主讲人：
黎新平，北京大学经济学院特聘研究员
主持老师：
（北大经院）王熙
参与老师：
（北大经院）王一鸣、刘蕴霆、王法、李少然、巩爱博
（北大国发院）黄卓、张俊妮、常晋源
时间：
2026年10月16日（周五）
10：00-11:30
地点（线下）：
北京大学经济学院301会议室
主讲人简介：
黎新平，北京大学金融工程实验室执行主任，北京大学经济学院研究员，斯坦福大学经济学博士，主要研究方向为量化投资、资产定价、机器学习及人工智能技术在金融中的应用。曾在国际货币基金组织从事新兴市场贷款计划定价的研究，并在美国纽约大型对冲基金任资深策略师和投资经理，国内公募基金任量化投委会主席、量化投资总监及基金经理。
报告摘要：
We develop a unified framework that links analysts endogenous research choices, forecast accuracy, and market reactions. We posit that analysts choose optimal research effort after observing private signals to maximize their utility, while investors place greater weight on signals with higher precision and lower weight on information that is redundant with other analysts' signals. Research effort improves forecast precision, strengthening the price response to forecast news. We test these model implications using a large language model to measure research effort documented in Chinese equity reports from 2016 to 2025. We find that greater effort is associated with smaller forecast errors and stronger price responses to forecast surprises. A strategy that buys stocks with high precision-weighted forecast revisions and sells those with low revisions earns an annualized abnormal return of 17.27%, compared with 12.84% for the simple-average revisions.
