The Design of Insurance Contracts When Liability Rules Are Unstable

Abstract

The Design of Insurance Contracts When Liability Rules are Unstable ABSTRACT Recent troubles in liability insurance markets have been associated with a destabilization of liability rules. A nautral policy response is seen in proposals for legislative reform of liability rules with the aim of increased stability. An alternative track explored here examines the design of insurance contracts. The effects of judicial and legislative process on the insurance pooling function are examined in a rational expectations setting. It is then shown that the conditions for closure of policies characterized as premium are weaker than for traditional contracts given instability of liability rules. Finally, it is shown that the insurance market has responded to recent crises by organizational and contractual innovations that embody the random premium design. These responses include the emergence of mutual like forms of organization and the policy. Baruch Berliner, 1982, outlined the ideal conditions for the functioning of an insurance market. Amongst other conditions is a requirement that there be a large number of independent loss exposures. This condition ensures the operation of the Central Limit Theorem and the insurer's risk will essentially be diversified away. Failure to satisfy this condition appears to be at the heart of many of the recent troubles in liability insurance markets. For example, the recent liability insurance crisis in the United States appears to be a response to a destabilization of the legal system.(1) Insurers argue that they are able to insure the liabilities of clients arising under an unchanging set of liability rules, but they cannot insure against changes in the rules themselves. Liability insurance markets have responded with different designs of insurance or risk bearing contracts and with changes in the forms of organizations that bear risk. For example, liability insurers have introduced the so called contract which, in effect, shortens the period over which changes in the level of expected loss may commonly affect all outstanding policies. A more dramatic change has been the shift from stock forms of insurance organization to mutuals, group captives, industry pools and reciprocals. These structures share the risk among a group of firms (or individuals) commonly exposed. Equity in the structure is held by the policyholders. Such structures are referred to in the current article as mutuals. These changes in contract or organizational design have a similar effect. The premium for any given period of cover is random. It is subject to retroactive adjustment on the basis of new information concerning the aggregate loss in the pool. For example, the mutual may pay a dividend (positive or negative) to its policyholders which is related to the aggregate loss in the pool. The policy holder buying a claims made policy will find that losses which may have arisen, but which have not been presented as claims, within the policy year will be priced in the future in a future insurance contract. The policyholder does not know how much he or she will have to pay to insure this year's liabilities. The premium is, in effect, random. The next section of the article describes how correlations arising both from the incidence of losses and the generation of new information can lead to undiversified insurance portfolios. This section also shows how judicial and legislative processes can lead to such problems. A subsequent section presents a formal analysis of optimal insurance purchase with random premiums showing the conditions under which random premium contracts are preferred. The final section addresses the contractual and organizational innovations which have emerged in theses troubled markets. The insurance markets described are mainly for corporate insurance buyers. …

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 — 65 citations exist upstream.

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