Research

Publications: 


Working Papers: 


Abstract: In this paper, we apply machine learning techniques to predict corporate bond returns all over the world. With a novel dataset, we find there is strong predictability of corporate bond returns in international markets. However, the documented factors that drive bonds in the U.S. market are substantially different from factors that impact bonds in non-U.S. markets, where downside risk and illiquidity are more influential. We further find there are cross-country differences in the degree of bond cross-country integration and bond-stock integration, but these two integrations in developed markets are on average higher than in emerging markets.


Abstract: This paper examines the effect of abnormal-temperature-related climate risk on bank loan pricing. Using a sample of syndicated loans from 35 countries or jurisdictions, we find that banks charge higher interest rates for borrowers with higher climate risk, especially for short-term loans. Our cross-sectional analyses reveal that borrowers’ voluntary climate risk disclosures in conference calls mitigate the effect of climate risk on loan spreads, and the moderating effect of climate risk disclosures is stronger when lead banks have less climate-risk-related lending experience and when borrowers have higher analyst coverage. In addition, the borrowing cost of high-climate-risk borrowers decreases in the U.S. after the SEC issued climate risk disclosure guidance, while the ESG disclosure requirements in other 19 countries, which are not climate-risk specific, do not alter the effect of climate risk on bank loan pricing.


Abstract: We examine whether climate risk disclosures in earnings conference calls affect analyst forecast accuracy. Using a sample of U.S. firms from 2002 to 2019, we find that the disclosure of physical shocks related to climate change is associated with smaller analyst forecast errors, while the disclosures of regulatory and opportunity shocks related to climate change have no significant effect on analyst forecast errors or dispersion. Cross-sectional analyses show that the effect of physical climate risk disclosures on analyst forecast accuracy is stronger after the SEC provided guidance on disclosures of climate risk and opportunities, for firms located in regions that are more likely affected by extreme weather events and have higher disaster costs, and for firms located in states where people have stronger beliefs in climate change. Our findings suggest that only certain type of climate risk disclosures in earnings conference calls help analysts make more accurate earnings forecasts. 

Work in Progress