Numerous psychological studies show that weather conditions affect people's mood and that mood states are correlated with people's subjective evaluation of future probabilities. In this paper, a new approach is developed and asset market data are employed to test the mood-subjective probability relation. Cloud cover and precipitation volume serve as two mood proxies. Our statistical analysis suggests that bad mood states are characterized by investors placing higher probabilities on adverse events.

Using asset market data, as well as theoretical relations between investors’ preferences, option-implied, risk-neutral, probability distribution functions (PDFs,) and index-implied, actual, PDFs, this paper extracts a time-series of investors’ relative risk aversion (RRA) functions. Based on results recently derived by Benninga and Mayshar (2000), these functions are used to recover the evolution of risk preferences heterogeneity. Applying non-parametric estimation on European call options written on the S & P500 index, we find that: (i) the RRA functions are decreasing; and (ii) the constructed risk preferences heterogeneity series is positively correlated in a static, as well as a dynamic, setup with a prevalent proxy for investors heterogeneity, namely, the spread between auction- and market-yields of Treasury bills.