تاثیر تعداد مداخله در بازار ارز: تجربه ژاپن

This paper examines the effects of the frequency of foreign exchange intervention on exchange rate volatility. Japanese intervention is characterized by differences in the frequency of intervention—there are high and low frequency intervention periods. Using the GARCH methodology, this paper models the changes in the yen/dollar exchange rate, with the frequency of intervention from April 1991 to December 2005 as the focal explanatory variable. The empirical results show that high frequency intervention stabilizes the exchange rate by reducing exchange rate volatility and that low frequency intervention is more effective.

This paper focuses on the relationship between the frequency of foreign exchange intervention and exchange rate volatility. I consider Japanese foreign exchange intervention from April 1991 to December 2005.1 Since July 2001, the Japanese Ministry of Finance has been disclosing daily intervention records from April 1991 to date.2 The largest Japanese intervention reported was the selling of 2620 billion yen (about 26 billion dollars at 100 yen/dollar) against the U.S. dollar on April 10, 1998. Japanese intervention is characterized by differences in the frequency of intervention—there are both high and low frequency intervention periods. Fig. 1 depicts the number of intervention days per year. High frequency intervention by the Japanese authority can be observed from April 1991 to May 1995 and from 2003 to 2004. On the other hand, June 1995–December 2002 and the year 2005 were periods of low frequency intervention. These differences in frequencies are caused in part by the policy stance adopted by the person who holds the office of the Vice Minister of Finance for International Affairs. For example, shortly after Mr. Mizoguchi acceded to the post of Vice Minister on January 14, 2003, there were 8 interventions in January itself, and interventions were implemented tri-daily during his term in office. Prior to Mr. Mizoguchi's inauguration, there was no intervention from June 2002 to January 13, 2003.3 Since July 2004, that is, after the next Vice Minister—Hiroshi Watanabe—took over, there have been no interventions in the exchange market.4 Therefore, the Japanese exchange rate system can be classified as a managed floating system, because the Japanese authority frequently intervenes in the foreign exchange market. Full-size image (7 K) Fig. 1. Number of intervention days per year. Note: This figure shows the number of intervention days per year. The data source is the homepage of the Japanese Ministry of Finance (http://www.mof.go.jp/feio/034_133.htm). For 1991, the number of days is calculated from April 1 to December 30. Figure options Previous studies on foreign exchange intervention have scarcely paid attention to the effect of intervention frequency.5 To begin with, it will be useful to examine the two effects of foreign exchange market intervention, namely, the effect of interventions (1) on the exchange rate level (first moment) and (2) on exchange rate volatility (second moment). Ito (2002) analyzed the effect of Japanese intervention on the exchange rate level, using the daily intervention records disclosed by the Japanese authority. He concluded that a Japanese intervention of 100 billion yen moved the yen/dollar rate by 0.1%, and a U.S. intervention of 1 billion dollars had a 5% effect. With regard to the frequency of intervention, Ito (2002) used the initial intervention variables that are of a certain magnitude if there is no intervention during the 5 preceding business days, and 0 otherwise. He reported that the initial intervention has a larger effect on the exchange rate level than do other interventions. Further, he stated that the intervention criterion is 125 yen/dollar. If the yen appreciates, that is, the exchange rate rises above 125 yen/dollar, the Japanese authority implements a yen-selling intervention; if not, it implements a yen-purchasing intervention. The exchange rate level desired by the Japanese authority can be considered to be 125 yen/dollar, as depicted in Fig. 2. This figure shows the relationship between intervention (vertical axis, 100 million yen) and the exchange rate (horizontal axis, yen/dollar) from April 1991 to December 2005 in the Tokyo Foreign Exchange Market. It is evident that during the 14-year span, 125 yen/dollar was the benchmark for intervention. Ito (2002) also finds intervention to be effective in all samples, excluding the 1991–1994 interventions. I believe that the interventions were not effective because the frequency of intervention was high during 1991–1994 and the magnitude of intervention was small. Thus, targeting a specific level was not the main purpose of the intervention during this period.6 Full-size image (5 K) Fig. 2. Exchange rate and the magnitude of intervention. Note: This figure shows the relation between intervention (vertical axis, 100 million yen) and the exchange rate (horizontal axis, yen/dollar) of the Tokyo Foreign Exchange Market from April 1991 to December 2005. The yen-selling interventions against the dollar are shown by a positive sign, and the yen-purchasing interventions are shown by a negative sign. Figure options Numerous papers focus on exchange rate volatility and intervention as effects of the second moment.7Dominguez (1998) and Nagayasu (2004) indicated that intervention generally increased exchange rate volatility. These results are surprising because the purpose of foreign exchange intervention “is to stabilize the exchange rate of the national currency.”8 Stabilizing the exchange rate could also be understood in terms of a reduction in exchange rate volatility. Thus, it is unlikely that the authority implements intervention in order to increase exchange rate volatility. However, the empirical results suggest that foreign exchange intervention increases this volatility. This paper is organized as follows. Section 2 examines how the frequency of foreign exchange intervention can influence exchange rate volatility. Section 3 includes a brief discussion of the data. Section 4 presents the model and results, and Section 5 is the conclusion.

The aim of this paper is to clarify the effect of intervention frequency on the foreign exchange market. The conjecture effect of intervention frequency on exchange rate volatility and the changes in exchange rate as a result of Japanese intervention from 1991 to 2005 are discussed. Empirical results suggest that the frequency of intervention has two different effects. First, frequent intervention stabilizes the exchange rate by reducing volatility. Second, low frequency intervention has a large effect on the exchange rate level. This implies that the authority has two policy objectives, namely, exchange rate level (first moment) and exchange rate volatility (second moment). The results may be the key to the “secrecy puzzle” proposed by Sarno and Taylor (2004). Most actual intervention operations in the foreign exchange market are secret interventions—the authority generally implements a secret intervention when the intervention is highly intensive. If the authority's objective is to reduce exchange rate volatility, it is likely to implement secret interventions with high frequency. On the other hand, if the objective is to change the exchange rate level, then the authority is likely to intervene with low frequency and by officially announcing its exchange rate policy. The relation between secret intervention and the frequency of intervention is still ambiguous.