Gibrat's law is augmented to produce a more general stochastic model of concentration consisting of growth, entry and exit processes. Empirical facts on growth, entry and exit reported elsewhere are employed to calibrate the model, which is then repeatedly simulated, generating concentration data for 280 hypothetical industries grouped under 14 types. The simulated data are used to investigate the importance of growth, entry and exit in shaping concentration development. Most important is systematic firm-level growth. Entry and exit are much less significant, as is random growth variability, which features prominently in previous stochastic models. Parallels with previous inter-industry research are drawn.