سیاست پولی هنگامی که تولید بالقوه قابل اطمینان نمی باشد: شناخت گمبل رشد از سال 1990s

The Fed kept interest rates low and essentially unchanged during the late 1990s despite a booming economy and record-low unemployment. These interest rates were accommodative by historical standards. Nonetheless, inflation remained low. How did the Fed succeed in sustaining rapid economic growth without fueling inflation and inflationary expectations? In retrospect, it is evident that the productive capacity of the economy increased. Yet as events unfolded, there was uncertainty about the expansion of the capacity of the economy and therefore about the sustainability of the Fed's policy.This paper provides an explanation for the success of the Fed in accommodating growth with stable inflation in the late 1990s. It shows that if the central bank is committed to reverse policy errors it makes because of unwarranted optimism, inflation can remain in check even if the central bank keeps interest rates low because of this optimism. In particular, a price level target—which is a simple way to model a commitment to offset errors—can serve to anchor inflation even if the public does not share the central bank's optimism about shifts in potential output. The paper shows that price level targeting is superior to inflation targeting in a wide range of situations. The paper also provides econometric evidence that, in contrast to earlier periods, the Fed has recently put substantial weight on the price level in setting interest rates. Moreover, it shows that CPI announcement surprises lead to reversion in the price level. Finally, it provides textual evidence that Alan Greenspan puts relatively more weight on the price level than inflation.

The performance of the US economy during the second half of the 1990s was outstanding. Rapid GDP and productivity growth were coupled with a very low unemployment rate and low-and-falling inflation. The sources of this performance are the focus of much of recent research (e.g., Krueger and Solow, 2001). Of course, many factors—including a possible decline in the equilibrium rate of unemployment (Staiger et al., 2001), changes in the structure of the labor force (Blank and Shapiro, 2001), acceleration in technological progress (Jorgenson and Stiroh, 2000; Oliner and Sichel, 2000; Basu et al., 2001), and pure luck (Mankiw, 2002)—likely contributed to the success of the late 1990s. Nonetheless, few question that the Fed's policies were very important in supporting economic growth with low and stable inflation. Since output was above its historical trend and unemployment was near a historical low, one might have expected the Fed to raise interest rates given its history of past policy actions. Even though inflation was not increasing, a preemptive tightening of monetary policy such as the Fed undertook in 1994 seemed likely. Yet the Fed held interest rates essentially unchanged in the face of output's spurt without an increase in inflation. Indeed, the rate of inflation has trended down since the mid-1990s. Alan Greenspan agrees that monetary policy was expansionary relative to historical standards. He says, “… from 1995 forward, we at the Fed were able to be much more accommodative to the rise in economic growth than our past experiences would have deemed prudent” (Greenspan, 2004, p. 35). The aim of this paper is to explain why accommodative monetary policy in this period did not increase inflation or inflationary expectations. One possibility is that the expansion of the economy's capacity in the late 1990s was apparent at the time, so that monetary policy merely accommodated an increase in productive capacity. In retrospect, this explanation has some appeal. As the 1990s unfolded, it was, however, much less clear that the economy experienced an increase in capacity. As Greenspan (2004, p. 34) notes, “The rise in structural productivity growth was not obvious in the official data… until later in the decade but precursors had emerged earlier.” Though it is now clearly established that a burst of productivity occurred in the late 1990s, there was a sequence of positive surprises as events unfolded. In the context of these very substantial, favorable surprises, the Fed pursued what at the time appeared to be a very expansionary policy. We argue that the Fed was effectively committed to a price level path. We do not claim that the Fed was following a price level rule per se, but rather that correcting policy mistakes was an important ingredient in its policy actions. We use a commitment to price level stability as a modeling device to represent a general commitment to undo the consequences of policy mistakes rather than allowing the consequences of policy mistakes to have long-term impacts. We show that a policy rule that has an element of the error-correction performs better than a rule without such correction, especially when potential output is not observable. In an economy with forward-looking agents, having the price level as an objective serves to anchor inflationary expectations. Making this commitment can help keep inflation in check even if the central bank is pursuing what looks contemporaneously like a very expansionary policy. We do not ascribe strict adherence to rules to Alan Greenspan or other central bankers. However, Alan Greenspan and others acknowledge the usefulness of rules as the benchmark for decision making. Greenspan says: Indeed, rules that relate the setting of the federal funds rate to the deviations of output and inflation from their respective targets … do seem to capture the broad contours of what we did over the past decade and a half. And the prescriptions of formal rules can, in fact, serve as helpful adjuncts to policy, as many of the proponents of these rules have suggested. … (Greenspan, 2004, pp. 38–39). Although Greenspan indicates that rules are inadequate for dealing with singular events such as stock market crash or the aftermath of the September 11, 2001 terrorist attacks, he says, “A rule does provide a benchmark against which to assess emerging developments” (Greenspan, 2004, p. 39). Our perspective in ascribing a price level target to the Alan Greenspan Fed is very much “as if.” This perspective is also found in Taylor's (1993) classic description of interest rate rules. Moreover, the particular policy rules we consider are not the key for our analysis. To reiterate, what is critical is that policy makers have a commitment to reverse the consequences of policy errors. We consider a continuum of interest policy rules with price level targeting (complete error-correction) and inflation targeting (no error-correction) as two extremes. Intermediate cases are equally interesting as they approximate targeting an average inflation rate over several periods. Since these two extremes and intermediate cases are nested, we can easily contrast various regimes in our simple dynamic general equilibrium model and derive a flexible empirical specification that allows us to determine the weights (coefficients) in the policy reaction function. We provide evidence that the error-correction model fits the late 1990s experience. When the Fed in the late 1990s took the gamble that economic growth had increased, inflation and inflationary expectations remained in check because the public believed that the Fed would quickly counter inflation if the projected rapid growth in productive capacity did not materialize. We show that an inflation targeter, who would not reverse errors, would have had much more difficulty keeping inflation under control during the 1990s than a central banker who puts some weight on price stability when setting interest rates. The organization of this paper is as follows. Section 2 discusses the economic events in the 1990s and how they relate to policy. These facts motivate the analysis by showing how surprising outcomes were in the second half of the 1990s from a historical perspective. Section 3 lays out a standard model that allows us to study policymaking when potential output is not observed. In this section, we show commitment to a stable price level is consistent with the stylized facts about the late 1990s. We also demonstrate that having the price level in the interest rate rule has significant benefits. In Section 4, we present empirical evidence suggesting that the price level was a part of the Fed's decision rule during the Greenspan era. This econometric evidence suggests that the Fed was indeed pursing the type of policy that our theoretical model shows would perform well against the shocks of the late 1990s. First, we show that a Taylor rule augmented with a price level term fits the data well in the Greenspan era. Second, we show that that the price level reverts after a CPI announcement surprise. Third, we present textual evidence that Alan Greenspan's rhetoric emphasizes the price level as opposed to inflation. Hence, the theoretical and empirical results of the paper present a consistent explanation of the good economic performance during the period when the Fed took its growth gamble. In Section 5, we present conclusions.

In the late 1990s, Alan Greenspan and the Federal Open Market Committee took a gamble. They kept interest rates essentially steady in the face of extraordinarily strong output growth and record-low unemployment rates. If the Fed had followed its historical reaction to economic conditions, it would have raised interest rates. Instead the Fed kept interest rates low because it believed that the productive capacity of the economy had expanded. Though it is now clear that there was such a shift, at the time the evidence for it was quite limited. Why did the gamble pay off? Even though the Fed turned out to be correct about potential output growth, why did inflation not increase as private agents took into account the risk that the Fed was being too stimulative when they set their price and inflation expectations? This paper provides an explanation for the success of the Fed in accommodating noninflationary growth in the late 1990s. A commitment of the Fed to reverse policy errors had it been too optimistic about the capacity of the economy can explain why inflation remained in check despite historically loose monetary policy. In particular, the paper shows that a price level target, either strict or partial, can serve to anchor inflation even if the public believes the Fed is overly optimistic about potential growth. The expectation that policy errors will be reversed leads to very favorable outcomes. The Fed is able to gamble on its belief that capacity has increased without the public raising inflationary expectations because the public knows that the Fed will reverse itself in the event the gamble does not pay off. The costs of reversing itself in this case are also low because the price level target keeps inflation low even in bad realizations. So only a modest disinflation is needed to reverse policy errors. The paper shows that price level targeting is superior to inflation targeting in a wide range of situations. The analysis does not depend on strict price level targeting. In particular, our formulation nests inflation targeting and strict price level targeting within a specification that allows for partial adjustment of the price level target to past inflation. For price level targeting to have these favorable properties, sufficiently many agents must be forward-looking. If not, then there is no value to the commitment to reverse errors. The paper also provides econometric evidence that the Fed, in contrast to earlier periods, has recently put substantial weight on the price level in setting interest rates. A generalized Taylor rule puts weight on the price level gap as well as the inflation gap in the Greenspan era. In the pre-Greenspan period, there is no evidence of price level targeting. Moreover, the presence of the price level gap in the Taylor rule provides an explanation that does not rely on interest-rate smoothing or serially correlated policy errors for the significance of lagged interest rates in standard Taylor rules. Price level targeting induces error-correction that resembles serial correlation of interest rate decisions in specifications that do not control for the price level gap. We also find that the price level reverts significantly in response to CPI announcement surprises. This finding provides complementary evidence that the Fed has a policy to offset surprises rather than a policy of letting bygones be bygones. The degree of the price-level reversion matches almost exactly what we estimate based on the generalized Taylor rule. Finally, we perform a content analysis of the language of Alan Greenspan, Alan Blinder and Ben Bernanke. This analysis supports the claim that Alan Greenspan puts a relatively greater weight on the price level when compared to either Ben Bernanke, who advocates explicit inflation targeting, or Alan Blinder, who does not advocate explicit rules. Our analysis does not ascribe strict adherence to a price level target, or indeed any other rule, to Alan Greenspan or the Fed. Rather, it shows that there is both theoretical and empirical support for the contention that commitment to correcting past errors was a part of Federal Reserve decision making in recent years and that this commitment contributed to the superior performance of the US monetary policy in the late 1990s. Our nesting of inflation targeting and price level targeting shows that these two policies lie on a continuum. If an inflation-targeting central banker has a preference for being on average in the middle of an inflation target range, then inflation targeting will partially anchor the price level relative to a deterministic trend. Our model of partial price level targeting is a way to build in this tendency to offset past shocks in some inflation targeting regimes. Consequently, our analysis admits the possibility of a high degree of continuity between the policy followed under Alan Greenspan's and Ben Bernanke's leadership.