Evolution of Insurance: 19th Century Liability to Modern Risk Pooling - Prof. J. Boyd, Study notes of Credit and Risk Management

Download Evolution of Insurance: 19th Century Liability to Modern Risk Pooling - Prof. J. Boyd and more Study notes Credit and Risk Management in PDF only on Docsity! Insurance Industry Origins Insurance in some form is as old as historical society. Bottomry contracts were known to merchants of Babylon as early as 4000-3000 BC. Bottomry was also practiced by the Hindus in 600 BC and was well understood in ancient Greece as early as the 4th century BC. Under a bottomry contract, loans were granted to merchants with the provision that if their shipment was lost at sea, the loan did not have to be repaid. A higher interest rate was charged on such loans to cover the risk, and the additional interest was referred to as a “premium”. Ancient Roman law recognized the bottomry contract in which an article of agreement was drawn up and funds were deposited with a moneychanger. In Rome there were also burial societies that paid funeral costs of their members out of monthly dues. In the insurance contract of ancient Greece and other maritime nations in commercial contact with Greece, a ship owner or merchant who wanted protection would circulate a contract document outlining specifics of the voyage (type ship, cargo, and other facts) to those who wished to bear part of the risk and sign their names under the agreement (“underwriting”, the term has the same meaning today: one who accepts and rejects risks). In early-England, a convenient place for mariners and underwriters to meet in were coffee houses in London where the firm Lloyds of London originated. United States Benjamin Franklin organized the first American insurance company in 1752 as the Philadelphia Contributionship. The first life insurance company in the American colonies was the Presbyterian Minister’s Fund, organized in 1759. By 1820 there were 17 stock life insurance companies in the state of New York alone. Many of the early property insurance companies failed from speculative investments, poor management, and inadequate distribution systems. Others failed after the Great Chicago Fire in 1871 and the San Francisco earthquake and fire of 1906. There was little effective regulation, and rate making was difficult in the absence of cooperative development of sound statistics. Many problems also beset the life insurance business. In the era following the U. S. Civil War, bad practices developed: dividends were declared that had not been earned, reserves were inadequate, advertising claims were exaggerated, and office buildings were erected that sometimes cost more than the total assets of the companies. Thirty-three life insurance companies failed between 1870 and 1872, and another 48 failed between 1873 and 1877. In the late 19th century (beginning in the 1880s) in the U.S., increased frequency and severity of injuries and death associated with the intercontinental railroad and the industrial revolution led to a shift in legal and public sentiment toward the development of liability insurance as a means to compensate victims of negligence. Until the 1880s, liability insurance was nonexistent in both the U.S. and England because it was deemed to be against public policy, a way to avoid moral responsibility since negligent acts were indemnified. Since the early 20th century all major forms of insurance have enjoyed a steady growth in the United States. Worldwide Operations Today Because of the great expansion in world trade and the extent to which business firms have made investments outside their home countries, the market for insurance on a worldwide scale expanded rapidly in the 20th century. This development has required a worldwide network of offices to provide brokerage services, underwriting assistance, claims service, and other services. The majority of the world’s insurance businesses as of the end of 2010 were concentrated in North America at 30% (U.S. 27%, Canada, 3%) and various countries in Western Europe (U.K., France, Germany, Italy, and the Netherlands) at 25%, and the companies located there serviced a large part of the insurance needs of the rest of the world. Insurance Industry An insurance policy is a contract that binds an insurer to indemnify a client against a specified loss in exchange for the payment of a set premium. We may classify the private insurance industry by product into three segments: life insurance, health insurance, and property and liability insurance. Life insurance covers the financial risk of a policyholder’s death, and of superannuation (living too long). Under the simplest life insurance arrangement, a policyholder pays a premium to an insurer that provides death coverage for a finite length of time. More sophisticated forms of life insurance also serve as vehicles for personal savings, combining an investment with life insurance coverage. Health insurance products are designed for loss by sickness or injury (medical expense insurance), and for periodic payments to the insured who is unable to work because of sickness or injury (disability insurance). Property and liability (casualty) insurance products provide financial protection for property and for legal liability. Individuals, for example, frequently purchase insurance policies to protect against damage to their cars and homes. Similarly, businesses buy insurance for damage from fire and theft, as well as for unforeseen liabilities such as lawsuits. Social insurance, such as worker’s compensation, social security, and unemployment compensation are forms of insurance at the state or federal level to provide minimal levels of protection for people who may otherwise be unable to cope with fundamental losses. Social insurance is usually compulsory by law, requires contributions to be eligible for benefits, and is supervised by the government. In the United States at the end of 2009, there were approximately 1,106 life/health insurance companies and 2,737 property/casualty companies which generated total premiums of approximately $0.93 trillion dollars, with 55% of the premiums for life/health insurance, and 45% of the premiums for property and casualty insurance. The ownership of U.S. private insurance market (and the global private insurance market) is highly fragmented within each of the three segments of private insurance (no firm controls a high market share), although there is an overall trend of consolidation within the financial services industry, including insurance, which suggests a continued decline in the overall number of insurance companies in the future. According to the Insurance Services Office (ISO), concentration in the P/C insurance sector as measured by the Herfindahl-Hirschman Index increased from 229 in 1980 to 357 in 2008, and then dropped to 351 in 2009. The U.S. Department of Justice classifies any score under 1,000 as unconcentrated, while any score over 1,800 is concentrated. Studies by scholars over the last 20 years have concluded that the insurance industry is highly competitive, with no evidence of excessive profits, and that the rate of returns to firms within the industry has tended to be less than for other comparable industries with similar risk (Industry ROE’s for 2010 were approximately 7.5% for the Property/Casualty sector, 12.0% for the Health Insurance/Managed Care sector, and 5-8% for the Life Insurance/Annuity sector. The insurance business involves four distinct activities: 1) ratemaking and underwriting, 2) marketing and sales, 3) claims adjustment, and 4) asset management: 1) Ratemaking is performed by the actuarial department in large life insurance companies, and in smaller companies, by an actuarial consulting firm. In property and liability insurance, rates are determined from loss statistics provided by an advisory organization or accumulated by the individual insurer. A rate is the price charge for each unit of protection (premium = rate X units of protection purchased), which in life insurance, for example is $1,000 per increment. The resulting premium income charged to an insured must be sufficient to cover losses and expenses. Underwriters evaluate applications for insurance against company underwriting requirements in deciding whether to accept or reject the applicant. The underwriter usually consults the application itself, information from the agent or broker, credit reports, information bureaus, and physical examinations or inspections in determining whether to accept/reject an applicant. If no homeowners insurance (fire insurance) is available, each homeowner faces the following probabilistic payoff: Probability $Payoff (loss) 0.1 100,000 0.9 0 The expected loss for each home is 0.1 X 100, 000 + 0.9 X 0 = $10,000. The risk of this payoff, as measured by the standard deviation, is $30,000= [0.1 X (100,000-10,000)2+ 0.9 X (0- 10,000)2]1/2 In the event of a fire, however, an uninsured individual homeowner would not bear the expected loss, but rather the actual loss of zero or $100,000. Thus, if a homeowner wanted to be absolutely sure that she was protected against the financial hardships of a house fire, $100,000 of upfront capital would be required. If no fire insurance were available, this homeowner would have to retain the 10% chance that a catastrophe would occur. Now suppose that two homeowners enter into an agreement to split their total losses. Each will swap one half of the payoff from his/her house for one half of the payoff of the other’s house. If we assume the houses are not in close proximity to one another, then the probability of one house catching fire is independent of the other. Each of the two owners now faces the following new probabilistic payoff: Probability $Payoff (loss) 0.01 100,000 0.18 50,000 0.81 0 Because the probability of one house burning is independent of the other burning, the probability of incurring a loss of $100,000 (both houses burn) is 0.1 X 0.1 = 0 .01. Similarly, the probability that no loss will be sustained is 0.9 X 0.9 = 0.81. The expected loss remains at $10,000. However, the probabilities of the extreme outcomes--losing $100,000 or losing nothing--have decreased, and the standard deviation of this scenario, $21,213, is smaller than the one house standard deviation of $30,000. Intuitively you might suspect that if you add more and more people to the above two-person risk pool, the standard deviation will continue to decrease. This intuition is correct and can be confirmed from statistical laws. Assume there are N homeowners who pool their risks. Each swaps 1/Nth of her loss for 1/Nth of the loss from each of the other homeowners. Each homeowner still faces an expected loss of $10,000, but the uncertainty of this loss as measured by the standard deviation is inversely proportional to the square root of the number of items in the sample, or, $30,000/N1/2, or $21,213 for example above for N=2 People This result, you may recall, stems from the central limit theorem of statistics that indicates that the larger a sample, the more closely the sample mean approaches the underlying population mean and the smaller the margin of error becomes (intuition: the bigger the sample, the closer the sample mean will be to the population mean). This is good news for the risk averse individual, because this mathematical reasoning suggests that each member of a large group of people should only have to pay the $10,000 average for fire insurance and be confident that this will be sufficient to cover the housing fire losses of the group. It is important to note that insurance does not reduce the magnitude of losses (for N homes in the risk pool, the total losses on average will be 10% X N X $100,000), the probability of occurrence at 10% (excluding any moral or morale hazard effects), nor the uncertainty for an individual as to whether or not the event will occur (his/her house may among the 10% of homes in the risk pool that burn down). It does, however, eliminate the homeowner’s uncertainty of financial loss, because the purchase of the appropriate level of insurance, $10,000 in the above scenario, is the extent of financial loss that the homeowner will incur in the event that his house burns down (again, assuming no deductibles). Capitalization Requirements Looking at the same scenario above from the vantage point of the insurer, the key question is how much capital is needed to underwrite the risks that the firm assumes. In our house example above, to achieve absolute certainty that the company as a whole will be able to meet all possible claims against it would require a total capitalization of $100,000 X N, since there is a non-zero chance that all N homes could burn. However, if the group is willing to accept the chance that it may not be able to fully cover aggregate losses, the capital cushion does become a decreasing function of N. Consider, for example how to fund a mutual insurance company (i.e., an insurance firm owned solely by its policyholders and thus having no capital stock) such that it could fully cover all potential claims against it with a probability of 0.999999 (the company will fail one in a million times). This would require a maximum premium per homeowner of $10,000 (losses expected) + [4.75(standard deviations) X $30,000]/N1/2 (i.e., we know the population mean =$10,000, so the upper limit of the confidence interval will be $10,000 plus 4.75 standard deviations (population standard deviations) divided by the square root of the number of homeowners in the pool, “N”. Note, of course in practice you don’t necessarily know either the population mean or the population standard deviation, so a variation of the method above is used) A table showing the premiums that each homeowner would have to pay to fully fund a mutual insurance company of N homeowners, such that the firm could cover all possible claims with 0.999999 confidence is shown below; this table also illustrates the economic efficiencies to be gained by the pooling of risks. Notice that as N gets large, the process of diversification leads to the funding of a virtually riskless insurance company with premiums, that in the limit, approach their actuarially fair value, the value of expected loss ($10,000). N $ Premium for .999999 confidence 1 100,000 10 55,062 100 24,250 1,000 14,506 10,000 11,425 100,000 10,451 1,000,000 10,143 (This table assumes the probability of complete loss is 0.1 and the probability of no loss is 0.9) The statistical principles underlying the results above (central limit theorem) are very important to the insurer because insurance companies insure samples of populations, not entire populations. Knowing first, that if you take a large enough sample, regardless of the distribution of the underlying population, that the resulting sample mean distribution will be normal, and second, that the sample mean reliably estimates the population mean as more insured customers are added to the risk pool (customer base) are important principles used by insurance actuaries to set insurance premiums. The valuable service of risk pooling and risk transfer offered to society by insurance companies also creates another major advantage for society, optimizing the use of capital. Recall last class that we spoke of options available to individuals and businesses in the management of loss exposures: avoidance, retention, and transfer. If insurance were not available, the loss forecasting precision developed by insurance companies over many years and exemplified by the chart above would not be available to individuals and businesses (non-insurance), therefore they would have to crudely estimate and retain capital reserves (liquid securities or cash) to meet future loss exposures rather than invest them in their line of business. Insurable Risks We concluded last class, using a frequency/severity risk management framework, that insurance is most appropriate for high-severity/low-frequency events. We can build further on this general statement and identify four requirements that insurance companies seek in an insurable risk: 1. A sufficiently large number of similar (homogeneous) exposure units to make the losses reasonably predictable. (This is a key criterion for objective statistical analysis) 2. Losses produced by the risk must be definite and measurable. (Any other criteria is subjective, not objective, and both the existence and severity of a loss may be open to dispute) 3. Losses must be chance or accidental (Moral hazard or the intent to cause a loss is an old concern in insurance, dating back to the origins of marine, fire, and life insurance; intentional loss is not covered by an insurance contract) 4. Loss must not be catastrophic. (Numerous insurance companies have learned this principle, after the fact, to their detriment (bankruptcy). Diversification is the key to this principle, so that you don’t sustain a high frequency/high severity event. Example: An insurance company’s customers’ insured properties are all in Florida or New Orleans where Hurricane Andrew or Katrina hits and destroys them all!) Insurance Industry Financials In collecting premiums and holding funds to pay future losses, insurance companies act as investment intermediaries. The assets that an insurer has available to invest are derived primarily from the following: loss reserves, unearned premium reserves, and policyholders’ surplus (See next page for discussion of these terms). The table below is typical of the listings typically found on an insurer’s balance sheet. Assets Bonds $XX Liabilities and Net Worth Common stocks XX Unearned premium reserve $XX Preferred stocks XX Loss and loss expense reserve XX Real estate XX Reserve for taxes and expenses XX Cash XX Reserve for all other liabilities XX Premium balances receivable XX Total Liabilities XX Other assets XX Surplus (including retained earnings) XX ___ Total Assets $XX Total Liabilities and Net Worth $XX For the year 2010, actual proportions of financial assets of the life/health insurance industry and the property/casualty industry in the U.S. were: Life/Health Property-Liability Corporate/Foreign Bonds 39% 23% Government. /Municipal Bonds 13 40 Corporate Equities 27 16 Other 21 21 Total 100% 100% Note above that life insurers invest primarily in corporate bonds and equities (66% for 2009) which are consistent with the long term nature of their expected payouts for mortality. The property/casualty insurers, on the other hand, invest largely in high quality securities such as bonds (63% for 2009) which can be sold quickly to pay for claims resulting from major hurricanes or other catastrophes.