The welfare effect of income tax deductions for losses as insurance: Insured- versus insurer-sided adverse selection ¬リニ

a r t i c l e i n f o JEL classification: D82 H24 G22 Keywords: Income tax deductions for losses Informed insurer Adverse selection Kaplow (1992) shows in a complete-information environment that allowing income tax deductions for losses as partial insurance is undesirable in the presence of private insurance markets. This paper elaborates on Kaplow's finding by studying two extreme types of asymmetric information structures in private insurance markets: Either the insured or insurers possess superior information. It is shown that our derived result is consistent with Kaplow's if the insured have superior information; however, Kaplow's negative conclusion with respect to the income tax deduction will be overturned if insurers have superior information instead. A policy implication from our finding is that whether or not to allow an income tax deduction for losses needs to be more refined and, specifically, it should be tailored to the " adverse selection " information structures of private insurance. In an important paper, Kaplow (1992) asks the question: Should the government allow income tax deductions for uninsured losses as partial insurance? He shows that a tax system without such deductions Pareto dominates a tax system with the deductions. The intuition underlying Kaplow's result is that the availability of income tax deductions as partial insurance will distort and crowd out the purchase of private insurance, leading individuals to expose themselves to more risk than otherwise. Most OECD countries allow some form of tax deductions for personal losses such as medical expenses and casualty losses. 1 In view of this fact, Kaplow's negative conclusion with regard to the income tax deduction deserves careful study. Kaplow (1992) addresses his question in a complete-information environment, but argues that his finding remains robust in the presence of adverse selection. 2 In this paper, we elaborate on Kaplow's argument by studying two extreme types of adverse selection in private insurance markets. The so-called " adverse selection " in the insurance literature typically follows the seminal work of Rothschild and Stiglitz (1976). That is, those who are insured possess superior information to insurers, in the sense that the insured know the loss probabilities they face, but the insurers do not. Put differently, the insured's loss probabilities are their private information, unknown to the insurers. Recently, the issue of informed insurers has attracted considerable attention. This literature empirically detects some " adverse selection " information structures contradicting that of Rothschild …