In most macro-economic models of Canada, the United Kingdom and the United States, prices and wages are determined in a Phillips-curve framework. In the original formulation the rate of change of wages is dependent on the rate of unemployment only. In subsequent formulations effects of rate of change of consumer prices and of productivity have been brought in to explain the rate of change of wages. However, the main emphasis is still on the trade-off between inflation and unemployment. In this paper we argue that, while this approach may be reasonable for a short-term model, it is inappropriate for a longer-run analysis. We argue that this approach does not explain the trend rate of growth of wages (and prices), but only the movement around the trend, with reference to unemployment and such other factors. In the longerrun analysis the trend rate of growth is the important matter to explain, and this requires a reformulation of the prices and wages equations. An important piece of information we use for this reformulation is the fact that the wage share has been stable in a secular sense, and this implies that in the long run both price and productivity changes are passed on to wages.2 This leads us to a productivity theory of real wages which we find explains reasonably well the movements in wages in the post-war period in Canada. Since the real-wage equation implicitly assumes absence of "money illusion" in the labour market, it determines only the ratio of money wages to prices and not their absolute levels. The latter has to be determined with reference to money supply, level of real output, and so on. Since these equations are based on classical economic theory and we draw inspiration from the work of such eminent neoclassicists as M. Friedman and D. Patinkin, we call our approach a neoclassical approach to prices and wages to be distinguished from the Phillips (or Keynesian) approach. As we note below, when viewed in the framework of long-run analysis, there is still some trade-off between unemployment and inflation, but the degree of trade-off is not quite so great as that obtained by traditional analysis. The main message