Combining Observational and Experimental Data to Improve Efficiency Using Imperfect Instruments
研究如何用不完美工具变量（满足相关性但可能违反排他性约束的变量）结合实验与观测数据，提升估计效率。
The paper studies how to combine experimental and observational data using imperfect instruments—variables that satisfy the first-stage relevance condition, but violate the exclusion restrictions.