Recursive Modeling of Nonlinear Dynamics in UK Stock Returns

This paper presents results from recursive modeling of nonlinear dynamics in UK stock returns. A specification search suggests a two-state model and we demonstrate the ability of this model to capture time-varying volatility, skew and kurtosis in UK stock returns. An out-of-sample forecasting experiment confirms the strong statistical evidence of nonlinearity and shows that accounting for regimes leads to improved forecasting performance. Copyright Blackwell Publishing Ltd and The Victoria University of Manchester, 2003.