The persistence of inflation and the cost of disinflation

New England Economic Review, Jan-Feb, 1995 by Jeffrey C. Fuhrer

Some economic questions can be considered only within the context of an economic model. For example, "How costly would it be (in terms of lost jobs and output) to lower the inflation rate to zero?" is a question that, because it is counterfactual, can only be answered by creating an economic model that allows us to estimate the effects of pursuing counterfactual economic policies. Models are not all created equal, however, so that the answer to this question can vary widely depending upon the characteristics of the model used to address the question.

This article will argue that one cannot answer the above question accurately without using a model that properly captures an important feature of the real world: the persistence of inflation. In essence, the more persistence inflation exhibits, the harder monetary policy has to push on it to bring it down. The harder monetary policy has to push, the more it will disrupt the real economy, and the greater will be the cost associated with disinflating.

While there is wide agreement that inflation is persistent and that disinflations have been costly, the source of persistence and the reason for the cost are not widely agreed upon. As will be discussed below, persistence of inflation and the cost of disinflating may arise for several reasons, including the inertia that wage and price contracts impart to the inflation rate, the inertia that slowly adjusting expectations may impart to inflation, or the inertia that imperfect credibility may impart to inflation.(1) This study will demonstrate the importance of persistence in a model of inflation, and then consider each of the explanations given above.

Different sources of inflation persistence bear different implications for the conduct of monetary policy. If disinflations are costly because the Federal Reserve lacks credibility, then the Fed should determine whether and how it can improve its credibility. If persistence arises from other aspects of price-setting behavior, then monetary policy must accept the costs of disinflation unless these behaviors change.

I. Defining Persistence

What do economists mean when they talk about the "persistence" of an economic variable? Persistence refers to the tendency for a variable to stay away from its average level for an extended period when perturbed.(2) For example, when the unemployment rate deviates significantly from its "natural" rate, most economists would not expect it to return immediately. Similarly, this study will show that historically, when inflation has deviated from the rate that the monetary authority desires, its return to the desired rate takes quarters or years, not weeks or months. Failure of the model to incorporate this "inflation persistence" can produce misleading policy prescriptions.

The persistence of inflation in the prices of goods and services contrasts sharply with the lack of persistence in the inflation in prices of financial assets. A leading example is the price of stocks traded on the New York Stock Exchange. The rate of change in a basket of stock prices, which averages out idiosyncratic movements of individual stocks, shows little or no persistence. A graphical comparison of these two qualitatively different behaviors is displayed in Figure 1, which shows the monthly percentage changes in the consumer price index (CPI) and the Standard & Poor's composite stock index. The difference in the volatility of the two series is obvious and striking. Stock prices are about as likely to rise as to fall markedly from month to month, regardless of which direction they were headed last month. The CPI changes only a bit from month to month, and the tendency for positive (negative) changes to be followed by positive (negative) changes is pronounced.

Common-sense economic reasons can be given for the difference in the behavior of these two types of prices. The prices of financial assets may be thought of as the valuation that the financial markets place on the expected stream of returns to holding the asset. For equities, the price will reflect the (discounted) expected earnings that a firm will accrue over its lifetime, or the dividends that the firm is expected to pay to shareholders over its lifetime. Thus, the price depends on the market's expectations, which are free to change from minute to minute.

In contrast, the prices of goods and services - the prices of chicken and haircuts, for example - cannot move as freely as the prices of financial assets. While their prices may reflect in part the expectations of market participants, they also depend on the cost of inputs to production and the terms of contracts with suppliers and buyers. The largest of the input costs for most goods and services is the cost of labor, which varies slowly as salaries and benefits are adjusted, usually annually. Thus, it is unlikely that the average level of goods and services prices, as reflected in the CPI, for example, will exhibit the same flexibility as the average level of financial asset prices.