The Problem of Passive Observation

Abstract

Trygve Haavelmo's methodological manifesto “The Probability Approach in Econometrics” not only laid down the paradigm for the research pursued at the Cowles Commission, but also sets out a strategy for measurement outside the laboratory. His conceptualization of “passive observation” is still very useful for understanding measurements where intervention is not possible. Haavelmo's classic is very rich: it provided the framework for introducing probabilistic methods in econometrics and a profound discussion on invariance (autonomy). These subjects are well treated by various historians of econometrics. This does not, however, apply to the problem of passive observation. It is only mentioned, if it is mentioned at all, in relation to the discussion of autonomy, but that is it. This essay will give a reconstruction of his discussion of “factual” and “potential” influences, which provided Haavelmo the framework to discuss “Nature's experiments” and will allow us to discuss measurement outside the laboratory more generally.