The role of inflation regime in the exchange rate pass-through to import prices

Abstract

We examine exchange rate pass-through (ERPT) into aggregate import prices for nine OECD countries in view of Taylor's [Taylor, J., (2000). Low Inflation, Pass-through, and the Pricing Power of Firms, European Economic Review 44 (7), 1389–1408] suggestion that the degree of pass-through is dependent on the importing country inflation regime. Extending the standard mark-up pricing model under monopolistic competition to a setting where the pricing decision is dependent on the importing country inflation experience, we present strong empirical evidence for the inflation regime dependence of pass-through elasticities for the analyzed countries based on nonlinear estimation methodology. Our main result is that the pass-through is highly incomplete and positively correlated with the importing country inflation environment, when the empirical analysis is conducted using nonlinear threshold or smooth transition estimation techniques.

Introduction

Exchange rate pass-through (ERPT) has been declining in recent years (see e.g. Campa & Goldberg, 2005 and Goldberg & Knetter, 1997), and the famous Taylor (2000) argument for the reason of it is the low inflation regime in industrialized countries.¹ He has argued that firms adopting staggered prices are more likely passing through the exchange rate and other cost changes to import prices when inflation is high. According to Taylor's idea firms find it difficult to fully pass-through the exchange rate changes to their export prices because of the intensified worldwide competitive pressure and low and stable inflation regimes that we have seen in most of the industrialized countries recently. For example the U.S. inflation rate did not rise even though there were strong demand pressures in the economy in the late 1990s. In addition, the U.S. import prices did not rise much in the late 1980s either, despite the fact that the U.S. dollar depreciated drastically. Similar empirical evidence can also be found in the ERPT for Japan (see Otani, Shiratsuka, & Shirota, 2003).

ERPT has played a central role in the debates on monetary policy and optimal exchange rate regime. For monetary authorities it is important to know how the previous exchange rate changes have affected the current import prices and how the exchange rate movements will affect the inflation forecasts of the central bank. Hence, the knowledge of the degree of the pass-through is crucial for conducting appropriate monetary and/or exchange rate policy both in ex post and ex ante terms. If the pass-through were due to sticky prices, it would probably be dependent on the stance of monetary policy as suggested by Taylor. Furthermore, Ball (1999) has pointed out that actually the monetary policy rules might be dependent on the degree of pass-through, so the important question is whether looser monetary policy (causing higher inflation) might also cause more frequent changes in firms' export prices.

There is also some evidence that the pass-through from exchange rate changes to prices might be nonlinear. Devereux and Yetman (2010) examined over 100 countries and found that there is a positive, but nonlinear relationship between exchange rate pass-through and average inflation rate. The nonlinearity in their analysis comes from the finding that when annual inflation rises above some threshold level, there should be no further impact to domestic import prices, because all prices adjust continuously, and hence, the exchange rate pass-through is complete. Furthermore, some of the factors in the new open economy macroeconomic models give support to the nonlinear response to exchange rate shocks. A small change in exchange rates may leave the product prices unchanged in importing country currency terms due to, for example, menu costs or contract renegotiation costs. However, large and persistent shocks may trigger price adjustment for several firms and at the aggregate level this can be seen as a change in the pass-through elasticities.

In previous studies e.g. Al-Abri and Goodwin (2009) have also applied nonlinear empirical methods to ERPT analysis. They found that the degree of pass-through in the data set for 16 OECD countries improves clearly from the previously observed 50% average values when the threshold effects are taken into account in using cointegration techniques for the analysis of ERPT effects. However, they did not analyze the role of inflation environment explicitly in the spirit of Taylor (2000) suggestion at all.² Prior to our study, in addition to Devereux and Yetman (2010) the actual Taylor (2000) idea on the role of inflation environment in the ERPT analysis has most recently been explicitly studied e.g. in Sekine (2006) and Nogueira Júnior and León-Ledesma (2010). Sekine focused on two ‘stages’ of the ERPT, i.e., the first-stage pass-through that captures the impacts of exchange rate fluctuations to import prices, and the second stage, that relates the effects of changes in import prices to the changes in consumer prices. The main result was that in both stages the pass-through has declined over time for all the analyzed six major industrial countries, i.e., the U.S., the UK, Japan, Germany, France and Italy. The analysis was based on a time-varying parameter stochastic volatility model, so the empirical approach was somewhat different compared to ours, but the main result would seem to be similar, i.e., that the lower values of pass-through were clearly positively correlated with lower values of average inflation. The empirical results in Nogueira Júnior and León-Ledesma (2010) were not very much in favor of the Taylor suggestion of positive correlation between the degree of ERPT and level of average inflation. They used a state space model that allowed time variation in the ERPT and its dependence on the inflation environment. They estimated the model for 12 developed and emerging economies and found that there has been a smooth decline in the impact of exchange rates on domestic inflation, but that the hypothesis that lower inflation precedes the declining degree of ERPT was mainly not supported in the data from the beginning of 1980 until the last quarter of 2007. We differ from their approach in the form of the nonlinear modeling procedure (state space representation and Kalman filtering compared to the smooth transition or threshold modeling in our paper). In addition, we have a longer data set, and we concentrate solely on the data from the OECD countries, so our country set coincides with their set only for the part of Australia, Denmark and the UK. Our results show clearly stronger positive correlation between the degree of exchange rate pass-through and the inflation rate.

We extend the recent pass-through literature in two ways. First, we present a model that adds the importing country inflation regime as an explanatory variable for the pass-through coefficient. Specifically, we assume that exporting firms use the present annual inflation rate of the importing country as an indicator of future pass-through for their product prices. In other words, we assume that the firms' mark-up is time-varying rather than constant. This implies also the time dependence of the exchange rate pass-through elasticities. Basically, our model is a conventional mark-up pricing model but we suggest that a change from a high importing country's inflation regime to a lower inflation regime alters the impact of exchange rate pass-through to export prices of firms. We want to show that the inflation regime affects the exchange rate pass-through nonlinearly when assuming that the mark-up of firms is dependent on the inflation regime of the importing country.³ For the empirical analysis of this idea we use the nonlinear estimation methodology (see Granger and Teräsvirta, 1993, Teräsvirta, 1998 and Tong, 1990, among others).

Overall, our main idea obtains strong support in the data from 9 OECD countries. We find a positive but nonlinear relationship between import price changes and exchange rate pass-through and the hypothesis of partial ERPT applies for most of the analysed countries. Hence, our results support the Taylor (2000) suggestion of positive correlation between exchange rate pass-through and inflation. In addition, the pass-through coefficients do not vary as much across countries as in many other previous studies.

The remainder of the paper is organized as follows. In Section 2 we give the theoretical and empirical frameworks for the analysis of exchange rate pass-through in this study. Section 3 gives the description of the data and basic statistics. In Section 4 the most important results from the empirical analysis are presented. Finally, Section 5 concludes.