Gibrat's Law with Mild Nonrandom Growth

Gibrat's Law (GL) has repeatedly failed to gain full empirical confirmation in specific industries. This study offers a deliberately favorable opportunity for full confirmation in the truckload sector of the U.S. trucking industry where firms are highly homogeneous. As such, most nonrandom determinants of growth remain very similar for all firms, so significant differences in growth rates are not expected. Still, there is only incomplete support: (1) long term growth rates are not equal for all firms, but the differences are small and not size-related except for the smallest firms, and (2) the size distributions better approximate lognormal when the smallest firms are excluded, but in no case does the variance rise over time. This suggests that for most other industries, where nonrandom growth should be much stronger, GL would seem unlikely to play more than a minor role in portraying actual firm growth or the evolution of market structure.