AN ANALYSIS OF PROPERTY-CASUALTY INSURERS: Introduction 2

At least two prior studies identify and test for specific multijurisdictional tax plans. Collins, Kemsley, and Shackelford (1997) use data from U.S. tax returns to examine shipments of inventory from foreign parents to U.S. subsidiaries in the wholesale trade industry. Contrary to expectations, they find no evidence of transfer pricing abuses. Klassen and Shackelford (1997) show firms avoid taxes by shipping manufactured goods from states with favorable taxation. However, their inferences are limited because they rely on non-tax return data, aggregated at the state level.

This paper analyzes firm-level data, collected from the publicly-available annual regulatory reports used to compute tax bases and filed with multiple governments. The high quality of the data enables the analysis to move beyond simply documenting income shifting to identifying the allocations tax planners could manage to shift income and assessing whether these accounts are managed. Its findings should interest accountants, tax planners, and insurers. financial services

Specifically, we hypothesize that, in their statutory reports, insurers allocate premiums and losses from multistate policies to reduce total state taxes. To test this prediction, reported premiums at the insurer-state level, scaled by incurred losses, are regressed on state tax measures. If allocations are managed to avoid taxes, we anticipate an inverse relation between the tax rate and the premium-to-loss ratio, which is the industry’s standard measure of the price of a unit of coverage. Firm-specific prices are computed using premium and loss information from the annual regulatory reports filed with each state in which an insurer underwrites. Primary analysis is conducted on 12,573 insurer-state observations from 1993.