Shaun Hargreaves Heap and Yanis Varoufakis
Experimental economics has a long history, but in the last decade there has been a proliferation of dedicated laboratories yielding a large body of experimental results (see Roth (1988) for a review of this history). This chapter assesses the contribution of the experimental approach to the neoclassical canon. We set the scene in the next section with a brief discussion of the methodological role of experiments in economics. The third section focuses on the experimental evidence with respect to one aspect of neoclassical economics: its theory of decision-making. These two sections feed directly into the final sections which is concerned with the implications of these results for both method and neoclassical economics.
Neoclassical economics has empiricism to thank for a strong defense of its often puzzling assumptions and concepts. 1 By referring the critic to the “tribunal of the facts”, it blunts all deductive criticism, asking instead the simple question: “Does the theory predict well?” If to predict is also to explain, then any theory which is consistent with its own premises can be good and proper depending entirely on its predictive capacity.
Many a criticism has stumbled on this empiricist defense, and yet there has been an undeniable and growing discomfort with its use over the last forty years. For example, econometrics may have prospered as a result of the commitment to empiricism, but the doubts have been growing at the same time. Unlike statisticians working in natural scientific or biomedical fields, economists have failed to make their statistics equally respectable. Instead, there is more than a hint of fool’s gold and the whiff of alchemy as the regressions pile up (see Hendry 1980). Indeed, we may laugh at the suggestion of “lies, more lies and statistics”, but it is a slightly nervous laugh because few economists will not at one time or another have been frustrated by the failure of the econometric evidence to throw any real light on a major