Essays on estate taxation.

Abstract

Essays in this dissertation apply the optimal taxation approach to analyze various arguments for using estate taxation. In the first chapter, it is demonstrated that estate taxation implicitly provides longevity insurance. Consequently, it can be used to correct imperfections in the annuity market. The optimal linear estate tax rate is shown to be positive when annuities are imperfectly priced. A calibration suggests that, in practice, the insurance effect reduces the cost of rising revenue using the U.S. estate tax by between 4 and 30 percent. In a model with private mortality information and heterogeneity of skills, it is demonstrated that a redistributive income tax, Social Security and an estate tax should all be used at the optimum. The second chapter considers estate taxation in a multi-generational context with endogenous wealth accumulation and inequalities of both inheritances and skills. Assuming that skills are perfectly correlated within a family line, the paper characterizes the optimal steady-state estate and labor income taxes. Bequest taxation in the steady state should be governed by the same rules as other commodity taxes if private and social benefits to bequeathing coincide. Otherwise, inheritances should be treated analogously to income, so that a redistributive estate tax becomes optimal. The third essay shows that an estate tax may be a part of the optimal redistributive tax system if people are able to disguise their labor income as capital income. Without capital income taxation, tax revenue would be eroded by income shifting between different tax bases. The estate tax is then a useful addition, because when one unwanted distortion must be introduced, it is optimal to distort addititional margins. The final essay contains a switching regression approach to estimate both the strength and frequency of bequest motives. Two types of parameters are estimated: structural parameters of the utility function, and indicators of the presence of a bequest motive. Having children is found to be a poor indicator of having a bequest motive, while having a will is the strongest indicator. Approximately 60% of the population has a bequest motive, which is arguably strong but usually becomes inoperative if one lives long enough.