Comment on "Excessive Ambitions" (by Jon Elster)

In "Excessive Ambitions," Jon Elster criticizes a wide range of social science aspirations to understand a complicated and evolving reality. Some of his analysis is to the point, but some is flawed, as explained in my comments. Crucially, however, his conclusions on empirical modeling are diametrically opposite to what is required-the problem has been a serious lack of ambition. And this is precisely the area where Elster is most guilty of 'criticizing others on the basis of third-party authorities.' Notwithstanding 'pitfalls and fallacies in statistical data analysis,' heterogeneous, high-dimensional objects like economies, which are subject to large, intermittent, and usually unanticipated, shifts, require ambitious approaches to characterize their behavior. I will try and explain how more ambitious empirical objectives can be achieved by automatic modeling methods which enhance human capabilities in tackling complicated data problems. En route, I re-emphasize the closely linked explanation for forecast failure recently discussed by myself, Michael Clements and Neil Ericsson in this journal. That leaves open the key issue as to why unanticipated shifts occur, and I speculate on that lacuna in existing economic theories, most of which omit any discussion of the mean levels of variables, and almost none address why such means might shift.