Terrorism

A Perspective on some Economic Issues i= n the U.S. Insurance Market

Richard D. MacMin= n
Edmondson-Miller Chair
Katie School
Illinois State University
Normal, IL 61790=
United States of America
Phone: +1 309 438-7993
mailto:richard.macminn@ilstu.edu<= /a>

and

Patrick L. = Brockett
Gus Wortham Chair in Risk Management
MSIS = Department
College of Business
University of Texas
Austin, Texas 78712=
United States of America
Phone: +1 512 471-6816
mailto:brockett@uts.cc.utexas= .edu

December 2004

This article presents the personal views of the authors delineating prom= inent economic issues facing the U.= S. insurance market. It is divid= ed into two parts, risks involving mortality and morbidity, and risks involving non-life markets. Withi= n the life arena, the most pressing current economic issues in the U.S. involve social insurance issues, i.e.= , longevity risk implications for social insurance programs and issues resulting from t= he U.S. sy= stem of health care insurance involving private rather than public supply of coverage. In the non-life are= na, terrorism risk raises economic issues of availability, affordability and de= bate concerning whether the private insurance market can provide an appropriate market, or whether governmental involvement or even assumption will be economically necessary. A fin= al economic topic addresses the very controversial problem of how far private insurers can go to implement a new classification variable which, while act= uarially related to the risk, is not widely comprehended as physically or economical= ly related. The topic here is in= surers’ use of the credit history of the insured to predict their automobile and homeowners losses.

<= /o:p>

Longevity Risk=

It is typica= lly said there are three pillars used in protecting against longevity risk in retire= ment. The first is a tax financed govern= ment safety net, i.e., Social Security in the U.S.. The second is savings to private accounts by workers or employers (pensions) and the third is voluntary savi= ngs by the individual. Social Sec= urity has been financed on a pay-as-you-go basis and provides benefits for the duration of retiree lives. It= is adjusted for cost of living and occasionally the tax and benefit rates are adjusted. The last reforms we= re enacted in 1983 and the expectation then was that more changes would have t= o be made in twenty years, i.e., now.

As mortality rates improve and population demographics change there is increasing fiscal risk for social insurance programs like Social Security. The fiscal r= isk is due in part to the pay-as-you-go arrangements established for the programs, which, combined with aging populations, is putting pressure on the government. The mortality improvement is also putting pressure on individuals since the specter of outliving their resources, i.e., longevity risk, becomes ever more probable= .

The second p= illar for many is an employer sponsored defined benefits (DB) or defined contribu= tion (DC) plan. The contributions by employer and employee are the same for DB a= nd DC plans, but DB plans provide an indefinite payment to the individual upon retirement while DC plans provide the aggregated wealth of the portfolio ch= osen by the individual. Hence, DB = plans cover longevity risk of the individual by transferring this risk to the employer, while longevity risk under DC plans is borne by the employees. Because of increasing governmental regulations and economic costs, employers are increasingly moving away from= DB plans to DC plans, shifting longevity risk to employees. There are economic implications of= this shift for total social adequacy of retirement amounts since economic resear= ch has shown that individuals invest more conservatively (with correspondingly lower returns) than do institutions (such as DB pension funds), resulting in the same contributions producing less ultimate income for retirees. Thus, this shift weakens the suppo= rt of this pillar.

The third pi= llar is private savings, e.g. an individual retirement account. Historically, U.S. citizens tend to have the lowest individual savings rates in the industrialized world, making this a = very weak pillar to rest upon.

Financial ma= rkets do provide governments and individuals a means of transferring risks and resou= rces across time as well as allocating risk among individuals. The government does allow the soci= al security administration to transfer money forward via a trust fund when it = is running a surplus. However, according= to the trustees of the Social Security Administration that surplus will turn i= nto a deficit in 2018 and the trust fund will be exhausted by 2042. Moreover, economically the trust f= und is not a financial asset; it is a financial obligation of the Treasury, and to= be honored the obligation requires the government to either increase taxes, re= duce federal expenditures or issue new debt. Since the actual assets in the trust fund have already been spent, the call for paying benefits out of the trust fund will require “cashing in” Treasury notes, which is done by using tax revenues since these are the only source of revenue for the government. Thus, the enormous burden of paying social security has been transferred to the future, a resu= lt which may cause great political and social upheaval unless addressed soon.<= span style=3D'mso-spacerun:yes'> It has been estimated that to meet current benefit promises in the future will require enormous average tax ra= tes unlike any previously encountered.

Similarly, i= n forming the next two pillars, corporations and individuals do transfer money forward via security purchases to fund the retirement accounts. The instruments available, however= , do not allow corporations or individuals to easily hedge the longevity risk th= at they face. Longevity risk in conju= nction with interest rate risk has created problems for the annuity market. The immediate annuity market in th= e U.S. is approximately two billion dollars per year. As more and more baby boomers reti= re, annuity markets will grow as will the risk and consequences of underestimat= ing mortality improvements. New medical technology created as a result of discoveries made in the links between genetics and disease has the potentia= l to create those mortality improvements. The private sector pillars of the pension system are potentially at = risk without better hedging instruments.

The Bush administration is currently proposing changes in the social security that w= ould partially move the first pillar from pay-as-you-go to self-funded private accounts. The motivation is t= hat the funds in private accounts would be invested in the financial markets and generate greater rates of return. Even ignoring the substantial deficits that would have to be run to = make the transition to such a scheme, the plan ignores the shift in longevity ri= sk from government to individual. One unacknowledged question here is who is best suited to bear the longevity ri= sk, the government or the individual.

Financial in= struments can be constructed to alleviate the problems faced by corporations and individuals. A survivor bond = is one example of such an instrument. The survivor bond has a coupon that is proportional to the number of survivors = in a selected birth cohort; letting the cohort be the number of individuals turn= ing sixty-five in the year that the bond is issued, the coupon the following ye= ar would be proportional to the number in the cohort that survive to this year. The survivor bond payoff matches the liability of the social security administration and would be closely correlated to the liabilities of private DB pension schemes. The existence of a publicly held a= nd traded survivor bond would allow the market creation of an effective hedge against longevity risk; it would be a boon to the DB plans because of the hedging potential and further, it would help create greater scale in the annuity markets so that those with DC plans could purchase immediate annuit= ies with part of their funds to cover longevity risk. This economic issue is of paramoun= t importance.

<= /o:p>

Health Care

In 2002, whi= le health care spending in the U.S. reached $1.6 trillion and was 14.9% of GDP, it was still true that 43.6 mil= lion Americans did not have health insurance.&n= bsp; The uninsured constituted 15.2% of the population, up from 14.6% the previous year. Because indivi= duals and businesses pay for health care insurance in the U.S., the decision to drop co= verage is an economic decision. It is estimated that every percentage point rise in the cost of health care insur= ance reduces the number of insured by 300,000.&= nbsp; While the quality of the best health care available in the U.S. is n= ot surpassed anywhere and wealthy individuals from all over the world come to = the U.S. for optimal treatment, the attributes of the U.S. system which have resulte= d in this benefit have a detrimental side as well; the problem of adequate health care insurance coverage is large and growing larger. It’s not just a question of coverage. It is a question of= who should provide the health care coverage and how. The financing of heath care is lin= ked to its availability, and the financing is primarily through insurance, raising= the question of whether insurance is a right if access to health care is deemed= to be a right.

Financing of= health care may be provided by an employer, the individual or the government. The U.S. system of (primarily) em= ployer provision of health care financing through private insurance coverage was an historical accident. American= s have always been distrustful of a strong governmental intrusion into their perso= nal lives and have looked with suspicion upon governmental provision of health = care and its financing. This suspi= cious independence was of little economic importance until the mid 1930s when the great depression along with the advent of antibiotics and the rise of a hospital based health care system transformed the face medical care in the = U.S.. Additionally, during the wage and = price controls of World War II it was hard to get and keep labor, so firms implemented employee benefits to attract and keep workers, thus creating a private system of health care financing through employer group insurance. It was ruled that this type of deferred contingent compensation did not violate the wage and price controls, so it remained for competitive reasons. There = are also significant tax advantages for this arrangement since the Internal Rev= enue Service ruled that firms could recognize the health care premium paid on be= half of their employees as a business expense deductible from income, while employees do not pay taxes on this benefit. Finally, the National Labor Relati= ons Board ruled that employee benefits were an appropriate subject for collecti= ve bargaining union negotiations, which, together with the tax advantages are a strong reason why the schemata continued. The rub now is that employer heal= th care costs are expected to rise by 12% in 2004, making it the fifth straight year of double digit inflation in health care costs. This increases the health care com= ponent of labor costs after taxes, e.g., by 9% if the corporate tax rate is 34%, e= ven before any increase in the wage rate. Unless there is an equivalent rise in labor force productivity, this inflation increases the competitive firm’s marginal cost, reduces its supply and its profit. This i= s true for domestic as well as international firms but is especially relevant for = the international firms competing with foreign competitors that do not cover medical care. The economic implications for outsourcing of jobs and discontinuation of employee benefits are apparent and beg resolution. =

The individu= al provision of some health care is being pushed by the Bush administration. The intent is that individuals can purchase high deductible policies for catastrophic health care coverage and= use health savings accounts to pay normal medical expenses; the health savings accounts would allow tax free accumulation of funds for health care and wou= ld not impose as many constraints on the kinds of medical expenses incurred. While the plan promises added flexibility, it is not feasible for low income families and so provid= es no promise of expanded coverage there; it also stretches the notion of the consumer rationality beyond the point of credulity since each must anticipa= te and save for uncertain but normal medical expenses.

Government p= rovision of health care is also an option. The government has not chosen nor is it likely to choose that univer= sal health care option. It does, however, provide Medicare and Medicaid, i.e., health care for the elderly a= nd poor respectively. Medicare a= nd Medicaid accounted for about one third of the total health care cost of $1.6 trillion in 2002. This propor= tion will grow as the baby boom reaches retirement age and swells the ranks of t= he elderly and so although the country may not move toward universal health ca= re politically it will move in that direction demographically. In fact, Medicare and Medicaid wil= l become insolvent sooner than Social Security, meaning that the current system of governmental health care provision does not seem to be a solution to the fundamental issues in the U.S..

<= /o:p>

Terrorism

Once the con= cern of select nations, terrorism has now become an international problem of unprecedented proportions. The September 11^th losses, unlike nat= ural catastrophe losses, were unexpected and unfunded. The total insured losses = have been estimated at 30-58 billion USD, with total economic losses to New York in the = months following are estimated at 105 billion USD. The reinsurance and some insura= nce markets hardened; the markets had actually hardened prior to September 11th and the attack accelerated the cycle. Insuring some property became quite a problem after September 11th and terrorism exclusions became common place in 2002. The lack of insurance coverage, or extremely escalated premium costs presented substantial economic problems f= or businesses that had mandated insurance coverage in their contractual agreements.

Terrorism in= surance remains an important issue for the industry. Before the September 11^th, 2001 attack, commercial insura= nce contracts generally included terrorist coverage however that coverage was n= ot explicitly priced. After September 11th, however, special measures legislated in the Terrorism Risk and Insurance Act (TRIA) were required to assure continued risk transfer coverage, without wh= ich major industries such as commercial airlines and energy faced potentially halting operations or even bankruptcy; the American Airline industry is required to carry at least 750 million USD of coverage to operate. TRIA was enacted in November 2002 = and covers certified acts of international terrorism. Insurers pay 10% of the losses abo= ve a deductible and losses are capped at $100 billion. The act is scheduled to expire in December 2005 unless it is extended by the U.S. Congress; the deadline was = initially imposed to enable insurers time to build reserves sufficient to handle loss= es. The P/C industry’s surplus s= tood at $334 billion at the end of 1999 and $347 billion at the end of 2003 or a= bout 4% higher than before the attack. This small increase in surplus does not suggest the industry is capa= ble of covering a $50 to $100 billion terrorism loss without severe financial distress or that the scheduled December 2005 sunset provision provided enou= gh time or motivation for the requisite increases in reserves. The intention of the U.S. go= vernment of extricating itself from this back-stop insurance coverage is clear but t= he actuarial and economic issues must be resolved before the private insurance industry can again assume this mantle of coverage. The issue now is whether to allow = the act to expire or to renew it in some form.=

While the go= vernment has mandated that insurers make terrorism risk coverage available as part of insurance options, there is probably a mistaken belief among many businesses that this risk is only a problem in places such as Washington DC or major cities like Los Angeles, New York and Chicago; the anthrax laced envelope received in Boca Raton, Florida, however, showed that the target need not be large. The anthrax event in <= st1:State w:st=3D"on">Florida cost nea= rly $100 million so a mistaken belief could be devastating. A survey conducted by the Council = of Insurance Agents and Brokers in 2003 showed that less than 10% of small commercial accounts and less than 20% their medium to large accounts had bought the terrorism insurance offered by their insurers (http://www.irmi.com/Expert/Articl= es/2003/Ciab03.aspx). Some businesses observing high pre= mium costs have chosen to manage risks through increased preventative actions li= ke security enhancements, controlled access for visitors, increased paroles, a= nd stand off barriers. Others si= mply ignore the risks hoping this will not be disastrous. The insurance industry= is itself struggling with pricing issues; the modeling is subjective and heavy tailed catastrophic probability models varying widely in resulting premiums.<= /o:p>

Renewal of t= he TRIA remains an issue since underlying insurability questions have not been adequately addressed. Those questions in= clude: Is terrorism risk insurable? Is private insurance feasible for the transfer= of this risk? And is private or public insurance cover the best response? The market response concerning the first two questions was clear in 2002; terrorism exclusions in commercial contracts were common until TRIA required that insurers provide terrorism coverage late that year. The industry did not feel that the= risk satisfied insurability conditions available in standard risk management textbooks including measurability of the risk, affordability of premiums, a= nd the potential for catastrophic impact of losses on the insurer. The evolution of this market remai= ns in doubt.

Credit Scoring=

The most imp= ortant new development in insurance underwriting and rating in the U.S. is= the use of a person’s credit history to predict their level of losses. A formula is constructed using approximately 10-50 of more than 450 variables obtainable from an individual’s credit file to derive a summary score directly related to predicting insurance losses. = While the advent of high speed computers has now made the use of massive credit f= iles for predicting insurance losses a viable option, it has been known since 19= 49 that credit history is related to driving accidents. Tillman and = Hobbs (1949 The American Journal of Psychiatry) showed that drivers with bad credit history have repeated crashes at a rate six times higher than those with good credit history. A 1968 study of Washington state d= rivers (Insurance Institute for Highway Safety, 2003) showed, that within the grou= p of drivers who had a history of no automobile accidents, 64% had good credit w= hile 35% had bad credit, while among drivers with two or more accidents, 35% had= bad credit, almost twelve times the percentage of these same drivers with good credit.

Recent stati= stical and actuarial studies have confirmed this relationship, and forged the justification (based on Actuarial Standard #12 from the Actuarial Standards Board) needed to consider this as a classification variable. Monaghan (Casualty Actuary Soci= ety Forum 2000) showed strong relationships between credit score and losses= in both automobile insurance and in homeowners insurance. Further studies in automobile insu= rance by several other authors have also confirmed these findings. For example, in a study commission= ed by the Texas Legislature, Kellison and Brockett (2003 http://www.utexas.edu/depts/bbr/b= br_creditstudy.pdf) examined 175,647 separate automobile policies and examined the relationship between credit score and losses. Results showed that the correlation between the credit score decile = of the policyholder and the relative loss ratio experienced on policies in this decile was over 95, highly significant.&nb= sp; Moreover, the relationship between credit score and losses encompass= es information not already contained in other existing rating variables such as age, sex, martial status, driving history, etc. The correlation between credit sco= res and losses on homeowners insurance has been claimed to be as large as or la= rger than that for automobile insurance. This is a very significant development introducing a potentially important underwriting and rating variable.

In spite of = its success and clear proof of its correlation with losses the use of credit scoring is controversial since it has not been made clear to date why there is a correlation between credit score and insured losses. Several state legislatures in the = U.S. are considering banning or severely restricting the use of credit scores in classification because of this inability to reason why this process works (= thus potentially holding credit scoring to a higher standard than other actuarial rating variables).

A recent stu= dy by P. Brockett and L. Golden (2004) presents biological and psychobehavioral correlates of risk taking propensity and shows there are fundamental chemic= al and psychological characteristics involved in both financial risk taking and risky driving behavior (e.g., blood levels of serotonin, MAO and testostero= ne, and the personality traits of sensation seeking and novelty seeking common = to risk taking in both financial and driving). An assessment of credit risk can t= ap information about the biochemical and psychobehavioral make-up, and hence y= ield information concerning future automobile risk taking and loss propensity. This forthcoming article helps exp= licate reasons why credit scoring works, but the controversy concerning the use of credit scoring is likely to be hotly debated and unsettled in the U.S. for decades to come.