An Empirical Analysis of Short Term Interest Rate Models for Turkey

Interest rate is one of the most observed and forecasted variables in financial markets. Interest rates and the volatility of interest rates play a crucial role in pricing financial instruments. In this empirical study, we try to investigate which short term interest rate model is appropriate for Turkish data. In that regard we use monthly average of the central bank overnight interest rate. The date set covers the period from 1990:01 to 2008:07. We use the generalized method of moments to estimate the model parameters since it does not require a distributional assumption for the interest rate making GMM a robust estimation method comparing to maximum likelihood. Estimation results reveal that Cox Ingersoll Ross square root process and Brennan-Schwartz models perform better. A common feature of these models is that they both have heteroscedastic variances. In the study, we also analyze if the policy changes of Central Bank of Turkey had any effects on the interest rate process. We find that the volatility of the interest rate is not affected by policy change. However, the level of the interest rate is affected