TY - JOUR
T1 - Equilibrium analysis of volatility clustering
AU - Vanden, Joel M.
N1 - Funding Information:
I thank Franz Palm, two anonymous referees and my colleagues at the Tuck School for helpful feedback. This research was partially funded by a NASDAQ Foundation Fellowship. Research support from the Tuck School is gratefully acknowledged. This paper was previously titled “Volatility Clustering, Time-Varying Risk Aversion and the Pricing of Options”. Any errors are mine.
PY - 2005/6
Y1 - 2005/6
N2 - Volatility clustering is a pervasive feature of equity markets. This article studies volatility clustering in an equilibrium setting by generalizing the CRRA and CARA representative agent models of finance. In equilibrium, the market portfolio follows a volatility regime-switching process in which the volatility level is determined by the agent's local risk aversion. Using monthly data, the empirical tests reveal that at least four volatility regimes are necessary to fit the data. While one of the models explains the GARCH effects in the data, an analysis of the Euler equation pricing errors suggests that both models are likely misspecified. Since the models can be used to closely approximate any state-independent utility function, it is doubtful that there exists any representative agent equilibrium (with state-independent utility) that is consistent with the data. An equivalent interpretation is that the market portfolio price process is not a diffusion process of the type studied by Bick [Bick, A., On viable diffusion price processes of the market portfolio, J. Finance 45 (1990) 673-689] and He and Leland [He, H., Leland, H., On equilibrium asset price processes, Rev. Financ. Stud. 6 (1993) 593-617].
AB - Volatility clustering is a pervasive feature of equity markets. This article studies volatility clustering in an equilibrium setting by generalizing the CRRA and CARA representative agent models of finance. In equilibrium, the market portfolio follows a volatility regime-switching process in which the volatility level is determined by the agent's local risk aversion. Using monthly data, the empirical tests reveal that at least four volatility regimes are necessary to fit the data. While one of the models explains the GARCH effects in the data, an analysis of the Euler equation pricing errors suggests that both models are likely misspecified. Since the models can be used to closely approximate any state-independent utility function, it is doubtful that there exists any representative agent equilibrium (with state-independent utility) that is consistent with the data. An equivalent interpretation is that the market portfolio price process is not a diffusion process of the type studied by Bick [Bick, A., On viable diffusion price processes of the market portfolio, J. Finance 45 (1990) 673-689] and He and Leland [He, H., Leland, H., On equilibrium asset price processes, Rev. Financ. Stud. 6 (1993) 593-617].
UR - http://www.scopus.com/inward/record.url?scp=19644362303&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=19644362303&partnerID=8YFLogxK
U2 - 10.1016/j.jempfin.2004.04.007
DO - 10.1016/j.jempfin.2004.04.007
M3 - Article
AN - SCOPUS:19644362303
VL - 12
SP - 374
EP - 417
JO - Journal of Empirical Finance
JF - Journal of Empirical Finance
SN - 0927-5398
IS - 3
ER -