Türkiye’de Döviz Kuru Oynaklığının Modellenmesi: Ampirik Bir Araştırma
Sinem Kutlu Horvath, İpek M. YurttagülerDöviz kurunun denge değeri etrafındaki dalgalanmalara karşılık gelen bir kavram olarak döviz kuru oynaklığı kur riskinin başlıca kaynağıdır ve uluslararası ticaret, yatırımlar ve sermaye akımları başta olmak üzere makroekonomik istikrarı bozacak pek çok değişkeni olumsuz etkilemektedir. Bu bağlamda, döviz kuru oynaklığının ampirik olarak tahmini ve ölçümü yaygın ekonomik etkileri açısından üzerinde durulması gereken bir konudur. Döviz kurlarındaki oynaklığın temel makroekonomik değişkenler üzerinde yarattığı etkiler geniş bir araştırma alanı oluşturmuş, böylece teorik ve ampirik açıdan oldukça zengin bir literatüre de zemin hazırlamıştır. Çalışmamızda, 2003-2022 dönemine ait efektif döviz kuruverileri kullanılarak Türkiye için oynaklık ARCH-GARCH modelleme teknikleriyle tahmin edilmektedir. Çalışmadan elde edilen bulgulara göre, Türkiye için döviz kuru oynaklığının tahmininde GARCH(1,1)'in en uygun model olduğu sonucuna varılmıştır.
Modeling Exchange Rate Volatility in Türkiye: An Empirical Research
Sinem Kutlu Horvath, İpek M. YurttagülerExchange rate volatility is a concept that corresponds to the fluctuations around the equilibrium value of the exchange rate and is the main source of exchange rate risk as it adversely affects many variables that can disrupt macroeconomic stability, especially international trade, investments, and capital flows. In this context, empirical estimation and measurement of exchange rate volatility is an issue that needs to be emphasized in terms of its widespread economic effects. The effects of exchange rate volatility on basic macroeconomic variables have created a wide range of research, thus laying the groundwork for a very rich theoretical and empirical literature. This study estimates the volatility in Türkiye using theAutoregressive conditional heteroskedasticity (ARCH) and generalized ARCH (GARCH) modeling techniques alongside effective exchange rate data for the period of 2003-2022. According to the obtained findings, the study has concluded the GARCH(1,1) model to be the most appropriate model for estimating exchange rate volatility in Türkiye.
Exchange rate volatility is a term that refers to wide fluctuations around the long-term equilibrium value of the exchange rate. The transition to floating exchange rate systems has confronted both developed and developing countries with the uncertainty created by exchange rate fluctuations and the resulting risk problem. Volatility in exchange rates is the main source of exchange rate risk, and is of great importance in this context in terms of its macroeconomic implications. Exchange rate volatility is known to negatively affect many macroeconomic variables such as investment, production, consumption, and economic growth, especially in international trade and capital movements. Therefore, empirical estimation and measurement of exchange rate volatility is important for its pervasive economic implications. The effects sudden and unexpected fluctuations in exchange rates have on basic macroeconomic variables have created a wide field of research, thus laying the groundwork for a very rich theoretical and empirical literature.
This study firstly discusses the theoretical framework of the concept of exchange rate volatility and emphasizes the factors that causing exchange rate volatility, as well as the negative effects of volatility on macroeconomic variables. In order to shed light, the study then provides examples from the literature examining exchange rate volatility in Türkiye.
The study’s econometric analysis section estimates volatility in the Turkish economy using effective exchange rate data for the period of 2003-2022. The volatility of many financial time series, including exchange rates, is not constant over time, and recent studies have revealed variance as a measure of volatility to not be constant. As such traditional time series models that accept variance as constant are understood to be insufficient for modeling volatility. In this framework, the study will develop modeling techniques such as autoregressive conditional heteroscedasticity (ARCH) and generalized ARCH (GARCH) models that are based on the assumption of changing variance and that are suitable for the dynamic structure of financial markets. The ARCH and GARCH models allow volatility to change over time and have been widely used for modeling exchange rate volatility. To estimate exchange rate volatility, the study will first ensure the stationarity of the series, then it will apply the ARCH Lagrange multiplier (ARCH-LM) model to determine whether the variance in the error terms is constant. The study then examines the stationarity of the effective exchange rate series together with the autocorrelation function and the cartesian graph; it then performs the unit root test, with any non-stationary series at that level being made stationary by taking the difference. The study determined the most suitable autoregressive integrated moving average (ARIMA) model as a result of the partial and autocorrelation functions of the series. The ARCH effect was investigated regarding the error squares of the determined ARIMA model to determine the volatility of the exchange rate series. Finally, as a result of the analysis made for modeling the exchange rate volatilityin Türkiye, the most appropriate model was determined to be the GARCH(1,1) model. In order to determine the reliability of the model, the study re-performed the ARCH-LM, after which the volatility in the model was seen to have disappeared. Accordingly, the
GARCH(1,1) model has been concluded to be a model that eliminates the effects of exchange rate volatility.
