Testing for one-factor models versus stochastic volatility models

摘要: This paper proposes a testing procedure in order to distinguish between the case where volatility of an asset price is deterministic function itself and one it or more (possibly unobservable) factors, driven by not perfectly correlated Brownian motions. Broadly speaking, objective generic one-factor model stochastic model. In fact, no specific assumption on functional form drift variance terms required. The proposed tests are based difference two different nonparametric estimators integrated process. Building some recent work Bandi Phillips (2003) Barndorff-Nielsen Shephard (2004a), shown that test statistics converge mixed normal distribution under null hypothesis factor diffusion process, while diverge multifactor models. The findings from Monte Carlo experiment indicate suggested has good finite sample properties.