AN ANALYSIS OF PROPERTY-CASUALTY INSURERS: Background and Hypothesis Development 3

States provide little guidance concerning the allocation process. The process is sufficiently murky that Harrington (1984, 614) contends any state allocations “depend on arbitrary assumptions that might make the resulting figures meaningless.” Private discussions with state revenue officials suggest tax audits primarily focus on determining whether all income is reported in some state. The allocations across states, which enable income shifting and are the focus of this study, reportedly receive less attention. Limited allocation regulations and enforcement lead to the paper’s hypothesis, stated in its alternative form:

H1: The premium-to-loss ratios that multistate insurers report to state tax authorities are decreasing in the marginal tax rate applied to their underwriting in the state. credit

Despite several factors suggesting Schedule T allocations are conducive to tax management, a countervailing implicit tax effect, inducing a positive relation between prices and taxes, biases against finding tax avoidance in this setting. Assuming competitive conditions and no tax-motivated management of the Schedule T allocations, the price of insurance should be bid up (down) in states that tax property-casualty insurers less (more) favorably until the after-tax returns to investing in underwriting are identical across all states. To achieve such equality, before-tax returns increase in state tax levies (Scholes and Wolfson 1992; Shackelford 1991; Guenther 1994). To the extent premiums and losses reflect adjustments in before-tax returns, our estimates of tax-motivated income shifting will be understated.