When Industries Become More Productive, Do Firms?

This paper investigates two explanations for why industries might become more productive over time. The first explanation, termed the real productivity case,' is one in which firms become more productive and this leads to more productive industries. The second explanation, termed the rationalization case,' is one in which firm productivity is constant, but productive firms expand while less productive firms either shrink or exit. Each case has very different implications for factor markets, long term growth prospects, and public policy regarding productivity. Further, one can only distinguish between these two cases with plant- or firm-level data. We investigate the empirical relevance of the two cases using the Chilean manufacturing census. We find that the rationalization case explains much of the measured increase in industry productivity. When industry productivity fails, the rationalization case appears much less important. We also contribute to the applied econometric literature on productivity estimation as we show that the value-added production function is especially well-suited to a simple extension of recent methods developed by Oiley and Pakes.