The problem of estimating the parameters of structural equations when only equilibrium positions can be observed. For example, in the market for a particular good, if demand conditions vary and supply conditions do not, comparing prices and quantities at different times allows us to determine the supply equation; if supply conditions vary and demand conditions do not, we can estimate the demand equation; but if both supply and demand conditions vary, regressing quantity on price tells us nothing. The identification problem can be resolved only if either theory or the results of other studies inform us that some explanatory variables affect one side of the market but not the other.