مداخلات بانک مرکزی و نرخ ارز : تجزیه و تحلیل با داده های فرکانس بالا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|24627||2000||14 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of International Financial Markets, Institutions and Money, Volume 10, Issues 3–4, December 2000, Pages 349–362
We use high frequency data for the mark–dollar exchange rate for the period 1992–1995 to evaluate the effects of central bank interventions on the foreign exchange market. We estimate an unobserved component model that decomposes volatility into non-stationary and stationary parts. Stationary components in turn are decomposed into seasonal and non-seasonal intra-day parts. Our results confirm the view that interventions are not particularly effective. The exchange rate moves in the desired direction for only about 50% of the time, and often with a substantial increase in volatility. The model suggests that the two components, which are affected the most by interventions, are the permanent and the stochastic intra-day.
Central banks may decide to intervene in the foreign exchange market for various reasons and with various styles. In a fixed exchange rate regime, the central bank must intervene to maintain the fixed international price of the currency. In a dirty float, the central bank has the option to intervene and often does so in an attempt to correct imbalances in the current account. Current proposals to extend a target zone regime to the main currency blocks of the world, the dollar, the euro and the yen, are based on increasing the role of central banks in the foreign exchange market. It is, therefore, of importance to understand the consequences of central bank intervention. The style of the intervention may be described by a combination of the following characteristics — public versus secret interventions, sterilized versus non-sterilized interventions. Sometimes, central banks publicize their interventions trying to influence the market to follow a specific direction. Interventions are often sterilized to minimize potential conflicts with other objectives of monetary policy. Previous literature concerned with measuring the effects of interventions has given various results. Dominguez (1998) analyzes a long time series of daily data in the context of various GARCH specifications to conclude that interventions have a significant effect on volatility, but the sign changes over time. Sometimes, interventions stabilize and some other times destabilize the exchange rate. Chang and Taylor (1998) use high frequency data on exchange rates and interventions for their analysis and conclude that intervention has a very short effect on volatility. On the other hand, Le Baron (1999) uses daily data to conclude that technical rules are effective especially in periods of central bank interventions. We examine the issue in the context of high frequency data, offering some methodological modifications and an extended data set. In particular, we use a sample of exchange rates recorded at the 30-min frequency for the period 1992–1995. This data set is considerably longer than the one used by Chang and Taylor (1998) and may allow for a more significant analysis. We perform an analysis with both daily and high frequency data. The daily data are aggregated from the high frequency data to retrieve a non-parametric estimate of volatility. The advantage of daily data is that information on interventions is also available at the same frequency. As far as direct use of high frequency data is concerned, the empirical evidence cumulated so far suggests the importance of properly taking into account the periodic intra-day dynamics. To take into account intra-day seasonals, we use an unobserved component model developed by Beltratti and Morana (1999). The thorough comparison with other methodologies reveals that the unobserved component methodology has several advantages. First, it fares well in terms of producing a series, which satisfies the theoretical aggregation properties of GARCH models. Second, it forecast satisfactorily the future volatility within the day. Third, it is very flexible in describing episodes of volatility connected with announcements of macroeconomic variables. This last characteristic is very relevant here as the scope is evaluating the effects of interventions on volatility. Given the continuous changes taking place in financial markets, it is of great importance to use an econometric model, which allows for time-varying effects of interventions, as the one we use here. The plan of the paper is as follows. After this introduction, Section 2 reviews the literature on interventions. Section 3 presents the empirical analysis and its results, while Section 4 concludes. The main conclusions of our work are as follows. We confirm the existence of a positive effect of interventions on intra-day volatility, particularly on the seasonal component. Moreover, we also find that intervention increases permanent volatility in the short run, even though the effect seems to revert at longer horizons of 10–20 days. Also, the level of the exchange rate moves in the desired direction only approximately 50% of the times. The analysis, therefore, shows that interventions are often not very effective. The change in the level of the exchange rate is obtained sometimes with a substantial increase in permanent volatility.
نتیجه گیری انگلیسی
In this paper, we have used high frequency data for the analysis of the effects of interventions on the foreign exchange market. Our analysis is novel from two points of view. First, we use high frequency date in a structural time series framework that allows for both identification of transitory and permanent components and a more precise analysis of intervention effectiveness. Second, we have used high frequency data also to construct a daily proxy for each component of volatility given that the intervention variable is sampled at the daily frequency. We derive two main conclusions from the analysis. First, in our sample, interventions are not particularly effective. They achieve the desired change in the direction of the exchange rate only about 50% of the times. Of these, 50% imply a relevant increase in the permanent component of volatility. Second, the stationary component, which is affected the most by volatility, is the intra-day cyclical component. This confirms previous results obtained from the study of announcements in the foreign exchange market and cast doubts on the reasons and the effectiveness of central bank interventions in the foreign exchange market. If interventions imply doubtful effects on the trend of the exchange rate and an increase in volatility, then one might wonder why central banks intervene at all in the light of an objective function, which is regarded usually as including only inflation and a measure of economic activity. Interventions would make sense if the resulting effects on the exchange rates were strong enough to affect also the levels of direct interest to the central bank. There are a few possible directions for future research. First, it would be important to know more about the exact timing of the intervention to analyze the short run dynamics of volatility more precisely. Second, it would be interesting to extend the data set to include secret interventions by the central bank to compare the different effects of secret and public interventions. Third, it would be important to use data on temporally close interventions from other central banks, which can also affect the value of the currency. Fourth, it would be useful to evaluate the robustness of our estimate of permanent volatility in light of a competing model based on a fractionally integrated long run component. Fifth, it would be interesting to extend the data set to include interventions on more than one exchange rate. This would increase the number of observed public interventions and would allow for a more specific testing of hypotheses.