What Has Become of the "Stability-Through- Inflation" Argument?

The purpose of this article is to examine the status of a well known argument for a positive rate of steady state inflation. The original argument, by Vickrey (1954) and Phelps (1972), suggests that the economy is more able to dampen shocks, such as fluctuations in the real rate of return on risky capital investment, when inflation rates are positive. The notion is that the conditions required for stability are sturdier at higher rates of inflation, and thus at higher nominal rates of interest, making the economy as a whole less vulnerable to stochastic shocks. This has sometimes been forwarded as an argument for stability through inflation. Many economists have no doubt come away from this literature thinking, albeit probably vaguely, that a little inflation is a good thing. Echoes of this hypothesis can be heard today as many economists wonder aloud whether monetary policymakers should proceed to lower inflation rates. Our view is that the “stability-throughinflation” argument has not stood the test of time very well. On the one hand, the rational expectations revolution destroyed much of the argument’s foundation by insisting on agents that adjust their expectations very rapidly and hence immediately violate the stability condition derived in the previous literature. The down side of the rational expectations innovation was that it left theorists arguing that a likely outcome was a stationary equilibrium on the high-inflation side of the Laffer curve. On the other hand, more sophisticated treatments of the adaptive expectations hypothesis, mostly appearing in the recent literature on learning in macroeconomic models, have found less tendency toward instability at low inflation rates. They suggest, instead, that relatively low inflation rates were associated with stability in the learning dynamics, and relatively high inflation rates were associated with instability in the learning dynamics. In these systems, higher inflation is often more variable inflation. In either case— the rational expectations case or the learning case—the notion that low rates of inflation can generate instability is, if not altogether absent, at least much less likely. Thus we conclude that the stability-throughinflation argument is effectively defunct. We proceed by summarizing the original stability-through-inflation arguments and the grounds on which they were criticized in the literature. We then summarize the rational expectations solution. The subsequent portion of the paper then turns to a discussion of the behavior of the model under what we regard as more sophisticated treatments of the adaptive expectations hypothesis. In the second half of the paper, we turn to recasting the argument in a version of the model with fixed real government deficits financed by seignorage revenue. The final section provides a summary.