Organizations and Markets

Robustness

18 July 2006 at 12:37 am Nicolai Foss Leave a comment

| Nicolai Foss |

Empirical economists tend to exalt the property of “robustness.” Robustness is a priori taken to be epistemologically virtuous. However, this is not so obvious as it may seem. As Kevin Hoover points out in the preface to the most recent (June) issue of the Journal of Economic Methodology, most of which is dedicated to a symposium on “Fragility and Robustness in Econometrics”:

The idea that robustness is an epistemic virtue across the entire range of measures is not obvious. Suppose that the efficient-markets hypothesis works well in the United States but not in China. Does that count against the efficient-markets hypothesis? Or does it instead point to a substantive difference between American and Chinese financial markets?

A lot of things in econometrics are closely related or depend on robustness, such as omitted-variable bias and extreme-bounds analysis, but robustness has seldom been explicitly discussed. The contributors to the symposium are John Aldrich, Aris Spanos, and Jim Woodward. For somebody who is not an econometrics buff, Jim Woodward’s “Some Varieties of Robustness” is likely to be particularly informative.

Entry filed under: - Foss -, Methods/Methodology/Theory of Science, Recommended Reading.