Forecasting and Finite Sample Performance of Short Rate Models: International Evidence n SIRIMON TREEPONGKARUNA School of Finance and Applied Statistics, Australian National University, Australia ABSTRACT This paper evaluates the forecasting and finite sample performance of short- term interest rate models in a number of countries. Specifically, we run a series of in-sample and out-of-sample tests for both the conditional mean and volatility of one-factor short rate models, and compare the results to the random walk model. Overall, we find that the out-of-sample forecasting performance of one-factor short rate models is poor, stemming from the inability of the models to accommodate jumps and discontinuities in the time series data. In addition, we perform a series of Monte Carlo analyses similar to Chapman and Pearson to document the finite sample performance of the short rate models when b 3 is not restricted to be equal to one. Our results indicate the potential dangers of over-parameterization and highlight the limitations of short-term interest rate models. I. INTRODUCTION The short rate is one of the key economic indicators that has attracted considerable interest from researchers over the past few decades. The short rate has several influential roles in the economy. For example, central bankers or policy makers use the short rate as a target instrument for implementing monetary policy to influence the macro-economy. Practitioners also use the short rate as a key element in their valuation process of all securities and derivatives. Hence, numerous attempts have been made to model the short- term interest rate. To name just a few, these include work done by Vasicek (1977), Cox et al. (1985), Dothan (1978), Brennan and Schwartz (1980), Chan et al. (1992), Gray (1996), and Ait-sahalia (1996). In many of these short rate models, the dynamics of the short rate is determined by the drift and diffusion functions. Unfortunately, many popular short rate models produce weak results when fit to different data sets. Difficulties in modeling the short rate stem from the nature of near unit root dynamics of the short rate, causing the estimated n We thank Yacine Aı ¨t-sahalia for providing his original dataset. Special thanks also go to Stephen Gray, Bruce Grundy, Michael Martin, David Pitt, and Tom Smith for providing numerous invaluable comments. r 2006 The Author. Journal compilation r International Review of Finance Ltd. 2006. Published by Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. International Review of Finance, 5:3–4, 2005: pp. 175–197