Measurement, Quantification, and Economic Analysis: Numeracy in Economics

John Smithin

In an important recent paper, Summers (1991) explores the question of why “formal econometric work” has been so apparently unpersuasive in settling macroeconomic controversies and in setting the agenda for research, in spite of the huge volume of intellectual and other resources which have been devoted to it. He makes the case that an alternative approach, “pragmatic empirical work”, has been and will be more fruitful.

In one sense, of course, Summers has simply written down explicitly in an academic journal what has been tacitly known, if not acknowledged, by much of the economics profession for some time. Given the incentive structure of contemporary academia the implicit goal of a large percentage of publication activity is mainly to demonstrate “technical virtuosity” (Summers 1991:146) rather to advance knowledge or engage in substantive debate on the issues. ¹ Summers contribution, nonetheless, is doubly significant in that it comes from a leading scholar at a major university, whose own research clearly meets or exceeds the peer-defined technical standards of the day.

In this chapter, however, it will be argued that in order to fully understand the linkages between economic theory and the “facts of experience” ² (linkages which formal econometrics is supposed to provide but has failed to do) it is necessary to push the debate on beyond the position taken by Summers—specifically that, in practice, the major examples of instances in which the macroeconomics research agenda has been seriously affected by empirical evidence have occurred when the prevailing theory of the day has failed to correspond with some very crude large-scale empirical generalizations which are nonetheless widely accepted. These empirical generalizations have typically been such that little or no formal technique, even at the level of Summer’s “pragmatic empirical work”, has been necessary to apprehend them, and have had maximum impact in situations when some competitor theory was in the wings which could (apparently) explain the empirical anomaly when the prevailing orthodoxy could not.

In making this case the chapter will draw on the work of Pheby (1988:

-365-