Empirical Evidence for State and Time Separable Preferences: Case Finland

Abstract

The study investigates the implications of modified utility specifications developed on the standard power utility assumptions for Finnish representative agent while breaking the state and time separable constraints. The estimations are carried out using returns on equity and bond returns with iterated GMM procedure. The results from Epstein and Zin (1991) and Campbell and Cochrane (1999) models show Finnish risk premia is time varying across the studied samples. We conclude Campbell-Cochrane model outperforms the competing models in producing plausible model parameters while suppressing specification errors. The diagnostic checks show model is able to capture variations in stock returns over time. It also commands a significant price of risk in cross-sectional regressions and even manages to do better than unconditional CAPM.