Academic journal article Journal of Money, Credit & Banking

What Makes the Output-Inflation Trade-Off Change? the Absence of Accelerating Deflation in Japan

Academic journal article Journal of Money, Credit & Banking

What Makes the Output-Inflation Trade-Off Change? the Absence of Accelerating Deflation in Japan

Article excerpt

THE ORIGINAL PHILLIPS CURVE WAS NONLINEAR: Alban W. Phillips (1958) estimated a nonlinear relationship between nominal wage inflation and the unemployment rate in the United Kingdom. Since that time, it has become standard to model the short-run Phillips curve as a linear relationship with a time-invariant slope. The present paper argues that this simplifying assumption is not as innocent as it seems.

Our paper assesses the empirical performance of two classes of models in which the slope of the Phillips curve varies endogenously over time. The model classes differ according to the set of variables determining the slope of the output--inflation trade-off.

In Ball, Mankiw, and Romer (1988) and Dotsey, King, and Wolman (1999), trend inflation is among the determinants of the Phillips curve slope. In these endogenous pricing models, the frequency of price adjustment depends on firms' optimizing decisions. For instance, a decrease in trend inflation causes firms to adjust prices less frequently, which in turn implies a flatter Phillips curve.

In Lucas (1973), the slope of the Phillips curve depends on the volatility of aggregate demand and supply shocks. For instance, if aggregate volatility decreases, a larger fraction of any change in the overall price level is misperceived by firms as being a change in their relative price. In that scenario, any change in aggregate demand has a larger impact on firms' production. Correspondingly, any change in output will be associated with a smaller change in inflation. That is to say, the Phillips curve flattens.

The working paper version of our paper also evaluates the performance of a third class of models. In papers such as Laxton, Meredith, and Rose (1995), the short-run Phillips curve is potentially nonlinear due to the presence of capacity constraints. De Veirman (2007) documents that this hypothesis is consistent with the data but performs poorly in comparison with the other two models.

To test the theories of endogenous variation in the Phillips curve slope, we gather evidence from Japan. The period 1991-2002 in Japan can be characterized as a succession of recessions, interrupted only by brief or limited recoveries. Standard estimates suggest that the output gap was negative for most of that period. Initially, inflation declined, with core consumer price index (CPI) inflation reaching the zero level in the mid-1990s and turning negative in the second half of the 1990s. After 1998, however, annual core CPI inflation remained fairly stable at moderately negative levels, reaching its trough at -0.79% in 2002.

As this paper documents, the fact that deflation remained surprisingly mild notwithstanding a relatively long period of negative output gaps presents a puzzle to anyone who takes a standard linear Phillips curve literally. This makes Japan a particularly interesting test case for assessing the nature of the output--inflation trade-off.

An advantage of using recent data for Japan is that, unlike the samples typically used in earlier empirical tests of the theories with endogenous variation in the Phillips curve slope, our sample includes a fairly large number of observations from the region of the Phillips curve at which inflation is near-zero or negative. The inclusion of low- and negative-inflation observations increases the chance that we will be able to distinguish the competing models empirically.

Throughout this paper, we use backward-looking Phillips curves in which expectations of current inflation are proxied by inflation lags. Our evidence does not directly speak about the behavior of Phillips curves in which expectations are at least in part forward looking. (1) Note that there is still some discussion in the literature regarding the empirical performance of marginal cost-based Phillips curves with forward-looking elements. (2) Our paper's starting point is that even a conventional backward-looking Phillips curve, with an ability to fit the data which is normally considered to be its relative strength, does not explain the data well in Japan. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.