Imperfect Enforcement of a Tax Threshold: The Consequences on Efficiency

In this paper, I build a model of firms' growth with heterogeneous entrepreneurs where audits for taxes are conditioned either on sales or employees and where entrepreneurs negotiate a bribe in exchange of a tax rebate when audited by tax inspectors. I show that the impact differential between the two auditing strategies depends on the regulation environment of a country. The model is calibrated and validated using firm-level data from Uganda. The model replicates well the Ugandan size distribution of firms as well as other key statistics relevant for the counterfactual experiments. Simulations show that, in a situation where Uganda could fully eradicate tax evasion and used the recouped taxes to subsidise production, productivity per worker could increase by as much as 45%. I also show that there could be a 16% annual gain in productivity per worker solely by conditioning audits on sales rather than on employees in Uganda.