What is identification?

Among economists, the conventional way of finding out what demand or supply functions look like is statistical: one gathers appropriate data and then asks what the relationship between them is. Early attempts to estimate demand functions, particularly by Moore (1914), led to a number of puzzles, chief among which was that some demand curves were apparently upward-sloping, in particular the demand for pig iron. This strange finding was ultimately explained in Working's (1927) elegant article, which marks the first statement of what has come to be known in economics as the ‘identification problem’. This is the question of when data drawn from observation of the world is able to identify a theoretical construct. It stands at the heart of applied economics, for economists are very fond of constructing theories, but perhaps less fond of testing or calibrating them with care. Working's explanation of why Moore's ‘demand curve’ for pig iron had the ‘wrong’ slope is still instructive, and has led to great improvements in the quality and usefulness of economic measurements.

Working's insight seems simple, but it has profound consequences. It is that economic theory says that observed prices and quantities are generated not by a demand curve alone, but by the intersection of a demand and a supply curve.