Property-Liability Insurer Reserve Errors: A Theoretical and Empirical Analysis

Abstract

Property-Liability Insurer Reserve Errors: A Theoretical and Empirical Analysis Introduction The management and reporting of income by property-liability insurers have long been legitimate concerns of regulatory and investment communities. Regulators have been concerned that investors and policyholders have fair representations of an insurer's financial status. Investors logically have been concerned with the yield and relative riskiness of their investments. When levels of risk and return are obscure to investors and regulators, stocks may not be valued properly, policyholders may not pay equilibrium-level premiums, and the ability of the firm to pay claims arising in the distant future may be difficult to judge. Detection of earnings manipulation is difficult at best, for the possibility of income management exists in every industry and differences in opinion develop over the proper measurement of income. The insurance industry is no different from other industries in its opportunities for income management. For example, the uncertainties involved in estimating future claim costs lead not only to reserve errors, but also to possible intentional income manipulation through the overstatement or understatement of reserves. Ultimately claim costs become known and the exact level of reserve errors is detected. Loomis (1984) reported that American Express received additional regulatory scrutiny in 1984 and was the defendant in an investor suit, resulting partially from reserving practices of an insurance subsidiary, Fireman's Fund, from 1978 through 1983. Reserve errors often are viewed either as a deliberate attempt by management to manipulate reported earnings, or as the result of sheer misestimation. Why management would choose to create reserve errors and whether reserve errors contribute to the value of the firm have not been developed theoretically; nor has there been much discussion of tax incentives arising from the use of reserve errors. Most of this article addresses the issue of why reserve errors arise. Specifically, the purposes of this article are to develop a theory of reserving practices of property-liability insurers and to test the theory empirically using a larger sample than do many earlier studies. It will be hypothesized that an insurer has as its objective the maximization of cash flow subject to smoothing constraints and uncertainty. Related Literature The literature on reserve estimation is exhaustive. Salzmann (1984), Peterson (1981), and Skurnik (1973), among others, have detailed numerous acceptable methods of estimating reserves. Salzmann illustrates the effects of various estimation techniques on the reserve estimates reported to regulators and investors. The studies concentrated primarily on reserving techniques, however, and made little mention of the effects of misestimation. Forbes (1969) examined auto liability reserving techniques for a sample of insurers, suggested simulation to aid in selection of optimal reserving techniques, and discussed the impact of reserving on financial statement valuations. Forbes (1970) also studied the regulation of loss reserving and found the regulation in need of reform. Anderson (1973) analyzed the effects of reserve errors on policyholders' surplus and found that reserve errors have a stabilizing effect on underwriting income. Balcarek (1975) and Ansley (1979) further related inadequate loss reserving to dismal underwriting experience in the 1970s. Ansley hypothesized that inflation was one of the major reasons for inadequate reserves in the 1970s. The investigation of reserve errors was furthered by Smith (1980), who tested whether insurers manage loss reserves to smooth underwriting results. Smith found that the incidence of overreserving and underreserving errors was not random for a sample of property-liability insurers in the auto liability line. Weiss (1985) continued the studies of Smith for a sample of insurers underwriting auto liability policies. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 110 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.