Fritz Machlup, “Why Bother with Methodology?” (1936)

Fritz Machlup’s 1936 Economica article is a brief single-author methodological note in economic theory, written after debates over Lionel Robbins’s definition of economics and Ralph Souter’s resistance to disciplinary “regimentation.” Its scope is modest but pointed: Machlup does not construct a full philosophy of economics, but shows that method matters for ordinary causal arguments.

Simple reasoning about a few simple economic propositions will show whether or not methodological reflections can do anything for the economist’s work.

The article turns on three hypothetical propositions: an abundant wheat crop lowers wheat prices; dearer labour and cheaper capital induce labour-saving inventions; endangered banks receive central-bank credit. Robbins would count only the first as “pure economic theory” and treat the others as data; Souter would stress that all three are hypothetical and assumption-dependent. Machlup accepts the logical similarity, but asks whether they have the same degree of general validity within an economic conceptual scheme.

Are these three statements imposed, so to speak, by “economic laws”? Are they of equal economic relevance? Are they equally true or false, equally necessary or probable?

Machlup first rejects a crude distinction by motive. Wheat growers, engineers, manufacturers, and central bankers can all be understood as responding to incentives: price, profit, duty, reputation, or job security. Nor does he deny that any of the three statements may fail under changed assumptions. A farm pool may buy the surplus wheat; inventors may fail despite effort; central bankers may refuse expansionary credit. The question is not whether the statements are conditional, but what sort of conduct they presuppose.

His central conceptual move is to rank assumptions by the anonymity and generality of the behaviour involved, a vocabulary he links to Alfred Schütz. A wheat-price fall depends on the highly anonymous conduct of many competitive sellers. By contrast, invention depends on uncertain creative success, and central-bank action on the decisions of a few identifiable people.

The type of the competitive seller is not so likely to disappear from the economic stage as the type of the benevolent central bank manager.

This permits Machlup to side partly with Robbins without accepting any crude hierarchy of “dignity.” Political intervention, induced invention, and central-bank policy are economically relevant and may be caused by economic conditions, but they enter analysis as lower-generality data unless their special assumptions are made explicit.

According to the previous considerations it is by no means a “crime against logic”, and indeed has its definite usefulness, to hold that only statement (1) is based on pure economic theory, while statements (2) and (3) are best regarded as data for the application of further economic analysis.

The practical force of the argument appears in his example from international transfer theory. A theorist may claim that foreign-loan withdrawals create an overwhelming demand for foreign exchange and that the trade balance cannot adjust. But this silently assumes central-bank credit to debtors; without that policy, debtors’ cash balances limit the demand. The mistake is methodological: a low-generality institutional assumption has been smuggled into a causal chain as though it were a high-generality market proposition.

“Data” should be stated expressis verbis, because they are factors of a lower degree of general validity within the particular scheme.

Machlup’s thesis is therefore a defense of methodology as disciplined attention to the hidden status of assumptions. Its relevance lies in making economists mark where anonymous market behaviour ends and institution-specific, political, inventive, or personal conduct begins.

The failure to make distinctions between statements of different order often has serious practical consequences. For such reasons we ought sometimes to bother with methodology.