In his seminal 1974 paper, Barro noted that Ricardian equivalence may fail if agents face undiversifiable, idiosyncratic risk and governments finance tax rebates with higher future tax rates [5]. Higher tax rates reduce the dispersion of future, after-tax income, decrease the precautionary demand for savings and hence tend to stimulate present consumption. Barro concluded, however, that the stimulus to consumption would be minor and would likely be offset by induced reductions in future labor income. The recent revival of interest in the precautionary savings motive has refocused attention on this Tax-Insurance Effect (TIE) [2; 8; 24; 27]. In particular, two highly influential papers, one by Barsky, Mankiw and Zeldes [6] and the other by Kimball and Mankiw [19], argue that the Tax-Insurance Effect generates consumption responses roughly equivalent to those obtained from Keynesian models where agents ignore future tax liabilities entirely. Although the existence of a precautionary demand for savings and, by implication, a TaxInsurance Effect, has received strong empirical support [9; 15; 21], the magnitude of the effect remains largely unknown. The two most difficult empirical issues involve quantifying the degree of uninsured, idiosyncratic risk faced by typical agents and determining the appropriate form and shape of agents' utility functions. On these issues, the profession is far from a consensus. Studies from the labor economics literature [1; 17] suggest idiosyncratic risk may be large and persistent whereas a recent paper by Heaton and Lucas [16] finds idiosyncratic risk to be small and largely transitory.' Similarly, the vast literature surrounding the equity premium puzzle highlights the profession's lack of consensus over the degree of risk aversion exhibited by typical agents. In this paper, I circumvent some of these controversial issues by extending Barsky, Mankiw, and Zeldes' model to include an explicit labor supply decision. The introduction of a labor supply decision allows me to bring evidence from the literature on optimal income taxation [22; 25] and the literature on the welfare cost of taxation [3; 4; 7] to bear on these issues. The idea is simple: Under the conditions that make the Tax-Insurance Effect large (i.e., very risky income distributions and a high degree of risk aversion), agents will place great value on tax-induced