What do we know about the real exchange rate? A classical cost of production story

This paper argues that exchange rate models rooted in the theory of Purchasing Power Parity (PPP) and balanced trade are fundamentally mis-specified, as evidenced by the disjuncture between: (1) the empirical evidence, which largely refutes PPP; and (2) the empirical result that 'real' productivity shocks are associated with observed secular trends in exchange rates. In the former case we have a theory without convincing evidence, and in the latter case we have empirical evidence in want of a consistent theory. If looked at from the perspective of a 'cost of production' theory of prices, such empirical results might not be so theoretically anomalous. So-called 'real' variables (especially productivity and unit labor costs), let in through the side door as 'shocks' to PPP equilibrium, may in fact be part and parcel of the formation of prices of production on an international scale through capitalist competition. The primary conclusion is that the empirical evidence supports a cost of production theory of the terms of trade and the real exchange rate. The empirical evidence in support of the Balassa-Samuelson model of the exchange rate is re-interpreted in this light. In this interpretation, parity holds only in terms of rates of return on investment which, in the classical tradition, are presumed to equalize across industries internationally.