Among the many pieces of good fortune that have blessed my life was learning international economics, during my very early 1980s grad-student days at NYU, directly from Fritz Machlup. And one of Machlup’s many superb contributions to economics is his 1964 essay “The Mysterious Numbers Game of Balance-of-Payments Statistics.” (My colleague Dick Wagner recently reminded me of this article.) With characteristic thoroughness and clarity, Machlup explains the relevance of the fact that international-trade and -finance accounts are not facts of nature like the speed of light or the atomic weight of lithium but, rather, conventions that reflect human decisions (1) to classify and measure the monetary consequences of certain exchanges (here, particularly ones that cross political borders) and (2) about how to classify and measure these exchanges.
To understand the significance, therefore, of, say, a current-account deficit for some period of time requires an understanding of the reasons – the theory – that guided the decision to classify X exchange in one way rather than another way. Here’s Machlup:
The allocation of an individual item to a particular account is a matter of judgment, and the arrangement of particular accounts as between balance and offsets is a matter of economic theory, or at least a matter of interpretation of observations in light of theory [p. 142].
Machlup then chose a (then) relatively recent year – 1951 – to show how different classifications by different agencies (e.g., the I.M.F. and the U.S. Department of Commerce) can lead to strikingly different dollar figures.
The variety of figures supposedly standing for a very concrete experience, namely, the balance of payments of the United States in the year 1951, is impressive. The range is enormous: from a surplus of over $5 billion to a deficit of about $1 billion. Let us note that this is the sort of thing that so many people apparently seek when they say, “let us look at the facts”; these are the “empirical data,” the “records of observation,” which are taken as established facts and are to be explained with the aid of theory or, according to some, are supposed to suggest theories, or to test and verify theories. It should be understood, however, that these so-called observations, data, or facts really presuppose and contain a substantial amount of theory and that changes in the theoretical presuppositions or preconceptions may result in drastic changes in the observations, of the empirical data, of the supposedly stubborn facts [p. 147; original emphasis].
There may well be good reasons for government to measure, say, the total dollar volume of a country’s imports and exports of goods and services during a year. (Or, by the way, there may not be a good reason to do so. Choosing to measure such a thing itself is an exercise in theory, or reflects some theory, whether or not the decision-makers are consciously aware that that is what they are doing.) Just because, however, measuring the annual dollar value of imports and exports of goods and services is justified by reason A (but what, exactly, should be classified as “imports,” “exports,” “goods,” and “services”?), it does not follow that this measurement would have been justified by reason B. That is, perhaps (say) domestic business people gain on net from knowing such numbers; it does not necessarily follow that these numbers are relevant for economic theory.
As readers of this blog are all too aware, I loathe the common misunderstanding – even by many economists – of the “trade deficit.” Machlup’s superb essay points to one reason for my reaction.