Threshold Cointegration between

 stock index and exchange rate

 

by Thian  Kanokpongsak  (slow – sex)

 

                    เป็นอีกหนึ่งเรื่องที่มีความน่าสนใจในตัวของมันเอง โดยเป็นหัวข้องานวิจัยของ คุณ เธียร กนกพงษ์ศักดิ์ (นักศึกษาปริญญาโท มธ. หลักสูตร MIF) ซึ่งกำลังหมกมุ่นกับมันอยู่  เนื้อหาและความน่าสนใจของหัวข้อนี้จะเป็นอย่างไรลองอ่านดูนะครับ อ่านแล้วขอความร่วมมือช่วยแสดงความคิดเห็น (ถ้าทำได้นะครับ) เพื่อนำไปปรับปรุงให้ดีขึ้นต่อไป

Introduction

 

                  The interdependent of the movements in the stock prices in some countries and others is obvious. As consequences, it is like a required activity for all investors around the globe to follow the news and impacts from others countries all along.

                    Consistent with the norm of every market, it is like a routine for Thai investors to always look at the movements in the stock exchange of Japan via Nikkei index of Tokyo Stock Exchange, and use it as the leading indicator to predict how the stock market will move; the relationship of these two markets is interesting. This might be enough reason for two stock markets comovements to be presence but, however, one could argue that the comovements in the stock market might occur because of the relationship in international trade among the countries. The degree of the exchange rate movements, via the import/export, might influence the performance of the stock market in others countries. To answer this question, this paper tries to study the relationship between the stock markets and the exchange rate among these countries.

                       The main reason why this paper studies the relationship between Thailand and Japan is that Japan holds the largest portion in Thai Foreign Direct Investment, FDI, for a long period of time. Because of that highly related real sector, Japan and Thailand should depict some relationship of the exchange rate and the stock prices.

                       Most of the previous papers studied the relationship between stock indexes and exchange rates movements by assuming the linear relationship. There’re more reliable to use the non-linear model to study the time series behavior of the exchange rates as suggested by David (2008). As the standard unit-root test assumes the linearity and symmetric adjustment of the variables, it is not seem to be realistic in explaining the real world situation. This study, by using the threshold cointegration technique, not only gives the answer of the evidence for the long-run relationship but also will focus on the insight about asymmetric adjustment that might occur in the causal effect of the variables. Then this could open the room for us to find the relationship of the exchange rate and the stock price in greater details.

                        This study tries to figure out the long-run non-linear relationship between stock indices THB/JPY. By employing the threshold Cointegration concept, we could study whether the stock prices and exchange rate have the relationship that will trigger on and off when the degree of divergent reaches certain threshold levels and disappears if the magnitude of difference is trivial. This threshold cointegration concept was initiated by Balke and Fomby (1997).

                        There are a considerable number of previous studies about the relationship of the stock prices and the exchange rate. In addition, there’re also a number of different models and methodologies employed as a tool to address the relationship. The contribution of this paper is that I will use the models that could depict the asymmetry and also not assume linear relationship, and then the result will be more realistic and reliable then assuming symmetry and linear of the model.

                        The studies that found the evidence that the relationship between stock prices and exchange rates exist are, for example, that of Yau and Nieh (2009), Anita and Murli (1995), Golaka and Samanta (2003), Homma et al. (2005), and Doukas et al. (1999).However there’re also some evidence against the relationship between the variables such as the study of Nieh and Lee (2001).

Theoretical Framework

 

                  There’re two basic underlying concepts about the relationship of stock prices and exchange rate, i.e. traditional approach and portfolio approach.

                   The traditional approach states that the depreciation in domestic currency against the foreign currency makes that country goods more competitive and then have positive effect on the stock market. If this concept holds true, the effect of the currency on the stock price should be found as the exchange rate influence the stock price and the effect is positive, the depreciation in domestic currency should causes the stock price in that country to increase. This concept suggested that the relationship between stock prices and exchange rates should be explained by the real economy.

                     On the other hand, the portfolio concept believes that the good stock markets make the foreigners demand more domestic assets, including stock, and then the stock market should influence the exchange rate as the foreign investors demand more domestic asset, the domestic currency should appreciates. From this point of view, we expect the stock market to causes exchange rate to move and when the stock prices increase, we predict that the currency should appreciate.

                       In this paper we will study not only the comovement in the stock indices but also the comovement between both THB/JPY and movement in each country stock market, Stock Exchange of Thailand, Tokyo Stock Exchange. Then we could consider which is the mechanism that explains the relationship between Thailand and Japan.

Data and Methodology

 

                   In this paper, monthly return of the stock index will be extracted from Datastream database, which is the SET index from Thai stock exchange and Nikkei index from Japan Tokyo Stock Exchange. Period covered in the study will be from Jan 1990, to Dec 2008 because the 1997 crisis and also 2002 internet bubble should be included into the period of study in order to study the comovement occurred in each events. The main reason why monthly data should be employed is that it will average out the possibility of abnormal short term fluctuation.

                    In this paper, I will study the relationship between stock indices and exchange rates assuming asymmetrical non-linear relationship between the two. Then, to test for the unit root, the basic ADF test and the exponential smooth transition autoregressive (ESTAR) will be employed. The ESTAR models, as represented in Kapetanios (2003), allow us to reasonably test for the nonlinearity of unit root of the variable. I will use this technique to check for the stationarity of the THB/JPY, SET index, and NIKKEI. The ESTAR process can be shown as follow:

 (ขออนุญาติใช้รูปนะครับ เพราะผมขี้เกียจพิมพ์สมการ)

 

                         

Enjoy it