Printer Friendly

The transactions cost theory of insurance: contracting impediments and costs.

The Transactions Cost Theory of Insurance: Contracting Impediments and Costs


It is well known that the "insurance-as-pooling-of-risks" paradigm, valuable as it is, cannot explain all demand for insurance. In the risk-management literature, for instance, it is often argued that large profit-maximizing firms should self-insure minor losses in order to avoid the loading costs of insurance. However, property and liabilities of relatively low value are often insured. Another phenomenon that is difficult to understand as pure risk pooling is the purchase of insurance policies that cover the replacement of buildings when replacement costs exceed expected flows of future returns (Doherty 1985, p. 277). Individuals also tend to over-insure. With loading charges of 20 to 40 percent of premiums, one would expect less insurance and larger deductibles than observed (Friedman, 1974, Pashingian, Schkade, and Menefee 1966, and Stuart 1983).

Smith and Warner (1979), Mayers and Smith (1981, 1982, 1987), and Main (1982, 1983) argue that individuals in frictionless capital markets would adjust their portfolios so that there would be no demand for a resource-consuming insurance industry. They also argue that with well-functioning capital markets, insurers would have no obvious comparative advantage over banks or corporate firms in diversifying risks. Following Modigliani and Miller (1958), Mayers and Smith (1987) argue that, if corporate insurance purchases affect the value of the firm, they must do so via taxes, contracting costs or the impact of financial policy on the firm's investment decision. Mayers and Smith (1987) conclude that the corporate insurance purchases are motivated by: low-cost claims administration services provided by insurers, assistance by insurers in assessing the value of safety and maintenance projects, improvements in the incentives to undertake investments in safety and maintenance projects resulting from insurance, increased efficiency in the allocation of risk-bearing among the corporations' claimholders brought about by insurance, and a reduction in the firm's expected tax liability.

This article is based on these seminal contributions to a new theory of insurance. If focuses on the contracting costs problem, which appears to be fundamental to an understanding of insurance. It is argued that insurance may be purchased by all agents (including widely held firms and risk neutral individuals) because of the impossibility of complete contingent claim contracts. The theory is outlined first, several examples are discussed, then conclusions are drawn.

Insurance and the Costs of Contingent-Claim Contracting

The general proposition here is that the insurance industry supplies services that reduce transaction costs in trade. Traders faced with the problems of contracting realize that a complete contingent claim contract that regulates all possible outcomes is infeasible because of the costs of identifying all events, negotiating, pricing, documenting that contracted liabilities have occurred, and enforcing the contract. The presence of these costs means that contingent-claims contracts are subject to a credibility problem: the parties cannot take for granted that the obligations of the contract will be fulfilled.

Contingencies that are too costly to regulate contractually may be insured. When the risk is insured, it is sold (at a negative price) to a third party, the insurer. (1) To illustrate, assume that a manufacturer and a shipper enter into a contract for the transport of the manufacturer's machinery. Their contract specifies the price and date of delivery as well as some other basis details. However, the contract does not cover contingencies such as fire, storm, and explosion risks because they have too little knowledge of the risks to be able to price liability and precautions and because the liable party may not be financially able to fulfill the agreement if a serious accident occurs. Instead they add a clause stating that the cargo must be insured during transportation. Accordingly, a whole package of contingencies is transferred to an insurer.

The insurance industry specializes in writing contracts on insurable risks such as potential losses due to fires, storms, traffic accidents, and third-party liability. A general characteristics of insurable exposures is that a large number of specific risky events can arise but the probability that a specific event will occur is very low. Low probabilities may give a specialized insurance firm a comparative advantage in writing contracts and in dealing with realized accidents. (2) The insurer obtains information on the actuarial relation between damages, the presence of safety devices, level of deductibles, co-insurance, and costs of various claim settlement procedures (Mayers and Smith, 1981). Thereafter, the insurer may offer different policies at premiums that vary with actuarially expected costs. An important property of the insurance contract is that it gives the insurer latitude to adjust or withhold claims in accordance with contractual provisions.

An insurance policy contains a list of restrictions and limitations that specify the liabilities of the insurer and the involved parties. For instance, a shipping policy may state explicitly that the insurer is liable except when the shipped goods are improperly packed, which would make the packing firm liable, or when the shipper is negligent, which would make the shipper liable. Hence the services of the insurance industry include writing, pricing, and enforcing liabilities as defined in the policies. These are largely legal functions.

Suppose a risk-management expert specializing in low-probability events were to advise the parties on how to draw up a contract and also assisted with safety projects and claim adjustments. Would this form of market organization work as well as insurance? Because of the low probability of each specific event, the credibility of information is poor. Consequently, information on the value of the risk manager's advice is unreliable. This implies that there is a principal-agent problem whereby risk managers may shirk. Of course the principal-agent problem might be reduced by transferring the risk to the risk-manager who would then also receive an ex ante payment covering expected claims and administrative costs. Such a contract that makes the risk-manager the residual claimant is in essence an insurance contract. Mayers and Smith (1982, p. 288) give similar reasons why insurance is preferred in favor of services by independent consultants.

It is interesting to note the similarity between this approach and the theory of the firm presented by Alchian and Demsetz (1972). The risktakers (the insurer and the entrepreneur) have similar functions and incentives. They monitor the policyowner and the employee respectively, and they maintain the right to interpret the contract (to settle claims or to hire and dismiss workers). A prerequisite for insurance (employment) is that the policyowner trusts the insurer (the employer). To be credible, the insurer (employer) must appear solvent, and have a long-term interest in fulfilling contractual obligations. Due to expected participation in the market the proper treatment of policyowners (employees) becomes important, and contributes to the policyowner's (employee's) trust in the verdict.

The inseparability of credibility, claim adjustments, and the pricing of liabilities thus explains why these services are sold together in insurance packages. It also explains why the insurance industry is an expert on claims administration, safety, and maintenance projects.

So far it has been assumed that the initial problem is pricing liability and safety measures. The credibility problem, however, can equally well be used as the starting point. Assume for instance, that the shipper is willing to accept liability for property damages during transportation, but that the manufacturer questions the credibility of the shipper and requires some form of security. For the sake of simplicity, suppose there are two options: a bank that issues a guarantee or insurance. Which will be preferred? The comparative advantage of a guarantor is due to its credibility, its ability to price the risk accepted, and its ability to limit moral hazard. The solvent insurer with long experience of property and liability damages thus has a comparative advantage in the coverage of such risks. Similarly, the bank with substantial reserves and experience in the evaluation of credit risks obtains a comparative advantage in insuring risks of bankruptcy. The manufacturer may therefore require both a guarantee from the bank in case of the shipper's bankruptcy, and transportation insurance.

There are several interesting differences between the pooling-of-risks theory and the transaction costs theory. First, in the pool-of-risks theory, moral hazard is treated as a complication. In the transaction-cost theory, moral hazard is the raison d'etre for insurance. In other words, if risk sharing had no impact on the controls taken, any credible risk-neutral party would be able to bear the risk for a premium equal to the expected accident cost plus loading, and there would be no demand for insurance as a regulator of levels of control. Second, in the transaction costs theory, economies of scale of the insurance industry are not attributable to the presence of large pool for the spreading of risks. The comparative advantage of the insurance industry is due to its specialization in writing and enforcing contracts on low probability events. However, the size of the insurer's assets is important for its credibility. Third, while the transaction costs theory of insurance explains why risk-neutral traders may purchase insurance policies to reduce contracting costs, the pooling-of-risks theory explains why diversification is demanded by the risk-averse. The theories are thus complementary rather than competitive; insurance both eliminates risk by pooling and reduces transaction costs.

Further Applications

Thus far, the convenient example of transportation insurance has been treated. The theory can also be applied to many similar contractual situations. For instance, consider a mortgage contract in which the borrower and a bank agree on interest rate, loan initiation fees, and a building as collateral. Low-probability contingencies such as liability if the building is damaged by fire are not included in the mortgage contract. The bank might cover such perils by adding a risk premium to the interest rate. Such unconditional coverage, however, is normally inefficient because of possible moral hazard. Requirements for safety-devices and other precautions might then also be included in the contract. The bank may therefore consult a specialist or trade the risk with an insurer. The principal-agent problem at low probability events like fire, storm, and water damages explains why such perils are usually insured.

In addition to simplifying contracts between the two primary trading parties, insurance may be beneficial to parties other than the two primary ones. Continuing the mortgage example, assume that the building is an apartment complex and that the borrower (landlord) purchases insurance against fires and other perils. Tenant-landlord contracts may then also be simplified. It is costly for them to contract on liability in relation to fire, water damage, and other causes of loss, and if they neither contract on the risks nor insure, there is potential for conflicts ex post of accidents.

To emphasize the potential importance of these gains in the form of contract facilitation, consider the fact that homeowners purchase homeowners insurance with premiums that exceed expected actuarial losses by 30 percent or more. Several studies suggest that deductibles are too low given reasonable assumptions about the level of risk aversion [Friedman (1974), Pashigian, Schkade, and Menefee (1966), and Stuart (1983)]. It is hardly possible to explain this demand solely by reference to the pooling-of-risks theory. One reason for the over-insurance may be the additional benefit that insurance facilities trade with banks, tenants, and others who may have claims in the event of accidents. Assume, for example, that a $100,000 house loan from a bank entails a yearly interest of 10 percent and an insurance premium of $200, including a $60 loading charge. Although 30 percent may sound like a high loading rate for diversification alone, the resulting loading may be small compared with the combined benefit of risk diversification plus benefits in obtaining contracts with the bank and others.

The transaction cost motive for insurance also provides an explanation for observed purchases of insurance that cover replacement of factories when replacement costs exceed the expected flows of future returns (Doherty, 1985). In a world with no contracting costs (including enforcement costs), the owner(s) would sign complete contingent claim contracts with bondholders, workers, customers, and suppliers. Because of moving costs and firm-specific capital, these contracts would include payments to workers (damages) in the event of factory closure. Confronted with a replacement decision, the firm would then have to compare the expected return from replacement with the damages paid if no replacement were made. If damages were sufficiently high, replacement would be optimal even if the investment cost exceeded the expected flow of future returns. In reality, of course, comprehensive contracts are costly. An alternative contracting technique that can lead to the same outcome as under a comprehensive contract is for the firm to purchase replacement insurance. The insurance policy reduces contracting costs and enforcement problems that would arise because of potential conflicts between shareholders who have no incentive to reinvest in a new factory and other claimholders.

Further evidence supports the view that transaction costs are a central motive for insurance. Consider the details of actual insurance policies; most of the text of an ordinary property and liability policy consists of a list of restrictions, conditions, and limitations that define the liability of the insurer and the involved parties under a large number of contingencies. These conditions are similar to conditions that might have been included in the trader's contract, if the traders had themselves contracted on these contingencies. Moreover, contractual clauses frequently require trading parties to purchase insurance. Shipping contracts and standard loan contracts normally include a clause requiring that properly and liability is insured. The same is true of rental and lease contracts. Covenants are commonly attached to bonds that require the issuing corporation to purchase insurance; thus instead of writing a contingent-claim contract that includes all precautions required to make the bond a safe asset, an insurance policy is purchased against a subset of contingencies. (3) Insurance clauses have a long tradition in the construction industry (Bunni, 1986). Franchisors often require explicitly that franchisees are insured. Perhaps more tellingly, these requirements appear to be independent of whether the insured party is risk-averse--insurance is universally required to reduce costs related to contracting, control and conflicts.


The demand for insurance by large, widely held, firms cannot be satisfactorily explained by the pooling-of-risks theory of insurance. Consumers also tend to over-insure. Here it is argues that traders sell (at a negative price) a subset of the contingencies to an insurer because of the transaction costs. The advantage of transfering certain risks to an insurance firm is due to the firm's specialization in writing contracts for low probability risks and to its credibility. Credibility depends on the firm's experience, its funds and its expected future in the market. The combined need of credibility and special skills in pricing risks and in claim adjustments gives the insurer a comparative advantage relative to riskmanagers that supply separate riskmanagement services.

Qualitative evidence supports the theory. An ordinary transportation or property and liability insurance policy is a standard form contract for low probability risks. An insurance clause is a common supplement to traders' contracts--a requirement independent of the parties' risk-aversion.

(1) A risk may also be left without consideration. It such a risk is realized the liability must be settled ex post, e.G. by a public court. Hence insurance is also an alternative to costly litigation, (Skogh, 1989).

(2) A number of empirical studies show that the minimum efficient scale in handling the insurable risks is relatively large, see e.g. Allen (1974), Cummins (1977), and Skogh (1982).

(3) Smith and Warner (1979) interpret insurance clauses as a mean of reducing agency costs between bondholders and equityholders.


[1] Alchian, Armen A. and Demsetz, Harold M., 1972, Production, Information Costs and Economic Organization, The American Economic Review, 62: 777-95.

[2] Allen, R.F., 1974, Cross Sectional Estimates of Cost Economies in Stock Property-Liabibility Companies, The Review of Economics and Statistics, 56: 100-103.

[3] Bunni, N.G., 1986, Construction Insurance (London: Elsevier Applied Science Publishers LTD).

[4] Cummins, David J., 1977, Economies of Scale in Independent Insurance Agencies, The Journal of Risk and Insurance, 44: 539-53.

[5] Doherty, Neil, 1985, Corporate Risk Management. A Financial Exposition. (New York: McGraw Hill).

[6] Friedman, B., 1974, Risk Aversion and the Consumer Choice of Health Insurance Option, Review of Economics and Statistics, 56: 209-14.

[7] Main, Brian G., 1982, Business Insurance and Large, Widely-held Corporations, The Geneva Papers on Risk and Insurance, 7:237-47.

[8] Main, Brian G., 1983, Why Large Corporations Purchase Property-Liability Insurance, California Management Review, 25: 84-95.

[9] Mayers, David & Smith, Clifford W., 1981, Contractual Provisions, Organizational Structure, and Conflict Control in Insurance Markets, Journal of Business, 54: 407-34.

[10] Mayers, David & Smith, Cliford W., 1982, On the Corporate Demand for Insurance, Journal of Business, 55: 281-95.

[11] Mayers, David & Smith, Clifford W., 1987, Corporate Insurance and the Underinvestment Problem, The Journal of Risk and Insurance, 54: 45-54.

[12] Modigliani, Franco and Miller Merton, H. 1958, The Costs of Capital, Corporation Finance and the Theory of Investment American Economic Review, 48: 333-91.

[13] Pashigian, B. P., Schkade. L. L. and Menefee G. H., 1966, The Selection of an Optimal Deductible for a Given Insurance Policy, Journal of Business 39: 35-44.

[14] Skogh, Goran, 1982, Returns to Scale in the Swedish Property-Liability Insurance Industry, The Journal of Risk and Insurance, 49: 218-28.

[15] Skogh, Goran, 1989, Contracts, Insurance and Litigation. Departments of Economics, University of Lund. Memo.

[16] Smith, Clifford W. and Warner, J. B., 1979, On Financial Contracting: An Analysis of Bond Covenants, Journal of Financial Economics, 7: 117-61.

[17] Stuart, Charles E., 1983, Pareto-Optimal Deductions in Property-Liability Insurance: The Case of Homeowner Insurance in Sweden, Scandinavian Actuarial Review, 227-238.

Goran Skogh is Associate Professor of Economics at Lund University, Sweden.

The author gratefully acknowledges the financial support of the Forsakringsbolaget Pensionsgaranti (FPG). Thanks are due to Karl Borch, Charles Stuart, and Emilio Venezian for comments on an earlier version of the study.
COPYRIGHT 1989 American Risk and Insurance Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Skogh, Goran
Publication:Journal of Risk and Insurance
Date:Dec 1, 1989
Previous Article:Tax reform and retirement income replacement ratios.
Next Article:An empirical evaluation of two methods for estimating economic damages.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters