Pigou's parable of the belching factory imposing an externality on the neighboring laundry has elicited more controversy than one could have expected from such a simple situation. Ronald Coase claimed that the Pigouvian solution of taxes and subsidies was demonstrably inefficient while William Baumol recently defended Pigou by presenting a situation in which a tax placed upon the factory (without taxation or compensation to the laundry) optimizes resource allocation under pure competition. What if the externality affects the factory itself? Suppose the reduces the work efficiency of employees and causes ill health. It might be conjectured that nothing new is gained by adding the wrinkle of poisonous air which sickens factory workers, since the factory both generates the and is affected by it. By analogy, a city where everyone works in factories generating air pollution (which affects only the city) might be deemed consistent with Pareto optimality, presuming the factories compensate individuals for the air pollution with higher wages. 1 Similarly, recent policy discussions have advocated exporting air pollution by importing pollution-producing goods. Little concern has been given to the welfare implications since the republic of smoke would be experiencing the profit as well as the pollution from production. There are even proponents of the view that some regions of the United States ought to be kept quite clean, while others are allowed to become highlv polluted. Their reasoning is that

In his comment, Mark Ladenson contends that substantial differences in wages still exist between the North and the South even after accounting for differences in regional prices. We find his results objectionable on three grounds: a) his arbitrary exclusion of available data, b) his choice of a low budget cost of living as a deflator, and c) the possibility of heteroscedasticity which would render his testing procedure invalid. Ladenson objects to four of the five cities we used as our southern sample, all previous empirical work notwithstanding. These five cities were used in our analysis because they were the only five cities for which the necessary 1963 data were available. Similarly, the five cities included in our Northeast sample were the only Standard Metropolitan Statistical Areas (SMSA) for which complete 1963 data existed.' If we consider 1967 data, there are many more SMSAs in both regions than are utilized by Ladenson. (Our data sources are identical to Ladenson's: Census of Manufacturers 1967 and Handbook of Labor Statistics 1970.) In this study we use all the data available by: 1) examining wage differentials throughout the entire United States, 2) offering alternative definitions for the Northeast and South, and 3) using different deflators to measure differentials in the regional cost-of-living. The statistical techniques used are similar to those employed by Ladenson and in our previous paper, p. 934, except that we have added two additional dummy variables to account for the additional geographical regions, the North Central and western regions of the United States. Dummy variables, then, exist for the South, North Central, and West of the United States; consequently we are comparing wages in these regions to wages in the Northeast.2 Industry 21 of the Standard Industrial Classification