The intent of air pollution regulations is to induce/coerce plants to limit source emissions that lead to air pollution, so that localities can meet specific, nationally-established air quality standards. The level and strength of air quality regulations vary intentionally across U.S. counties, depending on whether or not the county is in attainment of the national standards. Plants in non-attainment counties face much stricter regulations and more active enforcement, raising the costs of operating plants in those counties. Regulators have also tended to concentrate their attention on the largest polluters, with smaller establishments facing little or no regulatory pressure. While this regulatory strategy has generally been very successful in curbing many forms of emissions, it has had other unintended and potentially costly impacts on firm decisions, which we explore in this paper.

These unintended side-effects could include relocation of polluting industries from more to less polluted areas, downsizing of plants in sectors with historically large plant sizes, and relative proliferation of small-scale enterprises in some industries. There may also be changes in investment patterns: the timing of investments in existing plants, and the decision as to whether and when to open new plants or close old ones. Some of these effects may be socially beneficial: moving polluting industries to attainment counties could mean moving to less densely populated areas, lowering total social costs from pollution. Others may be less socially beneficial, if firms are tending to establish smaller, less efficient plants in order to avoid regulatory attention. This paper does not attempt to quantify these broader social welfare impacts, but instead focusses on identifying and quantifying key effects of air quality regulation on plant decisions.

The data used in the analysis come from the Longitudinal Research Database (LRD), maintained by the Center for Economic Studies of the U.S. Census Bureau. The LRD links data on firms and plants over time. We focus on data from the Census of Manufactures, conducted every five years over the 1963-1992 period. This allows us to track new plant openings and plant closings, along with plant size, at five year intervals, during the period of time when federal air quality regulations were being implemented and strengthened, as well as a comparison period (the 1960s) when there were no federal regulations. We estimate the differential responses to regulation

across different industries, different sized plants, and different firm types (single- vs. multi-plant firm). We then examine how these differential responses are reflected in changes and differentials in location and investment patterns.