We use a large administrative tax-returns panel dataset merged with tax audit database to estimate the effect of real-world operational tax audits on subsequent tax behavior. Our identification strategy and the institutional setting that we consider enable us to address potential endogeneity related to non-random selection of taxpayers to be audited. We find a positive and lasting effect of audits on subsequent reported income. However, in line with theoretical predictions, taxpayers do not increase tax compliance when the tax authority does not assess a positive additional income. Our results are robust to a variety of specifications and samples.

No abstract

Italy, which ranges among the OECD countries with the highest share of shadow economy, has adopted in 1998 a peculiar audit scheme (Studi di Settore), for small and medium enterprizes and self-employed. This scheme is based on a particular interaction between the Tax Agency and taxpayers, where the Tax Agency unveils part of the information used to develop its audit rule. We study this scheme by means of a simple theoretical model and test it using a sample of 23,000 firms in manufacturing sectors in 2005 tax year.A number of theoretically relevant relations are confirmed. In particular, reports made by taxpayers seem to be positively associated to the firm's size. When taxpayers know that the probability to be audited decreases, they tend to report less. Other factors which are expected to influence the behavior by taxpayers, have no or ambiguous impact on reporting behavior.

Income tax evasion by small firms has been seldom investigated mostly because of lack of data. In this paper we use a large data set produced by the Italian Revenue Agency for this project to analyse a recent policy to contrast business income tax evasion. Since 1998 Italy has adopted a method to audit small businesses (Studi di Settore), which defines the probability of a tax audit based on presumptive and reported levels of sales. In 2007 a letter campaign was implemented by the Italian Revenue Agency aimed at reducing manipulation of reports by threatening that if the "anomaly"was repeated with the 2008 tax declaration, the probability of a thorough tax audit would have drastically increased. By using difference in difference with matching methods on a sample of about 50,000 treated firms and 95,000 controls, we find that the letter campaign had a positive and statistically significant average effect on treated firms. A cost-benefit analysis of the policy suggests that the letter campaign generated a net increase of revenues of about 140 million euros.