Price Changes and Optimum Taxation in a Many-Consumer Economy

This paper follows recent work on the welfare effects of small perturbations from an initial general equilibrium with some distortions. Dixit (1975) and Hatta (1977) have analysed the one-consumer economy, and found several simple prescriptions for policy changes to improve welfare. Guesnerie (1977) has studied the many-consumer case when constant returns to scale prevail or profits are taxed at 100%. Diewert (1977) considers profits as returns to artifically defined fixed factors, and allows taxation of profits at any specified rates. In both of these last two papers it turns out that the simplicity of welfare-imoroving policies is lost in the many consumer case. Increases inb a Bergson type social welfare function depend very crucially on the distribution of ownership of fixed factors. General formulae mean little, and we must consider very special cases in order to obtain concrete results.Improvements in the Pareto sense are even harder to generate by simple formulae. In this paper I broadly follow the approach of Diewert, but try a different tack by concentrating on the question of when Bergson or Pareto improvement is not possible, thus obtaining local necessary conditions for welfare optimality Pareto efficiency of the initial equilibrium. This approach helps shed new light on aa question that has been much discussed (eg Dasgupta and Stiglitz (1971) and 1972)) : if profit taxation alone can yield enough revenue to finance government expenditures, is it optimal to rely on that alone and leave the commodities untaxed? The approach has added advantages of allowing a naturally paralled treatment of welfare optimality and Pareto efficiency, and of tying together diverse previous models and results.