The Rise and (Recent) Decline of Mathematical Economics

This paper presents a measurement of the degree to which economics has become more mathematical and quantitative over the period, 1948-1984. Several points should be made about our procedure. We chose 1948 as the starting date for measuring the spread of mathematics in economics on the grounds that Samuelson's Foundations was published in the previous year.' Purportedly, this book more so than others, paved the way for the expansion of mathematical economics. We proxy the extent of mathematics in economics with a count of numbered equations (per page, per year) in the American Economic Review (AER) over this period. This is a convenient measure for an obvious reason it is something, the main thing, that can be counted with respect to the mathematical content of economics articles. It is not a particularly good measure for a number of reasons. An equation count per se jumbles mathematical economics and econometrics, as we counted both types of equations., We might better say therefore that we are measuring the extent of quantitative rather than mathematical economics. We are also mixing apples and oranges. As a general rule, better math is leaner math. Papers with fewer equations may therefore be "more" mathematical than papers with lots. This means that our measure is biased to an extent that we cannot determine. Short of reading every mathematical paper and making a subjective judgement in this regard, we simply have to live with this problem. Moreover, we measure equations by volume year, and we include the May Proceedings issue in our count. Our measure thus includes refereed and unrefereed contributions. Finally, we have only counted the growth of mathematics in the AER. Whether the pattern of the AER reflects the profession at large is an open question.