تنظیم الگوهای دستمزد و سیاست های پولی: شواهد بین المللی

Systematic differences in the timing of wage-setting decisions among industrialized countries provide an ideal framework to study the importance of wage rigidity for the transmission of monetary policy. Synchronization in wage-setting decisions is prevalent in Japan and the United States, yielding varying degrees of wage rigidity within the year; instead, in France, Germany, and the United Kingdom decisions are more uniformly spread over time. Exploiting within-year variation in the timing of wage-setting decisions in these economies, we find support for the long-held but scarcely tested view that wage rigidity plays a critical role in the transmission of monetary policy.

A wide body of empirical evidence suggests that monetary policy has an important effect on the behavior of real variables at business-cycle frequencies. Most theoretical models that seek to identify the connection between nominal causes and real effects posit some form of nominal rigidity in wages and (or) prices.1 Empirical evidence assessing the extent of nominal wage rigidity and its relevance in the transmission mechanism from monetary policy to real variables is, however, regrettably scarce.2 This paper attempts to partially fill this empirical void by providing a study that exploits differences in the effective degree of nominal wage rigidity within and across countries. Synchronization of wage-setting decisions varies significantly across advanced economies. In Japan, the best-known example of synchronization of wage-setting decisions, the majority of firms set wages during the first and second quarters of the calendar year in what is known as “Shunto”, (or spring offensive), and wages remain in place until the following year. In the United States, the available evidence suggests that a large fraction of firms set wages once a year, typically at the end of the calendar year. In contrast, wage-bargaining renegotiations in Germany take place throughout the year, and contracts tend to last one to three years. Theories of the transmission of monetary impulses to real variables based on wage rigidity would hence predict that, other things equal, monetary policy innovations in Japan should have a larger effect when the shock takes place in the second half of the year, that is, after the Shunto has occurred and wages are relatively rigid. In the United States, the effect should be larger when the shock occurs in the first half of the year, as wages tend to be reset at the end of the calendar year. However, in Germany, where there appears to be little bunching in wage-setting decisions within the year, the effect should not vary with the quarter in which the shock takes place. The aim of this study is to test whether these predictions find support in the data. More precisely, this study assesses whether the response of the economy to monetary policy shocks differs according to the time of the year in which the shock takes place and whether this difference can be reconciled with the observed variation in the timing of wage-setting decisions. To this end, this study introduces quarter dependence in an otherwise standard, recursive VAR setup and analyzes the empirical impulse responses of aggregate variables to a monetary policy innovation in five large and highly developed countries. The countries considered are France, Germany, Japan, the United Kingdom, and the United States. The focus on these countries is related to the extant literature on central banking practices: The wider consensus in the literature on the monetary instruments used by these countries’ central banks provides a natural baseline from which we deviate to study the potential for seasonal dependence in monetary policy effects.3 Our empirical exercise has a “difference-in-difference” flavor, in that it tests for potential differences in the effect of monetary policy across quarters of the calendar year for each of the countries considered, and then relates the findings across countries to each country's degree of wage rigidity over the calendar year.4 The findings indicate that, for both Japan and the United States, there are, indeed, important differences in the response of the economy to monetary policy shocks that depend on the timing of the policy innovation. These differences, in turn, can be related to the differing degree of wage rigidity across the calendar year. Specifically, a monetary policy innovation in Japan that occurs during the first or second quarter – that is, during the Shunto period in which wages are being reset – has a relatively small effect on output, whereas an innovation in the third quarter – that is, immediately after the Shunto – has a remarkably large effect. The pattern is reversed in the United States: A monetary policy innovation in the first half of the calendar year has a significantly larger effect on output, whereas an innovation in the second half has a relatively small effect. Again, this pattern conforms well with the degree of wage rigidity in the United States, which is high in the first half of the year and low in the second half. In sharp contrast, in Germany, France, and the United Kingdom, where the degree of wage rigidity is more uniform within the year and the contracts are of longer duration, the quarter in which a monetary policy shock takes place appears to be less relevant. Our findings for the United States essentially replicate those in Olivei and Tenreyro (2007). This paper extends their empirical analysis to test whether the degree of synchronization in wage-setting decisions also matters for the transmission of monetary impulses in countries other than the United States. Overall, the findings complement and reinforce their conclusion that wage rigidities can play an important role in the transmission of monetary policy. The remainder of the paper is organized as follows. Section 2 briefly describes various pieces of evidence on wage-setting patterns and the policy strategies used by the countries’ central banks. Section 3 presents the empirical method and introduces the data. Section 4 describes the dynamic effects of monetary policy on different macroeconomic aggregates. Section 5 discusses the robustness of our findings, and Section 6 provides concluding remarks.

Our main conclusions have been amply foreshadowed. The empirical analysis indicates that the degree of bunching of wage-setting decisions matters for the transmission of monetary policy to the real economy. In Japan, wage setting has conformed to a synchronized pattern in the form of the annual Shunto and the associated process of collective bargaining. In the United States, various sources of anecdotal evidence point to a significant fraction of firms adjusting wages in the last quarter of the calendar year. One critical implication of this synchronized annual wage setting is that if preset wages are important in accounting for the connection between monetary policy and real activity at business cycle frequencies, then the transmission of a monetary impulse to the real economy should differ according to when the impulse occurs within a calendar year. Specifically, a monetary policy shock in Japan that occurs in the first part of the year, that is, when the Shunto is taking place, should have a small impact on output, since this is a period of relative wage flexibility. In contrast, a shock occurring later in the calendar year should have a larger impact on real activity, because at this time of the year wages are relatively rigid. In the United States, the timing should be reversed: A monetary policy shock in the first half of the calendar year should have a relatively large impact on real activity, whereas a shock in the second half should have a smaller impact. An empirical analysis of the transmission of monetary policy shocks to the real economy based on a quarter-dependent VAR supports this claim for Japan and the United States. We contrast the empirical findings for Japan and the United States with those for Germany, France, and the United Kingdom. In these latter countries, synchronization in wage setting has been lower, with wage bargaining more uniformly distributed across the calendar year and wage contracts lasting for longer than a year in some instances. Correspondingly, the response of activity to a monetary policy shock has been more uniform across quarters. This paper makes no claim as to whether synchronization of wage changes is preferable to uniform staggering. This is a problem that has been studied in the past, and the general finding of this literature is that synchronization is the equilibrium timing in many simple Keynesian models of the business cycle.43 Yet, the new generation of Keynesian models has glossed over this finding and assumed uniform staggering as both a convenient modeling tool and an essential element in the transmission mechanism of monetary policy shocks. This paper notes that while uniform staggering may be a realistic assumption for some countries, it is not for others. For these other countries, the empirical implications of non-uniform wage staggering can be important and should be taken into consideration from a modeling standpoint. Finally, it is interesting to note that there appears to be some synchronization in the timing of wage and price changes. In this respect, January is the month when the frequency of adjustment for wages and prices is relatively high in some countries.44 The extent to which this relationship is causal, with the seasonality in wage changes imparting seasonality to price changes, is a topic that deserves more research.