A review of the seasonal affective disorder hypothesis

This review of the SAD hypothesis of Kamstra et al. (2003), hereafter KKL (2003), isolates four new problems. First, the KKL (2003) statistical model does not test the KKL (2003) SAD hypothesis. Second, KKL (2003) do not properly interpret their results. Third, KKL (2003) incorrectly specify the sign of the SAD effect in the winter. The revised SAD hypothesis is that hours of night have a negative influence on stock returns in the fall and in the winter. Fourth, the statistical tests do not support either the KKL (2003) hypothesis or the revised SAD hypothesis.

► Isolates four new problems with the seasonal affective disorder (SAD) hypothesis. ► Offers a new corrected formulation of the hypothesis. ► Compares alternative measures of SAD. ► Controls for country effects by panel data regressions using 5 large economies. ► Finds no empirical support for the old or the revised SAD hypothesis.