Optimal Taxation under Income Uncertainty

Xianhua Dai

Optimal taxation under income uncertainty has been extensively developed
in expected utility theory, but it is still open for inseparable utility function
between income and e ort. As an alternative of decision-making under uncertainty, prospect theory (Kahneman and Tversky (1979), Tversky and Kah-neman (1992)) has been obtained empirical support, for example, Kahneman and Tversky (1979), and Camerer and Lowenstein (2003). It is beginning to explore optimal taxation in the context of prospect theory, for example, Oswald (1983), Tuomala (1990) in conventional setting without utility interdependence, and Kanbur, Pirttila, and Tuomala (2008) for separable value functions between income and effort. It is challenging in the prospect theory to treat with optimal taxation for inseparable value function between income and effort. In this paper, we model taxation under income uncertainty by sufficiently considering government's risk aversion and individuals' loss aversion. We obtain its sufficient condition for the first order approach to general value functions including inseparable value function between income and effort, hence generalizing Oswald (1983), Tuomala (1990) to optimal taxation with utility interdependence, and Kanbur, Pirttila, and Tuomala (2008) to inseparable value functions between income and effort.

Key Words: Optimal taxation; Income uncertainty; Moral hazard; Expected
utility theory; Prospect theory; Risk aversion; Loss aversion.
JEL Classification Numbers: D81, H21.