Does Weather Matter?

Abstract

We use semi-parametric bin tests, regression analyses and copula modeling techniques to identify the relationship between temperature and stock market returns. After examining 25 international stock markets, we find that the negative correlation is statistically significant in individual countries, i.e. the higher is the temperature, the lower the stock returns. However, we fail to find joint significance of temperature effects across markets after correcting for market comovement by seemingly unrelated regression. We also find negative temperature effects on returns are robust to different measures of daily temperature. Both constant-dependence and time-varying-dependence conditional copula models are employed to analyze the general dependence between temperature and stock market returns. The copula results show that the negative relation remains after controlling for autocorrelations, GARCH effects and non-normality and the dependence between temperature and stock market returns is relatively stable over time.