光华会计新才智论坛第四期

时 间：2024年9月26日（周四）9:30-11:30

地 点：柳林校区诚正楼645、腾讯会议767-519-140

主持人：陈磊 教授（现代会计研究所所长）

时间：9:30-10:10

报告1：Currency Carry Trades: A New Perspective from Risk-Neutral Moments

报告人：刘毅（罗格斯大学）

摘要：The forward-premium puzzle exists in the exchange rate market, where the forward exchange rate is a biased forecaster of the future spot exchange rate. In this study, we aim to explain the forward premium puzzle from the risk-neutral moments perspective. First, we confirm the existence of the forward-premium puzzle by conducting the back-test of the carry trade strategy and running the cross-sectional uncovered interest parity test. Second, we demonstrate that the forward-premium puzzle can be explained by the option implied risk-neutral moments (variance and kurtosis). These findings support the perspective that forward premiums are due to risk compensation required by investors for bearing volatility and jump risks of FX investments. Finally, we constructed highly profitable carry trade strategies through the double adjustments for risk-neutral variance and kurtosis. This paper contributes to a better understanding of high currency carry trade returns and provides new insights for future research.

时间：10:10-10:50

报告2：Can Machine Learning Improve Abnormal Accrual Estimates?

报告人：谷雨（罗格斯大学）

摘要：Traditional accrual models often misclassify normal and abnormal accruals due to Type I and Type II errors. This study aims to enhance the construct validity of abnormal accrual estimates by applying regression tree-based Machine Learning (ML) algorithms, which are adept at uncovering the intricate relationships between total accruals and their determinants. By analyzing a dataset of U.S. public firms from 1988 to 2017, we compare abnormal accrual estimates derived from traditional linear regression models against those from ML models. We find that abnormal accrual estimates obtained from ML models demonstrate lower Type I and Type II errors. This study contributes to accounting literature by integrating advanced analytics into the classic accounting paradigm of accrual estimation, offering a more precise and replicable method for future research. Enhanced abnormal accrual estimates afford researchers, practitioners, and regulators a more reliable means to assess financial reporting and audit quality.

时间：10:50-11:30

报告3：The Persistent Response from Option Liquidity to GameStop Short Squeeze

报告人：杨睿歆（罗格斯大学）

摘要：We examine the impact of the short sale constraint on option liquidity when short sale cost rises during the GameStop short squeeze. We find that option market makers lower their inventory risk by cutting liquidity provision. Besides, the short sale cost influences the option liquidity over time after the GameStop short squeeze. We also discuss the impact of retail trading in the relation between short sale cost and option liquidity. Our findings suggest that the impact of the short sale constraints on option liquidity is significant over time.