As a graduate student a long long time ago I studied some of
Haavelmo's papers (mostly in econometrics courses). His influence on
econometrics has been profound and yet not sufficiently recognised. He
was perhaps the first economist to look at causal relations (a no-no in a
positivist's world) seriously.
______________________________
Source: http://en.wikipedia.org/wiki/
Judea Pearl wrote "Haavelmo was the first to recognize the capacity of economic models to guide policies" and "presented a mathematical procedure that takes an arbitrary model and produces quantitative answers to policy questions". According to Pearl, "Haavelmo's paper, The statistical implications of a system of simultaneous equations,[4] marks a pivotal turning point, not in the statistical implications of econometric models, as historians typically presume, but in their causal counterparts."[5] Haavelmo's idea that an economic model depicts a serious of hypothetical experiments and that policies can be simulated by modifying equations in the model became the basis of all currently used formalisms of causal inference. It was first operationalized by Strotz and Wold (1960)[6] who advocated ``wiping out selected equations, and then translated into graphical models as ``wiping out incoming arrows.[7][8] This operation has subsequently led to Pearl's do-calculus[9][10] and to a mathematical theory of counterfactuals in econometric models.[11][12] Pearl further speculates that the reason economists do not generally appreciate these revolutionary contributions of Haavelmo is because economists themselves have still not reached consensus of what an economic model stands for, as attested by profound disagreements among econometric textbooks.
______________________________
This issue of Econometrics Theory seems to be exclusively
devoted to the work of Haavelmo. I have just looked at the TOC and it
looks like a treasure trove. Should keep me busy for some time.
I can do no better than quote a passage from Wikipedia:______________________________
Source: http://en.wikipedia.org/wiki/
Judea Pearl wrote "Haavelmo was the first to recognize the capacity of economic models to guide policies" and "presented a mathematical procedure that takes an arbitrary model and produces quantitative answers to policy questions". According to Pearl, "Haavelmo's paper, The statistical implications of a system of simultaneous equations,[4] marks a pivotal turning point, not in the statistical implications of econometric models, as historians typically presume, but in their causal counterparts."[5] Haavelmo's idea that an economic model depicts a serious of hypothetical experiments and that policies can be simulated by modifying equations in the model became the basis of all currently used formalisms of causal inference. It was first operationalized by Strotz and Wold (1960)[6] who advocated ``wiping out selected equations, and then translated into graphical models as ``wiping out incoming arrows.[7][8] This operation has subsequently led to Pearl's do-calculus[9][10] and to a mathematical theory of counterfactuals in econometric models.[11][12] Pearl further speculates that the reason economists do not generally appreciate these revolutionary contributions of Haavelmo is because economists themselves have still not reached consensus of what an economic model stands for, as attested by profound disagreements among econometric textbooks.
______________________________
No comments:
Post a Comment