The positive correlation between average establishment-level productivity and measures of competitiveness is oft-observed but still controversial, despite its implications for competition policy and its centrality in the debate over the origins of the productivity effect of trade liberalization. This paper considers two competing explanations for the existence of the correlation: a causal relationship between competition and productivity, historically known as "X-inefficiency", and a market-level dynamic selection story that has gained ground in the trade literature. This paper demonstrates that the two effects are econometrically separable. Two empirical approaches are developed: a regression of quantiles and a selection correction procedure derived from a model of Markov-perfect industry dynamics. Both methods are applied to the ready- mix concrete industry, and results suggest that the correlation is almost exclusively attributable to X-inefficiency.
