Optimal Inheritance Tax under Temptation

In this paper we derive the expression for optimal inheritance tax when agents' preferences are subject to temptation and self control problem. We consider a dynamic stochastic model as in Piketty and Saez (2013) where agents are heterogeneous in terms of bequest motives and labor productivities. In such a setup we show that the optimal inheritance tax rate decreases with the level of temptation, and thus it works as an incentive mechanism that leads to more bequests and makes succumbing to temptation less attractive. In fact, when temptation is acute, a subsidy may be justified at any percentile of bequest received. This holds independent of the variation in the models used in the literature as well as the assumption of labor elasticity. The study also reveals some interesting observations. Though from the point of view of incentives, this result has the same essence as in Krusell et al. (2010) where temptation justifies a subsidy on capital, we show that unlike their other policy prescription, the long run equilibrium does not demand a constant subsidy. Thus, even under temptation and self control issue, the standard Chamley - Judd result which recommends zero capital tax in the long run is still valid. However, in a setup that is comparable to Farhi and Werning (2010), our paper shows that in the presence of temptation and self control, if dynamic efficiency holds, optimality always requires a subsidy independent of whether social welfare function puts zero or positive direct weight on the children. This is in direct contrast to Piketty and Saez (2013). A calibration using the same micro data used by Piketty and Saez (2013) shows that the drop in inheritance tax is significant in the presence of temptation and self control.