The Impact of State Taxes on Self-Insurance: Sales tax

The Impact of State Taxes on Self-Insurance: Sales taxIn addition, documentation that insurers engage in premium tax avoidance is prima facie evidence that insurers believe that they could bear some tax costs. Consistent with insurers bearing state tax incidence, Petroni and Shackelford (1999) report that insurers allocate a disproportionately large percentage of the premiums from multistate policies to less heavily taxed states. To our knowledge, no study documents that retailers engage in tax plans to avoid sales taxes, consistent with retailers avoiding the incidence of the sales tax.
Finally, this study differs from the sales tax studies along a research design dimension. The sales tax studies generally assess the effect of taxes on prices and assume that quantity changes in response. Unfortunately, unlike the consumer goods examined in the sales tax studies, which are homogeneous and numerous (e.g., Besley and Rosen, 1999, compare the prices of thousands of Big Macs, Kleenex, and similar goods across U.S. cities), insurance policies are customized financial contracts with extensive contingencies. Thus, comparing the rates of insurance policies across states would not be a fruitful exercise.
An alternative is to compare profit margins across states, e.g., total premiums in the state divided by all costs of providing coverage in the state. Unfortunately, data limitations prevent us from observing this profit measure. Its numerator, premiums, is biased. Petroni and Shackelford (1999) document that the premiums disclosed in insurers’ annual accounting reports are managed to reduce premium tax liabilities. Its denominator, total expenses, is unobservable. The only expense reported in the annual accounting reports for each state and by line of insurance is nominal incurred losses. Other expenses are segregated neither at the line of insurance level nor at the state level.
Bradford and Logue (1998) use the premium-loss ratio to measure price in their analysis of the impact of the Tax Reform Act of 1986 on the insurance industry. Although they are unimpeded by any cross-state shifting, they conclude that non-tax factors, such as interest rate swings, overwhelm the effect of Federal taxes on premium-loss ratios. Their inability to isolate Federal tax effects using intertemporal profit-loss ratios support the notion that profit-loss ratios lack the necessary power to calibrate the effects of state insurer taxes on self-insurance.
Consequently, we estimate the impact of state taxes on the market for property-casualty insurance by evaluating the relation because taxes and quantity, i.e., insurance coverage, as measured by incurred losses. On a more positive note, unlike the sales tax studies, which generally ignore quantity effects, this paper can address issues of social welfare concerning selfinsurance. Thus, this paper’s inability to estimate tax-price relations and its focus on tax-quantity relations is not necessarily a weakness, but simply a difference from most extant sales tax literature.