ASYMMETRIC IMPACT OF EXCHANGE RATE TO ISTANBUL STOCK MARKET: A NONLINEAR ARDL APPROACH

The stock market plays an important role for financial organizations and portfolio managers in modern economics. Volatility in stock market return is one of the essential tools between lenders and borrowers that help them to assess the risk of portfolios and predicting the return of future investment’s income. This paper investigates the relationship between the stock market and the exchange rates of Turkey. BIST100 index is employed as a proxy for stock returns and TL/US$ and TL/EURO (currency rate of Turkish Lira against US Dollar and EURO) are considered for exchange rate exposure. The series is on monthly data over the period 2008m10- 2021m03. We use the NARDL and ARDL mythology of the time series model to determine the short-run and long-run impacts of the exchange rates on the İstanbul Stock Market. Our findings indicate that there is a relationship between both exchange rates (US Dollar and EURO) and stock returns. This relationship is positive, asymmetric, and statistically significant for both the long and short run.


INTRODUCTION
Predictability of volatility of stock returns helps portfolio managers to minimized risks of investment. One of the important factors which help financial organizations and portfolio managers to the predictability of the stock market is changes in the exchange rate. Although there is not especial consensus on the relationship of fluctuation in the exchange rate and stock market there is wide research about the effect of exchange rate and interest rate on stock markets. Those researches are mainly based on "good market approaches", "portfolio balance approaches" and "Asset Market approach". The first one was developed by (Dornbusch & Fischer, 1980) that emphasize on moving in exchange rates effect on the international competitiveness of firm by the effect on income and cost of borrow in foreign currencies. According to "portfolio balance approaches" which were developed by Frankel, J. A. (1993) focused on the role and effect of capital account transactions on determining the relationship between the exchange rate and stock market returns. The asset market approach is a method that emphasizes the value of an asset based on the selling price of similar assets.

LİTERATURE SURVEY
Theoretically, there are three models to explain the relationship between exchange rate and stock returns which are basically different and conflicting models (Moeljadi & Fauziah, 2015).
I. "Flow-Oriented" that presented by Dornbusch and Fischer (1980): According to (Dornbusch & Fischer, 1980 through international operation (by effecting on net export, consolidated report. Asset assessment/ liability assessment…) and domestic operation (by the effect on the price of input and output…) that both of them have an effect on cash flow, profitability, and decision making within the company which ultimately effects to stock supply and demand and stock price (Ball & Brown, 1968), (Evelyn, 2010). Therefore, there is a positive relationship between the stock market and the exchange market. This relationship and causality run from exchange rate to stock market.

II.
The theory of "Stock-Oriented" presented by (Frankel, 1983 and1984): Frankel believes that changes in stock prices cause to change in the exchange rate. Stock price changes affect an investor's reaction in the stock market (capital flow) that which will cause changes in supply and demand of local currency and exchanger rate. In other words, if the domestic stock price decrease (rises), will persuade investors to (sell) buy more domestic assets by (buying) selling foreign assets in order to obtain local currency. Increased demand for domestic currency will lead to an appreciation of the domestic currency (Moeljadi & Fauziah, 2015). The exchange rate of the local currency will appreciate foreign currency and there is a negative relationship which it runs from stock market to exchanger rate.

III.
Asset Market approach: according to this model due to different variable which has an effecton both exchange rate and the stock market, there is no interaction or very weak association between the exchange rate and stock market (Suriani, Jamil and Muneer, 2015).
When we look at the literature, there are studies in three fields. For example (Abdalla & Murinde, 1997) found a long-run relationship between the two variables in four Asian countries for the period 1985-1994 but there is no causality exists in Pakistan and Korea. While they found causality in India and the Philippines.
The study of (Bahmani-Oskooee & Sohrabian, 1992) is one of the first studies which employs causality methodology to examine the relationship between stock prices and exchange rates in the US. The study provided evidence that there is causality between the S&P 500 index and the effective exchange rate of the Dollar, at least in the short run. While (Ajayi, Friedman, and Mehdian, 1998) employ daily stock indexes and exchange rates for a set of emerging economies to investigate causal relations between stock returns and exchange rates. The findings provide Granger causality between the stock and exchange rates in all the advanced economies such as Canada, Germany, France, Italy, Japan, US, UK. While no consistent causal relations are observed in the emerging economies such as Taiwan, Korea, Philippines, Malaysia, Singapore, Hong Kong, Indonesia, Thailand. (Pan, Fok and Liu, 2007) found a high correlation between exchange rates and stock markets in seven Asian countries. The same results were get by (Kiymaz, 2003). Kiymaz, (2003) considered the efficiency of Turkish firms against the exchange rate. According to this study Turkish firms especially textile, machinery, chemical, and financial industries are highly exposed to foreign exchange risks. (Ojaghlou, 2020) focused on fundamental economic factors to the İstanbul stock market. According to this study, variables are cointegrated and there is a positive and statistically significant asymmetric longrun relationship from the nominal exchange rate and other factors to Istanbul stock market return. (Ojaghlou, 2020) suggests that the Istanbul stock market return (BIST-100) has consistent with the Efficient Market Hypothesis (EMH).  (Ajayi, Friedman, and Mehdian, 1998), (Suriani, Jamil and Muneer, 2015) and (Moeljadi & Fauziah, 2015) we set two models as follows:

DATA and METHODOLOGY
We carried out hypostasis by both USD (eq1) and Euro (eq2) 100 One of the is ARDL Bound Model developed by (Pesaran, Shin and Smith, 2001) which in our case for the effect of USD to BIST100 (eq1) can be written as follows: The second and third is NARDL and multiplier Model developed by (Shin, Yu and Greenwood-Nimmo, M, 2013) which in our case can be written as follows: Considering the asymmetric long-run relationship for the Nonlinear ARDL model: where xt : k × 1 vector of repressors decomposed as are partial sum processes of positive and negative changes in xt defined by and β pos , β neg are the associated asymmetric long-run parameters. The model can be written in an errorcorrection form as follows: The null hypothesis: ρ = θ pos = θ neg = 0 The steady-state of the model is: For the effect of EURO to BIST100 (eq4) is:
In the case of the effect of EURO to BIST100, in eq3, F-Bound (4.96) is greater than the upper bound of 99% critical value (4.94) and ECT is -0.11 negative and between -1 and 0. The coefficient of EURO is positive and statistically significant. Also in eq4 (NARDL model) coefficient error term (ECT) is (-0.124) which is negative and between -1 and 0) and it is statistically significant. The coefficient of Euro + is positive and statistically significant. Also, the coefficient of Eurois positive but it is not statistically significant.
To sum up, all coefficients of EURO and USD are positive and statistically significant (except Euro -). Therefore, there is a positive long-run relationship from exchange rates to BIST100.
There is a short run effect (90%) According to (Pesaran et al., 2001) the stability of ECT of the estimated models should also be empirically stable. Graphical representations of CUSUM are shown in figure1 and 2 for both models. According to figure1 the null hypothesis of specified ARDL and NARDL models cannot be rejected if the plot of these statistics remains within the critical bounds of the 5% significance level. The plots of all the CUSUM tests are within the red lines. Therefore, the stability of the estimated model and estimated coefficients are confirmed.

CONCLUSION
The main objective of this study is to analyses whether the exchange rate has a long and short-run effect on the Istanbul stock market returns or not. The series is on monthly data over the period 2008m10-2021m03. We use the NARDL and ARDL mythology of the time series model to determine the shortrun and long-run impacts of the exchange rates on the stock market.
The findings of the NARDL and ARDL Bound tests indicate that there is a relationship between both exchange rates (US Dollar and EURO) and stock returns. This relationship is positive positively asymmetric (NARDL), also liner (ARDL), and statistically significant for both the long and short run.