Cloud cover and expected oil returns

Abstract

Satellites can “sense” oil inventory, but cloud cover prevents observation, which reduces the flow of information into the oil market and creates uncertainty about information availability. The effects of the availability of such information on oil prices need to be thoroughly explored. Therefore, using time-series prediction, this paper examines the effects of the availability of satellite-based information on oil returns. The cloud cover above the floating roof tanks in major oil storage areas is measured to predict oil returns through regression approaches. The empirical results indicate that higher cloudiness in a week leads to lower oil returns in the following week. The predictive ability of cloudiness is significant from both in-sample and out-of-sample perspectives. The ability of cloudiness measures to predict oil returns can be explained by information uncertainty and information flow channels. The findings have important implications for asset pricing and risk management using big data.