s February 12 Policy Forum on Clinton’s Social Security reform plan.
In recent years the central bankers of the world’s most advanced economies have come close to achieving a goal that seemed unreachable two decades ago: the eradication of inflation. In the United States, for instance, the consumer price index (CPI) rose only 1.6 percent in 1998. Yet, just as the elusive goal of price stability has come within their grasp, the central bankers are getting cold feet: a little inflation, they now tell us, is a price worth paying to avoid an even worse menace: deflation, or falling prices.

Deflation. For many people the word conjures up images of the Great Depression, when prices fell dramatically throughout most of the world. The decline in prices was the counterpart of collapsed sales, widespread bankruptcies, and armies of unemployed workers. If a little inflation is the only guarantee against another calamity like that of the 1930s, then it is, indeed, a price worth paying.

The truth, however, is that deflation need not be a recipe for depression. On the contrary, a little deflation can be a good thing, provided that it is the right kind of deflation.

Since the disastrous 1930s, economists and central bankers seem to have lost sight of the fact that there are two kinds of deflation—one malign, the other benign. Malign deflation, the kind that accompanied the Great Depression, is a consequence of shrunken spending, corporate earnings, and payrolls. Strictly speaking, even in this case, it is not so much deflation itself that is harmful as its underlying cause, an inadequate money stock. The hoarding of money, or its actual disappearance (the quantity of money in the U.S. economy actually shrank 35 percent between 1930 and 1933), causes the demand for goods and services to dry up. In response, firms are forced to curtail production and to lay off workers. Prices fall, not because goods and services are plentiful, but because money is scarce.

Benign deflation is something else altogether. It is a result of improvements in productivity, that is, occasions when changes in technology or in management techniques allow greater real quantities of finished goods and services to be produced from a given quantity of land, labor, and capital. Because an increase in productivity is the same thing as a decline in unit costs of production, a productivity-driven decline in the prices of finished goods and services needn’t involve any decline in producers’ earnings, profits, or payrolls. Lower costs are matched by correspondingly lower consumer prices, not by lower wages or incomes. Such productivity-driven deflation is actually good news to the average breadwinner.

Benign deflation is a relatively unfamiliar concept both because modern economists devote relatively little attention to it (focusing all their discussions on inflation or on deflation of the malignant sort) and because monetary policymakers in the United States and elsewhere have prevented benign deflation from occurring throughout most of the 20th century. Thus, since World War II, the United States has witnessed frequent gains in productivity. The real (inflation-adjusted) unit cost of production of goods and services today is approximately half of what it was in 1945. Yet the general price level, instead of being half as high as it was after the war, has increased to over nine times its former level. Instead of allowing goods’ prices to fall along with their falling real costs of production, the Fed has artificially inflated those prices by pumping large amounts of money into the economy. The last time productivity improvements were allowed to be reflected, partially and temporarily, in a fallen price level was in 1955, 44 years ago. And that was not by design but by accident.

Yet ongoing, benign deflation is not just a hypothetical possibility. Many Western nations experienced something like it between 1873 and 1896, when the gold standard placed limits on Western governments’ ability to offset the effects of productivity improvements through monetary expansion. Although national price levels declined almost continuously from 1873 to 1896, causing scholars for a time to refer to the era in question as the world’s first “Great Depression,” every other economic indicator—prices, wages, profits, industrial output, trade—shows the period to have been one of unprecedented growth and prosperity. The period had its share of genuine depressions, to be sure, but those cyclical downturns were a result of faulty financial legislation. No harm seems to have come from allowing a downward trend in prices so long as that trend reflected ongoing gains in productivity.

The last decade or so has witnessed substantial gains in labor productivity, along with smaller gains in total factor productivity, making the present era one in which significant reductions in the nominal cost of living might have been painlessly achieved. The world’s central bankers have, however, been unwilling to let those reductions occur.

Central Bankers’ Misconceptions

Central bankers believe that deflation is always harmful, no matter its cause. They are convinced (1) that deflation is unfair to debtors, because it arbitrarily increases the real value of their debts; (2) that deflation means falling wage rates and increased un-employment; (3) that a stable, or slowly ris-ing, price level is the best means of avoiding booms and busts; (4) that deflation may mean artificially high real interest rates, because nominal interest rates can never go below zero; (5) that deflation is painful to sellers, who will therefore resist cutting prices; and (6) that a variable rate of deflation must be a source of avoidable entrepreneurial confusion and error. In truth, none of those beliefs is valid so long as the deflation in question mirrors the economy’s rate of productivity growth.

Consider first deflation’s effects on debtors. Suppose that over the course of one year overall productivity unexpectedly rises 2 percent and the general price level declines 2 percent. Then, although creditors will earn a higher real interest rate than they anticipated, debtors will have no reason to complain: although the real value of debtors’ obligations rises, so does their real income, while the nominal payments burden borne by them is unchanged. Debtors can, in other words, afford to pay higher real rates of interest, and might well have agreed to such rates in advance had they and their creditors both been equipped with perfect foresight. A monetary policy aimed at deliberately preventing prices from falling would, in this case, merely deprive creditors (and other people with fixed incomes, including persons living on Social Security) of their fair share of productivity gains being enjoyed by all other income earners.

Those who insist nevertheless upon a policy of price-level stabilization as the only equitable policy would do well to consider what such a policy entails in the event of a setback to productivity. When productivity declines, the only way to keep the price level stable is by shrinking nominal incomes. Debtors might then find it not only difficult but perhaps impossible to repay their loans. If it is admitted that debtors’ and creditors’ interests are best served by an increase in prices when productivity declines, then symmetrical reasoning suggests that those same interests are best served by allowing prices to fall as productivity advances.

The falsity of the belief that deflation must involve falling wages has already been alluded to. When productivity rises, so do workers’ real wages. In other words, nominal (money) wage rates rise relative to the cost of living. If the productivity gains are allowed to take the form of falling output prices, then money wage rates will remain stable or will rise modestly. (The second result obtains if prices fall at a rate equal to the growth rate of total factor productivity, which usually progresses more slowly than the growth rate of labor productivity.) Otherwise, if the monetary authorities insist on preventing the price level from falling, productivity gains must take the form of more rapidly increasing money wages.

Suppose, for example, that labor productivity grows at an annual rate of 3 percent, while total factor productivity grows at an annual rate of 2 percent. Then the real wage rate, which reflects labor productivity, should also increase at an annual rate of 3 percent. If consumer prices are allowed to decline at a rate equal to the rate of growth of total factor productivity, money wage rates will still increase at a modest 1 percent annual rate. If, in contrast, the authorities insist on stabilizing the CPI, money wage rates must increase 3 percent a year. Generally speaking, money wage rates are less “flexible” than output prices, so a policy of avoiding or limiting wage-rate adjustments by allowing prices to fall is less likely to be a source of labor-market “frictions” and consequent labor misallocation. (Of course, if productivity declines, labor-market disturbances are best avoided by letting output prices rise while leaving money wage rates alone, rather than by stabilizing output prices and thereby making a reduction in money wage rates necessary.)

The relative rigidity of input prices, and of the price of labor especially, compared to output prices is also reason for doubting the widespread belief that a stable price level best avoids booms and busts. Suppose, for example, that productivity grows more rapidly than usual. In that case, a zero-inflation policy requires a money growth rate sufficient to sustain a rate of factor-price inflation equal to the rate of productivity growth. If, however, factor prices are rigid, more rapid monetary expansion may at first succeed in swelling corporate earnings, without inducing any equivalent rise in factor prices. Firms’ profits will then be artificially enhanced. Speculators who fail to appreciate the temporary nature of those swollen profits will bid up stock prices, generating a boom. Eventually, though, factor prices will respond positively to expanded earnings (costs will catch up with revenues), so profits will shrink, and share values will decline. The boom will give way to a bust. Were the monetary authorities to stabilize, not the level of output prices, but the level of input prices (or, equivalently, the flow of nominal income or revenues), the boom-bust cycle might be avoided.

The claim that deflation is painful to sellers, so they resist cutting prices, may be valid for malign, demand-driven deflation, but it is not valid for deflation of the benign, productivity-driven sort. When overall spending on goods and services shrinks, sellers may resist cutting prices until they can negotiate corresponding cuts in costs, thereby preserving their profit margins. For that reason, product prices are often set according to “implicit contracts” promising some fixed percentage markup of prices above unit costs. But this practice, which accounts for the sluggish adjustment of product prices in response to changes in consumer spending, does not lead to sluggish price adjustment in response to changes in productivity. When productivity changes, so do unit costs of production, so adjustments in product prices are in fact required to preserve constant markups. Indeed, sellers actively seek ways to improve productivity just so that they can charge less than their rivals without sacrificing profits. Productivity-based price cuts are, in other words, a healthy aspect of the competitive process, which would occur routinely in most markets were it not for monetary authorities’ success in offsetting productivity gains with equal or greater additions to the flow of nominal expenditures.

The fear that deflation may mean artificially high interest rates, because nominal interest rates cannot be negative, is a red herring so long as the rate of deflation is never allowed to exceed the rate of productivity growth. Suppose, for example, that the nominal interest rate in an economy with constant total factor productivity and a constant price level is 4 percent. If monetary policy were to tighten to the point of causing a 5 percent annual rate of deflation, with no offsetting improvement in productivity, disorder would result: nominal interest rates would approach, but could not fall below, zero, despite the deflation, creating a surplus of loanable funds and a correspondingly deficient level of current spending. But suppose that deflation is only allowed to proceed so long as it is matched by corresponding productivity gains. In that case, a 5 percent annual rate of deflation would only be permitted in response to a 5 percent growth rate in total factor productivity. Such a high rate of productivity growth tends to be reflected in a correspondingly high equilibrium real rate of interest. The effects of productivity growth and of deflation on the nominal interest rate would then tend to offset one another. In other words, equilibrium nominal interest rates would almost certainly remain positive.

Does Deflation Confuse Entrepreneurs?

The last mistaken belief—that a stable price level best serves to avoid entrepreneurial confusion and error—arises from a failure to appreciate the fact that incorrect price-level forecasts are only one of several kinds of forecast errors that can frustrate entrepreneurship. Of course a stable and, hence, fully predictable price level is desirable in a world of unchanging productivity, because price-level fluctuations could never be a source of useful entrepreneurial information in such a world. But ours is a world in which productivity constantly changes, sometimes for better, sometimes for worse. In our world, unexpected productivity-driven price-level movements would serve to inform entrepreneurs of underlying changes in economic efficiency in the most transparent possible manner. If the price level is allowed to vary only with opposite changes in productivity, entrepreneurs’ inability to forecast every change in the price level is merely a reflection of the general impossibility of accurately forecasting productivity changes. Monetary authorities who imagine that they can avoid this source of entrepreneurial error by stabilizing output prices are deluding themselves: by stabilizing output prices, they merely succeed in destabilizing factor prices and nominal income, adding to instead of minimizing entrepreneurial confusion.

Probably all central bankers appreciate the importance of allowing particular money prices to reflect changing cost conditions in particular industries. Everyone understands that computers are being produced more efficiently today than they were a decade ago. That this information is most readily conveyed to consumers and entrepreneurs through a lower price tag on computers seems obvious. Yet central bankers refuse to extend this logic to cover general changes in productivity. In reality, not only computers, but also most other goods and services, are being produced at lower real cost (that is, with less land, labor, and capital per unit) today than was the case a decade ago. Between 1991 and 1997, overall unit costs of production in the private sector declined at an average annual rate of just over eight-tenths of 1 percent. So why shouldn’t the general price level be allowed to decline so as to reflect this fact? The answer is, for no reason at all, save central bankers’ unwarranted belief that deflation—even a little deflation—can never be in consumers’ best interest.

Productivity growth continues to put downward pressure on prices, and would result in actual deflation if only the Fed and other central banks would allow it. In-stead of aiming at zero inflation or (even worse) allowing prices to continue their gradual, upward drift, the world’s central bankers should allow price-level movements reflecting opposite changes in productivity. They can do this by stabilizing, not consumers’ prices, but the growth rate of nominal spending, allowing such spending to grow at a rate equal to the (expected) long-run growth rate of labor and capital, or (to judge from statistics for the last decade) at an annual rate of about 2 percent. This policy would lead to a very gradual downward trend in national price levels, interrupted by occasional productivity setbacks or “supply shocks,” like the OPEC-sponsored oil shortages of the 1970s. Because the policy requires that central banks always create enough money to offset hoarding (as reflected in reductions in money’s “velocity of circulation”), it poses no danger of a recurrence of a 1930s-style depression. Only benign deflation would be allowed: every decline in prices would be good news to both consumers and businesspeople.

The Ghost of the Great Depression

If the policy of mild, benign deflation being recommended here seems extreme, it is only because people have long been accustomed to rising prices, because of the lingering ghost of the Phillips curve (which purports to show a negative relationship between the rate of inflation and the rate of unemployment), and because the Great Depression gave deflation a bad reputation from which it has yet to recover. In this connection, it is worth noting how, prior to that episode, many of the world’s most famous economists, including Alfred Marshall, Dennis Robertson, Gunnar Myrdal, and Friedrich Hayek, viewed gradual deflation geared to productivity improvements as the least disruptive way to pass the benefits of economic progress along to consumers, including persons living on fixed money incomes. Their reasoning agreed with what used to be common sense: as goods become less costly to produce, their prices ought to fall.

The Great Depression dealt a near-fatal blow to such common-sense thinking about prices and the price level. A new generation of economists became so obsessed with avoiding the bad kind of deflation that they all but forgot about the good kind. Followers of Keynes advocated inflationary policies, which have been the norm ever since. Having paid penance for the Great Depression by suffering through six decades of inflation, it is time for us to revive old-fashioned logic concerning the potential benefits of deflation.

Recognition of the possibility of benign deflation should have a salutary effect on the thinking of the world’s central bankers. By helping them to overcome their fear of falling prices, it will encourage them to deal a deathblow to the worldwide scourge of inflation. But that is only the beginning. Once the possibility of benign deflation is fully appreciated, zero inflation itself will come to be recognized as an overly expansionary policy—that is, as a mere steppingstone on the way to something even better.

The Great Depression dealt a near-fatal blow to such common-sense thinking about prices and the price level. A new generation of economists became so obsessed with avoiding the bad kind of deflation that they all but forgot about the good kind. Followers of Keynes advocated inflationary policies, which have been the norm ever since. Having paid penance for the Great Depression by suffering through six decades of inflation, it is time for us to revive old-fashioned logic concerning the potential benefits of deflation.

Recognition of the possibility of benign deflation should have a salutary effect on the thinking of the world’s central bankers. By helping them to overcome their fear of falling prices, it will encourage them to deal a deathblow to the worldwide scourge of inflation. But that is only the beginning. Once the possibility of benign deflation is fully appreciated, zero inflation itself will come to be recognized as an overly expansionary policy—that is, as a mere steppingstone on the way to something even better.

This article originally appeared in the May/June 1999 edition of Cato Policy Report.