Carbon risk and return prediction: Evidence from the multi-CNN method

This paper investigates the carbon risk and its role in stocks’ return prediction by identifying information implied feature engineering. We predict stock returns with different neural networks, construct investment portfolio according to predicted reflect of stocks risks through relevant evaluation portfolio. Our Multi-CNN method can best collect on relationship types make full use graph structure data identify risks. With or without factor, market performance high-carbon industry is better than that medium-carbon industry, low-carbon worst. Moreover, our finding consistent both Chinese American markets. Investment should pay attention requires corresponding premium.