Reflections on Heckman and Pinto's Causal Analysis After Haavelmo""
Judea Pearl · 2013
Working paper. TECHNICAL REPORT R-420 November 2013 Reflections on Heckman and Pinto’s “Causal Analysis After Haavelmo” Judea Pearl University of California, Los Angeles Computer Science Department Los Angeles, CA, 90095-1596, USA [email protected] Abstract This paper reflects on a recent article by Heckman and Pinto (2013) in which they discuss a formal system, called do-calculus, that operationalizes Haavelmo’s conception of policy intervention. They replace the do-operator with an equivalent operator called “fix,” highlight the capabilities of “fix,” discover limitations in “do,” and inform readers that those limitations disappear in “the Haavelmo approach.” I examine the logic of HP’s paper, its factual basis, and its impact on econometric research and education. Introduction A forthcoming special issue of Econometric Theory, dedicated to Haavelmo’s centen- nial, will contain two papers on causation. The first is “Trygve Haavelmo and the emergence of Causal Calculus” (Pearl, 2013) and the second is “Causal Analysis After Haavelmo” by Heckman and Pinto (HP) . The HP paper is devoted almost entirely to the causal inference framework that I have summarized in (Pearl, 2013) and, in particular, to causal models that can be represented by Directed Acyclic Graphs (DAGs) or Bayesian Networks (Pearl, 1985; Verma and Pearl, 1988) and to the do-operator that acts on, and helps draw causal and counterfactual inferences from such models (Pearl, 1993, 1994, 2009; Spirtes et al., 1993; Strotz and Wold, 1960). This note reflects on the way HP present the do- operator, and highlights key features of the do-calculus that are not described in HP’s paper.