Year: 2019 January, Volume 28 No. 2
This study proposed a state-space model that allows time-varying weather effects on asset returns. It resolves the model misspecification of the unrealistic, fixed effect assumption commonly made by previous weather studies. The model was applied to examine the weather effects on Thai government bond returns from July 2, 2001, to December 30, 2015. Kalman filtering was used in the estimation. The study found that the weather effects were time-varying. They were wandering in the early sample period but disappearing in the later period. The effects were not co-integrated
with the market’s inefficiency levels.