New Tools for Economic Analysis
Impossibility Results for Nondifferentiable Functionals
This paper is part of a larger project supported by the National Science Foundation that develops tools for the analysis of economic data using a decision-theoretic framework.
It shows that a number of recent statistical problems, arising in the econometric analysis of game-theoretic models of firm interactions, the analysis of missing data problems in economics, and work in biostatistics on optimal treatment regimes, have a similar structure and face similar challenges.By isolating a common mathematical property of these problems (a type of non-smoothness), the authors show that some standard statistical techniques cannot be applied but also develop a framework for developing better methods of analysis.
Keisuke Hirano is APS Professor of Economics at the Eller College of Management, University of Arizona. Jack R. Porter is an Associate Professor of Economics at the University of Wisconsin.
Forthcoming in Econometrica.