Having good longitudinal identifiers is important in empirical microeconomics, since researchers often need to be able to observe the same unit over time to make causal inferences. However, firm identifiers in Statistics New Zealand’s Longitudinal Business Database can be “broken” by, among other things, changes in the legal status of the firm. This paper proposes a simple method for repairing broken firm identifiers, making use of existing plant migration data. We show that making such repairs materially reduces the apparent rate of business entry and exit, and allows real economic phenomena, such as small business incorporation, to be observed for the first time.