德勤－复旦会计论坛系列215-216期讲座信息

德勤－复旦会计论坛系列215期讲座信息

时间：2023年12月14日（周四）下午1:30

地点：史603教室

主持人：黄蓉李达三讲席教授

报告人：‪Bin Ke, Provost's Chair Professor (National University of Singapore)

题目：Can Machine Learning Enhance Management Control Systems? The Case of Employee Selection

摘要： Leveraging proprietary job application data from a financial institution, we investigate the efficacy of machine learning (ML) in enhancing employee selection processes. We demonstrate that ML models can not only emulate human recruiters' decisions in matching applicants with employers but also substantially reduce inefficiencies in these matches. Critically, our most effective machine learning models excel in identifying job candidates likely to demonstrate a lack of commitment to the firm, as well as those poised to deliver exceptional job performance. Employing Shapley values, we illuminate the distinctions in decision-making between human recruiters and our ML models. In addition, our top-performing ML models exhibit reduced biases compared to human recruiters, against disadvantaged groups. These insights emphasize the transformative role of machine learning in optimizing employee selection and enhancing management control strategies.

德勤－复旦会计论坛系列216期讲座信息

时间：2023年12月18日（周一）下午2:00

地点：史301教室

主持人：张新教授

报告人：Yong Yu, C. Aubrey Smith Professor (University of Texas at Austin)

题目：The superiority of earnings over cash flows in predicting cash flows available to investors over the long run

摘要：Recent studies have reignited the debate over whether earnings or cash flows better predict future cash flows. These studies and prior work generally focus on predicting future operating cash flows over the short run (one to three years ahead). In contrast, we examine the ability of earnings versus cash flows to predict cash flows available to investors after investment outlays over the long run (up to 20 years), which have a greater impact on future dividends than near-term operating cash flows. We show that earnings dominate cash flows in predicting free cash flows over the long run, and this superiority is attributable primarily to earnings’ inclusion of long-term investment accruals. We also find the superiority of earnings over cash flows and the importance of long-term investment accruals are more pronounced among firms with more volatile investment spending and longer investment cycles. Overall, our results help explain why earnings, not cash flows, are the predominant metric used for valuation and contracting purposes.