Last Saturday my advisor informed me that he never wanted to read my dissertation again, which was his way of saying he was ready to sign off. So I thought I would amuse everyone (well, me anyway) with a very brief recap of my findings. Let me assure you that there is no Mormon angle to this work, so if you are offended by the secular, feel free to move on.
I looked at the formal and informal (i.e., noncompliant) labor markets in Brazil. I was interested in how much it cost people, wage and benefits wise, to be illegal. I was also interested in the extent to which different kinds of illegal behavior go together. So if you are insufficiently productive to work for more than the minimum wage, how much more likely are you to evade payroll taxes? Fortunately, Brazil has extensive survey data with people very willing to talk about their illegal behavior. The workers, after all, are not directly punished, only the firm. All the results are for men between the age of 15 and 55.
I should mention that in the U.S., the textbook model is higher minimum wage = lower employment. But it is rather hard to get far in the U.S. data because illegal workers are difficult to survey and the minimum wage is comparatively low. In the Northeast part of Brazil, there are some groups where half or more of the workers are working for less than the minimum wage! This lets one estimate models that may apply to the U.S., but could never be estimated on U.S. data.
What I found was that the labor laws that created an informal market cause the informal workers to average about 25% less than they would earn if they were not breaking a law. Thus the “informality tax” is, on average, about a quarter of illegal’s salary. This is the cost of evading the law. Raising the minimum wage 10% increases the violations of the payroll tax law by about 2%, but in some groups as much as 8-9%. This suggests that raising the minimum wage can substantially increase the relative number of illegal workers. Also, there is no noticeable employment loss, those that would be unemployed just shift into illegal work.
Of course, all of this is based on a set of modeling assumptions and imperfect sample data, so there’s certainly room for error…