| Summary: | 碩士 === 輔仁大學 === 金融研究所 === 91 === Exchange rate is an important indicator of a country. Appreciation or depreciation of exchange rate will affect prices of trades and result in fluctuation of corporate profits. Investors who evaluate a corporate performance based on corporate profits. According to Efficient Market Hypothesis, if the security market is efficient, evaluation of a corporate will reflect on stock returns. The purpose of this paper is to explore the exchange rate exposure on Taiwan’s Electronic, Textile, and Financial industries.
Most of the researches use long-term sample periods. Meanwhile, many researchers further split off sample periods to study whether the exchange rate exposure will change between different periods. But they often arbitrarily split the research period into two or three sections. Based on above shortcomings, this paper employs panel data threshold regression model (Hansen, 1999) to split off research period objectively; that is to allow that regression functions may not identical across all observations and they may be divided into different classes based on the value of an observed variable. Besides, we also discuss current and lagged exchange rates which one is more proper to confer exchange rate risk. The empirical consequences are as follows:
1、Exchange rate risk does exist in Taiwan’s three major industries, and the coefficients are significantly minus signs. That means when N.T. dollar appreciates, the stock returns will rise up. Current exchange rates model is more proper than lagged exchange rates model. The exchange rate exposure between the electronic and textile industries is indifferent. It means that these two industries could be pooled together to discuss the whole exchange rate risk they face.
2、The test results show that three industries does exist single threshold value. No matter smaller or larger threshold value, three industries exchange rate risk coefficient present significantly minus signs. The electronic-textile industry shows if companies whose export ratios are smaller than threshold value will face larger exchange rate risk. The empirical result in financial industry also presents that, more exchange deposits the banks have, more exchange rate risk the banks face.
3、To compare panel data regression and panel data threshold regression model, we find that the panel data threshold regression model is more proper to explain exchange rate exposure in combined electronic-textile industry, however, to financial industry, the panel data regression is a better model to explain exchange rate risk.
|
|---|
