How should external validity problems be tackled? Because the philosophical and economic literature provides little help, I look for inspiration somewhere else: in the actual techniques used by experimental economists when they engage in applied science. What do economists do when they have to make sure that their laboratory results will work in the field?

This chapter is primarily descriptive in character. I focus on a celebrated application of game theory and experimental economics to solve a real-world problem: the construction in 1993–94 of a new market institution, by means of which the Federal Communications Commission (an agency of the U.S. government, FCC from now on) allocated a peculiar kind of good (telecommunication licenses). One of the key themes of this book – that experimental science is much more than theory testing – emerges forcefully again; but more importantly, this chapter provides the first concrete example of the procedures scientists use to bridge the gap between experimental and nonexperimental circumstances. In the chapters that follow, then, I try to draw some lessons from this case and generalize them to build a methodology of external validity inferences.

The market for telecommunication systems

A “decentralizing wave” hit the American economy in the eighties and nineties. A new political imperative prescribed the replacement, whenever feasible, of centralized, bureaucratic systems of allocation with market processes.