Tax evasion in a principal-agent model with self-protection

Gatekeepers have an increasing role in taxation and regulation. While burdening them with legal liability for misconducts that benefit those who resort to their services actually discourages wrongdoings — as will be clarified in the paper — an alienation effect can also arise. That is, the gatekeeper might become more interested in covering up the illegal behavior and in cooperating with the perpetrator. Such perverse effects are difficult to detect and to measure. This paper studies the problem with respect to tax evasion by firms, by building upon the classical Allingham and Sandmo (1972) model and by providing a more detailed description of the "concealment costs" than that available in the literature, which often simply makes assumptions about their existence and their functional form. The relationship between a risk neutral firm owner aiming at evading taxes and a risk averse gatekeeper is described through a simple principal-agent framework. The paper highlights the role of legal rules pertaining to liability for tax evasion in shaping the parties choices, since concealment costs vary according to whether the risk neutral principal or the risk averse agent are held responsible when tax evasion is detected. The main result of the analysis is that there are simple conditions under which one can easily infer whether harnessing the agent is socially beneficial.