Format of Original
Foundation Journal Public Finance
Public Finance - Finances Publiques
Original Item ID
Shelves: HJ 109 .P86x Mem Periodicals
This paper presents an empirical model distinguishing evaders who cheat by filing fraudulent income tax returns from those who do not file. Using a maximum-likelihood procedure that corrects for sample selection bias, and data from Michigan's amnesty program, we estimate a linear probability model which relates the probability of filing to various economic and demographic characteristics. The results indicate that higher true income and automatic withholding raise the probability of filing, while males and single individuals are less likely to file. The evidence regarding a grouping of occupations often thought to be association with evasion is inconclusive.