Volatility clustering: A nonlinear theoretical approach

This paper verifies the endogenous mechanism and economic intuition on volatility clustering using the coexistence of two locally stable attractors proposed by Gaunersdorferet al. (2008). By considering a simple asset pricing model with two types of boundedlyrational traders, fundamentalists and trend followers, and noise traders, we provide theoretical conditions on the coexistence of a locally stable steady state and a locally stableinvariant circle of the underlying nonlinear deterministic financial market model and shownumerically that the interaction of the coexistence of the deterministic dynamics and noiseprocesses can endogenously generate volatility clustering and long range dependence involatility observed in financial markets. Economically, volatility clustering occurs when neither the fundamental nor trend following traders dominate the market and when tradersswitch more often between the two strategies.