Abstract: We consider the problem of causal inference for panel data with general treatment patterns, a paradigm with broad applications in areas ranging from program evaluation to e-commerce. We propose a novel treatment effect estimator for this problem that we show to be rate-optimal and asymptotically normal under general conditions on the treatment pattern. In particular, our recovery guarantees are valid under a set of conditions on the treatment matrix which relate to its projection on the tangent space of the counterfactual matrix. Should our conditions be violated by an amount that grows negligibly small with problem size, no estimator can recover the treatment effect. Our work thus generalizes the synthetic control paradigm to allow for general treatment patterns. Our recovery guarantees are the first of their type in this general setting. Computational experiments with our estimator on synthetic and real-world data show a substantial advantage over competing matrix completion based estimators. Joint work with Andrew Li (CMU) and Tianyi Peng (MIT).

Bio: Vivek Farias is the Patrick J. McGovern (1959) Professor and a Professor of Operations Management at the MIT Sloan School of Management. His research focuses on the development of new methodologies for large-scale dynamic optimization under uncertainty, and the application of these methodologies to the design of practical revenue management strategies across various industries ranging from airlines and retail to online advertising. Farias is a recipient of the 2006 INFORMS MSOM Student Paper prize for a research paper judged to be the best in the field of operations management. A consultant in the finance industry, he most recently contributed to GMO LLC’s first successful high-frequency algorithmic trading strategy. Farias holds a BS in computer engineering from the University of Arizona, and a PhD in electrical engineering from Stanford University.