Regime and Underlying Inflation Dynamics: ¿Generalized Comovement or Relative Price Adjustment?

Inflation is usually defined as a generalized and persistent change in the price level. But this notion seems to be restricted to à la Cagan high inflations of a monetary origin, in which absolute price change prevails. It is not obvious, though, that inflation dynamics is always dominated by a strong comovement in sectoral prices. Nor that absolute price changes predominate over relative price adjustment. What is more, it is expected that in normal times, when inflation remains at a low level, transitory movements in prices explain a high portion of inflation variability, because inflation does not follow a trend but instead evidences erratic movements that on average tend to cancel out. Under high inflation, the presence of a trend in aggregate inflation reflects in a relatively higher importance of the common component in price variations. That is, more comovement in price adjustments, which not necessarily implies the absence of persistent, medium term relative price adjustments. Aggregate inflation is the result of multiple price setting decisions in response to changes in costs of production, demand conditions and policy signals. The type of response to these impulses depends on the environment where agents make these decisions. Inflation dynamics is, thus, “regime specific” because it depends on the way economic policy, in general, and monetary policy, in particular, operate. The regime dependence of inflation dynamics was initially noted by Fisher (1982), who emphasized the role loose monetary policies could have had in perpetuating supply shocks in the United States during the ‘70s. More recently, Ball and Mankiw (1995), also discussed how sectoral price responses to shocks could be influenced by the inflationary environment. Notwithstanding this long tradition in considering inflation dynamics as “regimedependent”, the concept of regime has remained fairly vague in the literature, partly because it is an “unobservable” associated to institutional factors that define a framework to interactions between economic policy and private agents. One can try to identify regime by the observable consequences of these interactions. In particular, it can be thought that the average inflation that characterizes an economy during a period of time, the trend inflation, can be a proxy for monetary regime. This is in fact what recent literature does (see Kiley, 2006; Blake and Fernández-Corugedo, 2006; and Ascari and Ropele, 2007; among others). But trend inflation as a criteria to identify regimes can be imprecise, because it ignores sectoral prices adjustment dynamics behind a specific inflation trend. This omission can lead to at least two types of miss-specifications of regime. On the one hand, in terms of underlying relative price dynamics, a regime can precede its positive manifestation when a shock takes place. On the other hand, typical relative price adjustment dynamics cannot be the same for all economies. For small open economies, frequently subject to external disequilibrium, it is economic policy itself that is an important source of relative price variability, usually through the correction of disequilibrium in the real exchange rate (RER). In this case, the tradable-non tradable price adjustment seems to be relevant to explain aggregate price dynamics. In contrast, for industrial economies shocks to energy and food often seem to be the prevailing source of inflation variability. In both cases, it is interesting to study how the transmission of aggregate shocks of different nature can change in terms of its generalization and persistence depending on the inflationary environment. With this aim, we study inflation dynamics and its relation with inflationary regimes in Argentina and the United States during the last forty years, when both countries experimented high, moderate and low inflations. The differences in size and level of development between both economies are very significant, as it is also the case for the type of shock to which they are typically subject to. In the United States shocks to energy and food prices have hit the economy not only during the ‘70s and ‘80s, but also during the ‘90s and ‘00s. In Argentina, policy shocks, usually associated to devaluations of the currency to correct external disequilibrium or to stabilize inflation using the exchange rate as an anchor, have been dominant. We found that there is far more than a generalized and persistent movement in prices behind inflation dynamics. It also reflects persistent relative price adjustment, different from the idiosyncratic, short-run noise. Our results also show that the identification of monetary regimes only by the trend inflation can be fairly vague and limited, as it ignores underlying relative price dynamics. On the one hand, the relative importance of relative price adjustment vis à vis generalized comovement in prices depends on the regime. Under high inflation, when nominal impulses are an important source of inflation variability, comovement acquires relatively higher importance than relative price adjustment. On the contrary, when inflation is low, the converse is true. On the other hand, the incidence of aggregate shocks on inflation dynamics generally manifests in the rise of the comovement between sectoral inflations, that is, the higher inflation, the more is the comovement persistence induced by aggregate shocks. It is to note, also, that the transition from regimes of moderate inflation to regimes of high inflation is a slow process: comovement increases as inflation rises, showing no discontinuities. Also, the different nature of typical aggregate shocks can impose distinct features to inflation dynamics. For Argentina we show evidence that the importance of the tradable-non tradable dynamics is common to all regimes. For the United States, the differential adjustment between energy and food prices and the rest seems to be relevant. These distinct features of inflation dynamics should be incorporated to the modeling of inflation with monetary policy purposes. They should also be taken into account when choosing a monetary policy relevant measure of core inflation: results suggest that an ex energy & food core seems relevant for the United States, but not at all for Argentina. Finally, our results show that under a low inflation environment supply shocks become more idiosyncratic. This is clear in the case of Argentina from the low pass-through on prices after 2002 devaluation vis à vis other devaluations occurred under higher inflation environments. In this sense, a general conclusion is that macroeconomic contexts of high inflation tend to impede conventional mechanisms of price adjustment to act, favoring aggregate shocks to induce a generalized comovement with a persistent impact on the price level.