We use data from over 20,000 firms in 60 countries to identify constraints on the growth of firms. We interpret managers’ answers to survey questions on the extent to which various aspects of their external environment inhibit the performance of their firm as measuring the shadow cost of constraints to their activities, not as direct measures of the constraints. These costs can vary with firm characteristics as well as with the magnitude of the constraints themselves. Our model reveals that, contrary to common practice, the importance of an obstacle to performance is not, except under very restrictive assumptions, measured by the coefficient on the reported level of the obstacle in a performance regression. We test the predictions of the model on the large firm-level dataset and show how the importance of different constraints varies across countries and how the cost of a constraint depends on the characteristics of the firm. We find that telecoms are less important, and taxes more important, as constraints on performance than the literature has previously identified.