Recent advances on testability in economic equilibrium models

The revealed preference program started by Paul Samuelson brought a Popperian view of what constituted true scientific discovery to economic theory.¹ In his Foundations, Samuelson states that “meaningful theorems” are hypotheses “about empirical data which could conceivably be refuted”.² He worried that by appealing to unobservable fundamentals such as preferences, the body of economic theory built upon the premise of rational individual choice lacked a scientific foundation.

While the program quickly succeeded in providing an empirical counterpart for the individual choice problem, its application to social outcomes, namely market interactions and games, proved more difficult, and only in the 1990s did the profession start to develop the testable implications of equilibrium concepts in these contexts. Carvajal et al. (2004) covered the early literature that followed the seminal contributions of Brown and Matzkin (1996) and Sprumont (2000). This survey is an update on the state of that literature covering the results of the last two decades. As the reader will see, the Journal of Mathematical Economics has played a key role in developing these ideas.