Abstract
When supply and demand are recursive, with uncorrelated cross-equation errors, least-squares estimation has no simultaneous-equation bias. Supply to a daily fish market is determined by the previous night's catch; hence this would appear to be a good example of a recursive market. Despite this, data from the Fulton fish market are treated in the literature, without adequate explanation, as coming from a market in which price and quantity are determined simultaneously. We provide the missing explanation, and in doing so reveal some issues about simultaneity that deserve better coverage in our textbooks and fuller consideration by applied econometricians.