I introduce two-factor discrete time stochasticvolatility models of the short-term interest rate tocompare the relative performance of existing andalternative empiricial specificattions. I develop anonlinear asymmmetric framework that allows forcomparisons of non-nested models featuringconditional heteoskedasticity and sensitivity of thevolatility process to interest rate levels. A newclass of stochastic volatility models withasymmetric GARCH models. The existing models arerejected in favor of the newly proposed modelsbecause of the asymmetric drift of the short rate,and the presence of nonlinearity, asymmetry, GARCH,and level effects in its volatility. I test thepredictive power of nested and non-nested models incapturing the stochastic behavior of the risk-freerate. Empirical evidence on three-, six-, and12-month U.S. Treasury bills indicates but thattwo-factor stochastic volatility models are betterthan diffusion and GARCH models in forecasting thefuture level and volatility of interest ratechanges.