در زمینه عوامل مداخلات "کوچک" و "بزرگ" بازار ارز: مورد مداخله ژاپن در دهه 1990

During the past 30 years, central banks have often intervened in foreign exchange markets, and the magnitude of their foreign exchange market interventions has varied widely. We develop a quantitative reaction function model that renders it possible to examine the determinants of “small” and “large” interventions. We apply the model to analyzing the intervention policy of the Japanese monetary authorities (JMA) in the yen/U.S. dollar market during the period from 1991 through 2001. To this end, we use recently released official data on the foreign exchange market interventions of the JMA. We find that the JMA tended to conduct large interventions when the yen/U.S. dollar exchange rate drifted away from an “implicit target exchange rate.”

During the past 30 years, central banks have often intervened in foreign exchange markets in an attempt to manage exchange rates. Therefore, it is not surprising that a substantial body of literature has analyzed the factors that led central banks to intervene in foreign exchange markets. When analyzing these factors, most researchers have regularly assumed that central banks decide merely whether or not to intervene in the foreign exchange market. As a consequence, relatively little is known about the factors that determine the magnitude of central banks' foreign exchange market interventions. In this article, we develop a quantitative model that aims at examining whether central bank take specific economic factors into consideration when deciding on the magnitude of their foreign exchange market interventions. Identifying the factors behind the magnitude of interventions can be regarded as important for at least two reasons. First, the empirical evidence on the intervention policies of the central banks of all major industrialized countries suggests that the magnitude of central banks intervention during the last 30 years has varied significantly. Hence, the determinants of foreign exchange market interventions cannot be completely understood as long as empirical research has not uncovered the factors that determine the magnitude of interventions. Second, it can be assumed that a large central bank intervention demonstrates the willingness of a central bank to calm “disorderly markets” and to correct an exchange rate level it finds inappropriate. This assumption suggests that understanding the determinants of the magnitude of central bank foreign exchange market interventions can improve our understanding of the factors central banks deem important when deciding how forcefully and to what extent they should intervene. In the empirical literature on foreign exchange market interventions, the factors triggering interventions are usually identified by estimating reaction functions of central banks. We follow a similar route and estimate a quantitative reaction function model that belongs to the class of so-called qualitative dependent variable models. Such models have recently been used in the empirical literature on foreign exchange market interventions of central banks by, for example, Baillie and Osterberg (1997a), Beine, Benassy-Quere, and Lecourt (2002), and Dominguez (1998).1 However, the qualitative dependent variable models that have often been used in this literature are in general not suited to shed light on the factors that determine the magnitude of central banks' foreign exchange market interventions.2 Therefore, we use an alternative qualitative dependent variable model. More specifically, we apply the ordered probit model. Recently, the ordered probit model has been successfully applied by, for example, Davutyan and Parke (1995) and Eichengreen, Watson, and Grossman (1985) to estimate reaction function models describing central bank discount rate policy. We apply the ordered probit model to analyze the intervention policy of the Japanese monetary authorities (henceforth referred to as JMA).3 To this end, we make use of the fact that the JMA recently released a comprehensive data set on their interventions in foreign exchange markets during the period from 1991 through 2001. This release of data on their foreign exchange interventions constituted a significant change in the information policy of the JMA. Until the beginning of 2002, the JMA did not publish any data on their foreign exchange market interventions. In fact, official intervention data were in general only made available to researchers mainly by the Fed and the Deutsche Bundesbank. For this reason, most empirical studies have focused on the central bank reaction functions and, thus, the determinants of the intervention activities of the Fed and the Deutsche Bundesbank (see, e.g., Almekinders & Eijffinger, 1994, Almekinders & Eijffinger, 1996 and Baillie & Osterberg, 1997a). Only a few authors, like Baillie and Osterberg (1997b) and Humpage and Osterberg (1992), had access to Japanese intervention data. Recently, Hillebrand and Schnabel (2003) and Ito (2002) have used the published official JMA intervention data to analyze the Japanese foreign exchange market interventions. We find that the variables commonly used in the empirical literature to estimate central bank reaction functions had statistically and economically significant explanatory power for the decision of the JMA to conduct large interventions. Specifically, we find that the absolute deviation of the yen/U.S. dollar rate from the “implicit exchange rate target” of 125 yen/U.S. dollar reported by Ito (2002) had a relatively strong influence on the propensity of the JMA to conduct large foreign exchange market interventions. This supports Ito's conjecture that the JMA had an implicit exchange rate target in mind when conducting their foreign exchange rate market interventions. Our results further indicate that, in the first half of the 1990s, the JMA tended to use “small” foreign exchange market interventions as a tactical instrument allowing for reactions to shorter run developments in the foreign exchange market. In the second half of the 1990s, large interventions dominated. We organize the remainder of this article as follows. In Section 2, we describe the data set we use in our empirical analysis. In Section 3, we explain the empirical central bank reaction function model that forms the basis of our study. In Section 4, we briefly motivate the set of explanatory variables we use to describe the interventions of the JMA in the yen/U.S. dollar market. We also present our estimation results. In Section 5, we offer some conclusions.

The results of our empirical study suggest that a widening of the absolute wedge between the yen/U.S. dollar exchange rate and the “implicit target” exchange rate of 125 yen/U.S. dollar exerted a statistically significant effect on the propensity of the JMA to conduct large foreign exchange market interventions. By contrast, the absolute deviation of the yen/U.S. dollar exchange rate from its moving average was more important for small interventions. We also find that the intervention behavior of the JMA changed when a new director general of the International Finance Bureau was appointed in 1995. While in the first half of the 1990s, the JMA conducted a number of smaller interventions, it conducted relatively large interventions in the second half of the 1990s. The broad picture that emerges from our empirical analysis leads us to conclude that the JMA primarily used larger interventions to combat deviations of the yen/dollar rate from the implicit exchange rate target. Moreover, our results suggest that particularly in the first half of the 1990s, the JMA tended to use smaller foreign exchange market interventions as a tactical instrument to smoothen the exchange rate path. This could have been based on a rationale that has to do with how interventions can affect the exchange rate. Given the fact that the interventions of the JMA were not made public at the time and that it may be that smaller interventions can be kept secret more easily than larger interventions, smaller interventions could have been seen as an instrument to affect the exchange rate through creating rumors that the central bank had conducted foreign exchange market interventions.