Nonlinear Dynamics of the Russian Stock Market in Problems of Risk Management

This paper studies the dynamics of the Russian stock market in 2000–2007 from the stochastic and chaotic viewpoints. Estimation of Lyapunov exponents for a number of Russian stock prices and indices suggests the absense of low-dimensional chaos. A more precise description of the market dynamics is offered by the stochastic approach, within which the best model was found to be GARCH(1,1) ~ t . Christoffersen and Berkowitz tests show that this model is better at estimating value-at-risk of trading positions than a benchmark model with independent Gaussian returns, and that systematic errors in risk assessment are quite small.