Font size:
Michigan auto insurance premiums are among the highest in the nation. The American Association of Retired Persons, in a recent survey, found that Michigan’s premiums were the second highest in the nation, behind only Louisiana. This, combined with a statutory requirement to purchase insurance, has led to legislative attempts to keep premiums down. Unfortunately, state lawmakers have pursued an approach that includes price controls, regulation of how premiums may be set, and requirements for insurance companies to provide specific types of coverage. As the famous Austrian economist Ludwig von Mises pointed out decades ago, this kind of government intervention, while well-intended, leads to unintended consequences that then lead to further government interventions, further unintended consequences, in a lengthy cycle with results that no legislator would have expected at the beginning.
Rather than attempting to regulate insurance company and individual behavior, Michigan legislators would much better serve the people they represent by examining why insurance premiums are so high in the first place, in order to address the problem at its source. A careful study of Michigan’s insurance market and the regulations governing it indicates that no-fault insurance and the legislative requirement for individuals to purchase unlimited personal injury protection are two important reasons for the increased costs of providing insurance coverage in Michigan. The good news is that it is possible to reduce these costs and reduce the number of drivers who take the risk of violating the law and do not purchase insurance.
Michigan auto insurance premiums are among the highest in the nation. The American Association of Retired Persons, in a recent survey, found that Michigan's premiums were the second highest in the nation, behind only Louisiana.[1] This, combined with a statutory requirement to purchase insurance, has led to legislative attempts to keep premiums down. Unfortunately, state lawmakers have pursued an approach that includes price controls, regulation of how premiums may be set, and requirements for insurance companies to provide specific types of coverage. As the famous Austrian economist Ludwig von Mises pointed out decades ago, this kind of government intervention, while well-intended, leads to unintended consequences that then lead to further government interventions, further unintended consequences, in a lengthy cycle with results that no legislator would have expected at the beginning.[2]
Rather than attempting to regulate insurance company and individual behavior, Michigan legislators would much better serve the people they represent by examining why insurance premiums are so high in the first place, in order to address the problem at its source. A careful study of Michigan's insurance market and the regulations governing it indicates that no-fault insurance and the legislative requirement for individuals to purchase unlimited personal injury protection are two important reasons for the increased costs of providing insurance coverage in Michigan. The good news is that it is possible to reduce these costs and reduce the number of drivers who take the risk of violating the law and do not purchase insurance.
The purpose of insurance is to transfer to an insurer the policyholder's risk of suffering a monetary loss should an unpredictable event happen. For example, fire insurance protects you from a large monetary loss should your house burn down. Insurance companies are able to provide such a service by correctly estimating the chances of such an accident happening and charging each customer an amount that is equal to the expected loss from an accident.
The expected loss is equal to the probability that a loss will occur times the cost of the loss. For example, if there is a 1 percent chance that your house will burn down during the year, and it will cost $100,000 if it does burn down, then the expected loss is equal to 1 percent of $100,000, or $1,000 per year. Most people are risk-averse and therefore will purchase insurance even if the premium is a little more than the expected loss, rather than self-insure or take on the risk themselves. This is generally the case because individuals cannot easily spread the risk of the loss on their own. While the chance of your house burning down is small, if it does you may not be able to replace it, so you buy insurance.
Insurance companies earn a profit by spreading the risk and charging a little more than the expected loss. If an insurer has 1 million customers, then it can expect that a large number of houses will burn down, but if it has calculated the risk correctly, then it can expect to have enough money from premiums to make the payments.
Insurers may charge premiums that are less than the expected loss if they can invest the premiums and the losses occur much later. Insurance companies primarily invest in fixed-income securities, which have a return that is generally less than the equity market, but carry much less risk. (Bonds still involve risk, including the solvency of the borrower and price volatility if the bond cannot be held to maturity due to cash flow problems.) Because the rate of return on investment is not risk-free, it is important that companies correctly estimate the probability and cost of an accident.
For an insurance company to cover its costs, and thus provide insurance in the first place, it must use whatever information it has available to help it better calculate expected losses from accidents-not only for its entire customer base, but also for smaller groups within its customer base that have different expected losses. For example, suppose that we know that people who have accumulated speeding tickets are more likely than drivers who have not received a ticket to have an accident. Successful insurance companies will use this information when setting premiums that are sufficient to cover their expected costs.
Conversely, were government policies to preclude auto insurers from using this information, then the insurers will take on more risk than they would otherwise have to, and would have to charge higher premiums for everybody. Such policies make it more expensive to provide insurance, and thus lead to higher premiums and encourage those who have low income, low expected losses, or both, to avoid purchasing insurance.
Another way to raise insurance premiums is to require insurance companies to offer only policies that provide high amounts of coverage. For example, if insurers could only sell fire insurance policies that had no deductible, they would have to charge much higher premiums in order to stay in business, since the loss to be covered would be greater than if the individual policy holders had the option of paying a deductible.
An important consideration to keep in mind when discussing insurance is the problem known as moral hazard. This is the problem that arises when a person is insured against a specific risk and the security provided by that insurance discourages him or her from taking optimal precautions against incurring an accident. For instance, if you have car insurance, then you may be less likely to drive as carefully as you would if you did not have insurance. (In the automobile insurance market this is likely mitigated by the fact that a car accident may cause physical damage or bodily injury that the driver may wish to avoid even if he was compensated for the financial cost of the accident.)
There is a related problem to moral hazard, which studies have found to have a significant effect on insurance premiums — third party payment. This problem particularly besets medical insurance. When an individual suffers an injury and the insurance company must pay all — or nearly all — medical costs, both the injured party and those who are delivering treatment will have every incentive to choose very expensive treatments even if the added value of those treatments is small. We have all seen the advertisement for the motorized wheel chair that we are told is "free" to individuals covered by Medicare. With a price of zero, people who have no idea what the cost of these wheelchairs actually is will be more likely to demand them.
This encourages the use of very expensive treatments and makes it difficult for insurance companies to figure out the expected loss from an accident because the insurer's liability is not limited. As noted above, the insurer must correctly estimate the probability and cost of an accident in order to set a premium that allows it to stay in business. If the insurer does not know what the liability will be then it will not be able to set premiums with any degree of accuracy, and in order to protect itself from bankruptcy will set premiums higher than if the liability were limited.
In a 2010 RAND Corporation study, economist Paul Heaton found that, "[I]n 2007, average total auto insurance premiums in Michigan were 17 percent higher than those in the rest of the country ($928 versus $795)."[3] (This is the latest available data.) Michigan has consistently had some of the country's highest insurance rates, ranking among the top 15 states every year since 2003. Heaton cites an Insurance Research Council report that estimated that 17 percent of Michigan drivers failed to purchase mandatory auto insurance in 2007 — the ninth highest total in the United States.[4]
Michigan's auto insurance market has several important characteristics, but two stand out. First, Michigan is one of 12 no-fault insurance states.[5] Under no-fault insurance schemes, drivers are limited in their ability to sue for recovery of non-economic losses, such as damages for pain and suffering. In return for this limitation, the driver's insurance company pays for economic losses, such as medical expenses and lost wages, regardless of who was at fault. The expectation was that by reducing the costly process of litigation, premiums would be lower than under the tort system, victims would be compensated more quickly, and fraud would be reduced.
Massachusetts enacted the nation's first no-fault insurance system in 1971, which met with considerable popularity in the beginning. Several states adopted no-fault systems in the 1970s, but then the bloom fell off the rose. In another 2010 RAND study examining the no-fault system, Heaton and co-authors James Anderson and Stephen Carroll found that no-fault states have higher premiums than tort law states, generally due to higher medical costs in no-fault states.[6]
A second major characteristic of Michigan's auto insurance industry is it is the only state that requires consumers to purchase unlimited personal injury protection (PIP), which provides extremely generous reimbursement benefits for medical costs and lost wages.
Anderson, Heaton, and Carroll ask why auto insurance costs more in Michigan. They found that property damage losses in Michigan are equal to, or lower than, those in other states, and that auto collisions in Michigan are less severe. They also found that consumer expenditures for all types of auto repairs are about what they are in other states. So there had to be another reason.
The authors developed a model that used 72 variables to capture claimant demographics, accident circumstances, and reported injuries, and found that it costs 57 percent more to settle a Michigan claim for personal injury than the same injury would cost in another state.[7] They also cite Fast Track data that show Michigan injury losses per insured vehicle were 40 percent higher than in the U.S. as a whole.
So why are injury costs so high in Michigan? The share of claimants who seek medical treatment after an accident was not appreciably higher in Michigan, but the mix of services was more costly. Michiganders are much more likely to use expensive treatments, such as hospital and emergency room services, X-rays, and CT scans, and to recover wage-loss payments. Anderson, Heaton, and Carroll conclude that this extensive use of medical services is a primary cause of Michigan's high premiums. A major reason for this high use of medical services is the unlimited PIP medical payments and generous wage-loss payments mandated under Michigan law.
A major weakness in Michigan's auto insurance system is the requirement for consumers to purchase unlimited personal injury protection, for several reasons. First, as any economist will attest, people respond to incentives. Once third-party payment is introduced, and there is no limit on how much the third party must pay, then there is every incentive for health care providers to choose expensive methods for treating injury, and there is no incentive for the patient to restrain expenditures.
Second, the lack of a limit to the insurance company's liability creates a good deal of uncertainty as to the expected loss from writing policies, and makes it difficult to price premiums efficiently. This uncertainty reduces the number of companies that would be willing to write policies in Michigan, particularly small insurers who cannot spread their risks nationally. In 1978, in an effort to address this problem, Michigan instituted a program to reduce this uncertainly for insurers — the Michigan Catastrophic Claims Association, a state-mandated reinsurance program that allows insurers to retain only the first $480,000 of a loss and to be reimbursed by the MCCA for the excess. This has resulted in its own problems, as discussed below.
Finally, consumers are not allowed to choose a policy that fits their risk preferences, and are forced to buy more expensive insurance than they would choose. This is a particular problem for low-income residents, who when faced with high insurance premiums may choose to evade the law and not purchase insurance. The fact that Michigan ranks so high in its number of uninsured drivers can be explained in good part by the inability of drivers to choose less expensive policies.
Insurance Information Institute President Robert Hartwig, in testimony before the Michigan House Insurance Committee, showed that Michigan's high insurance premiums are being driven by rising medical costs associated with auto accidents. The average no-fault PIP claim rose by more than 250 percent from 1998 to 2007, reaching $31,383.[8] Given the incentives by medical care providers to use expensive treatments, it is a problem that the state has no constraints on costs, such as medical fee schedules and treatment protocols.
Requiring unlimited personal injury protection is a very inefficient mechanism for dealing with catastrophic injury. It distorts the price of insurance, misallocates resources, and reduces consumer choice. The most effective way to deal with the high premium costs in Michigan is to allow consumers to choose a specified PIP limit — for example, $50,000, the level in New York State, which is the highest current threshold in the country. This would give consumers an incentive to monitor the cost of their treatment. It would also result in total premium savings, according to Heaton's research, on the order of 15 percent. In a 2007 study of more than 70,000 PIP claims in Michigan, actuarial expert Michael Miller found that about 94 percent of the claims were under $50,000. The claims that exceeded $400,000 made up 0.5 percent of the claims, but accounted for more than 42 percent of the PIP claims' loss dollars. He estimated that limiting Michigan PIP benefits to the $50,000 level would save 45 percent on the PIP premiums and 15.8 percent of total premiums.[9]
As noted, a key aspect of a healthy insurance industry is the ability to correctly forecast the probability of an accident as well as the loss associated with an accident. Eliminating the requirement for purchasing unlimited PIP coverage will allow insurance companies to better estimate the loss from an accident. In addition, insurance companies should be allowed to use the best models available for estimating the probability of an accident.
The Michigan legislature should not restrict the ability of companies to use the risk-management and underwriting tools they choose. There are more than 100 insurance companies competing for business in Michigan.[10] They have every incentive to use those variables that are most likely to predict the chance of an accident. Using variables that are not accurate will result in underwriting losses that will drive the company out of business. It is not possible for a legislative body to know which variables will be successful and which are not, and it has no incentive to get it right.
If lawmakers are concerned about the cost of premiums for certain subsets of the population, particularly low-income persons, they should not attempt to create an artificial subsidy for all drivers by restricting the ability of insurers to properly price premiums, and instead focus only on how to bring down premiums for that particular subset of the population.
When the Michigan no-fault law was enacted in 1973, it was recognized that certain people injured in automobile accidents would not be covered by policies that provided the medical and wage loss coverage that vehicle owners were required to purchase.[11] If those injured people were not in violation of the law, they would be compensated for economic losses. An example of people in this group would include a pedestrian who did not own a vehicle, who resided in a household with no insured vehicles owned by others, and who was injured by an uninsured or a hit-and-run driver.
The mechanism for coverage for such people is the Michigan Assigned Claims Facility (ACF), located within the Michigan Department of State. The Facility assigns the claims to "servicing insurers" that pay the medical costs and lost wages. There are currently 10 such insurers, who are reimbursed for their loss payments, loss adjustment expenses (LAE), and interest from a fund raised through assessments on the entire auto insurance industry and legally self-insured businesses. (A self-insured business is a company that acts as its own insurer by maintaining funds and managing its own claims.) These assessments are prorated based on each insurer's share of the total state premiums for no-fault and auto liability insurance and self-insured businesses' total share of self-insured vehicles. The costs of these assessments are, of course, passed on to the insurance customers in the form of higher premiums.
In the early years, these annual assessments were in the low tens of millions of dollars, but they have increased dramatically in recent years. In 1997 the assessment was for $37.7 million; by 2008 it had grown to $145.6 million — a 286 percent increase. During the same period, the assessment per Michigan vehicle grew from $6.04 to $20.66 — an increase of 242 percent, indicating that any growth in the total vehicles in the state was at most a minor factor in the increase in total losses, LAE, and interest reimbursed by the ACF.[12]
During the same period, the total annual paid PIP losses for all Michigan vehicles grew from $788.1 million to $1.7 billion, an increase of 119 percent.[13] Since the increase in ACF assessments is 2.4 times that in PIP claims payments, it is clear that something is going on with assigned claim payments. If this difference had been identified earlier and if steps had been taken to hold the increase in ACF assessments to the same rate as the growth in overall PIP losses, the 2008 assessment would have been $82.6 million — $63 million, or approximately $8.94 per insured Michigan vehicle, less than it turned out to be.
Unfortunately, the Facility, with only eight employees, one claims examiner, and over 3,000 files to monitor, is not equipped to determine what that "something" is. Health privacy concerns preclude insurance industry experts from viewing the files. Individual servicing insurers could analyze their own files, but none of the 10 has a sufficient share of total claims for its own analysis to be applicable to all the servicing insurers. Further, since the increase in payments might be, at least in part, the result of inadequate servicing carrier performance, an investigation by those insurers might be criticized as lacking objectivity.
However, there is a readily available source of expertise that could be tapped. The Michigan Auto Insurance Placement Facility (MAIPF) was established to provide auto insurance to applicants unable to obtain it in the regular market, usually because they are perceived as unacceptably high risks and are more like to have claims than the average driver. Managed by a board consisting of seven auto insurer representatives, two agent representatives, and two public representatives, it operates under the scrutiny of the insurance commissioner, a representative of whom attends all meetings. The commissioner approves the MAIPF plan of operation, its procedures, and its rates. The board has committees of company employees with expertise in claims, accounting, actuarial science, and underwriting.
Legislation has been proposed to transfer the administration of the ACF to the MAIPF. The proposal is supported by the Secretary of State and has been reviewed by the Office of Financial and Insurance Regulation, which raised no objection. Once enacted, the MAIPF can begin a detailed study of the claims experience of the ACF, identify the problem causing the unnatural increase in losses, and implement a solution. Reductions in assessments and insurance premiums should follow.
As mentioned above, the Michigan Catastrophic Claims Association was created by the legislature in 1978 to address the problem that many insurers were unable to afford reinsurance protection for the portion of unlimited PIP coverage that they were financially unable to retain as their sole responsibility. Under the MCCA's excess reinsurance, a primary insurer pays a reinsurer to cover the amount of a claim in excess of a fixed level, known as a "retention." When the costs of a claim exceed that retention level, the reinsurer pays that portion of the claim or reimburses the primary insurer. This prevents a primary insurer from being made insolvent by a single large claim.
Different retentions are appropriate for different insurers, depending on factors such as size. However, the MCCA legislation mandated participation by all insurers at a $250,000 level. This was too large for some insurers, which had to buy additional reinsurance to cover the gap between an appropriate retention and $250,000. Larger insurers had to pay for reinsurance at that level whether they needed it or not. In essence, this fixes prices for reinsurance. Consumers cannot benefit from lower premiums resulting from reinsurance savings that efficient primary insurers might otherwise obtain.
Legislation passed in 2001 implemented gradual annual increases in the MCCA retention, up to $480,000[14] effective July 1, 2010, then to $500,000 effective July 1, 2013, and thereafter increased biennially by 6 percent or by the same rate of increase as the consumer price index (CPI), whichever is less. None of these retentions is any more appropriate for any particular insurer than the original $250,000. The arbitrary nature of these increases is evidenced by the fact that in 2001 a legislature that was interested in adjusting the retention to reflect CPI increases would have raised the original $250,000 retention to approximately $800,000, rather than the $300,000 it actually mandated for 2002.
The MCCA is required to calculate annually the total premiums that will be needed to pay all claims that will be incurred in the ensuing year. This immediately proved difficult (something that private reinsurers already knew and that explains their reluctance to provide the reinsurance for an unlimited amount). That total is then assessed to all insurers according to the number of vehicles each insures. Technically the assessment is calculated on a "car year" basis — one year of coverage for one car. The assessment for the first year was $3.00. By the 10th year, the amount had increased to $22.67 — $15.77 to cover the year's anticipated losses, $6.81 to apply to prior year deficits, and $0.09 for MCCA expenses. In later years the assessment has ranged between $5.60 and $143.09. For the year beginning July 1, 2010, the largest assessment on record consists of $116.84 to cover the year's claims and $26 to reduce a deficit, and $0.25 to cover MCAA administrative costs.[15]
Some critics have contended that the volatility in assessments results from mismanagement or incompetence, but in fact it is caused by PIP coverage being unlimited. When an accident occurs, an insurer must establish a liability, known as a reserve, on its books equal to the total anticipated payments it will make on the claim. For serious accidents, this can amount to hundreds of thousands of dollars per year for many years in the future, often for the lifetime of the injured person. When notified of the loss the MCCA must establish a reserve for the amount in excess of the retention. If the MCCA set such reserves for the total amount that would be paid over time, the annual anticipated losses that help determine its required annual premium would be dramatically increased, resulting in enormous increases in that premium and thus in per-vehicle assessments as well.
These claims will be paid out over long period, so the MCCA does not need to collect all the money at once. Instead, it discounts the reserves based on anticipated investment income it will receive over the duration of the claim payments. This discount reduces the amount needed for assessments. Meanwhile, medical costs are likely to increase over time, so the MCCA includes these increases in its calculation, thus increasing the amount it might otherwise need to collect. Anticipated investment income normally exceeds the anticipated increase in medical costs, so the net effect is to reduce the estimate of premiums and assessments needed to pay all the claims. This discounting has a dramatic effect. In 2010 the future payments for 12,404 active claims are expected to exceed $71 billion, but the discounted reserves are about $12.7 billion.[16] The difference is billions of dollars that did not need to be collected.
Nonetheless, the rates assumed for investment income and medical inflation are necessarily estimates. No one can be certain of investment returns over the next several decades, and no one can be certain of how much medical costs will increase. No one can even predict the cost of claims in the next year. Conditions change over time. The MCCA must acquire as much information in these areas as possible and adjust its assumptions about interest rates, stock market returns, inflation, and claims trends.
There are additional unintended consequences arising from the methodology chosen by the legislature to determine the per-vehicle assessments each year. That assessment is determined on a car-year basis (a unit of measurement to which a rate is applied to determine a premium). Regardless of the type of motor vehicles insured by a company, the per-vehicle charge for a particular year will be the same for all vehicles. No consideration was given to the possibility that the actual claim cost for some classes of motor vehicles would vary significantly from the average charge imposed by the MCCA. One of those classes is historic vehicles, which were legally recognized as an exception in 2003. Under this exception, the charge for "historic vehicles" is 20 percent of that for other vehicles. The 20 percent charge was not based on any data about the actual costs of that class, but was a political concession to a special interest group whose vehicles obviously have smaller exposure to catastrophic losses.[17]
At the time of that legislative change, a more significant class with claim costs well below average — which had actually been measured by the MCCA — was brought to the attention of legislators. Commercial vehicles, based on an actuarial analysis of the period from September 1, 1996, through December 31, 2001, had an average annual MCCA-covered loss of $10.78 per vehicle. Private passenger cars had an annual average of $62.31. The annual average for motorcycles was $151.04.[18] But all three were charged the same annual assessment.
An example of the impact of this scheme is the 2002 MCCA assessment of $71.15 per vehicle. Had that assessment been distributed among the three classes in proportion to the MCCA claim costs of the previous five years, commercial vehicles would have paid $12.77, motorcycles $179.04, and private passenger vehicles $73.85.[19] Commercial vehicles subsidized the other classes to the tune of $58.38 per vehicle for a total overcharge of approximately $29 million, based on an estimate of 500,000 insured commercial vehicles.
The reason for the much smaller per-vehicle claim costs for commercial vehicles is that when an employee is injured in a commercial vehicle, and insurance is available from both the employer's workers' compensation and automobile policies, the workers' compensation policy pays first. The employer pays the cost of the injuries through its workers' compensation policy, but then, because of the MCCA subsidy, has to pay again for the injuries through its automobile policy premiums. The extra $29 million could have helped create a lot of jobs in Michigan. The legislature addressed the issue of commercial overcharges in 2003, but due to the opposition of insurers that wrote predominantly personal lines and that apparently did not want to see a $2.70 (3.8 percent) increase in the per-vehicle assessment paid ultimately by their customers, the subsidies continued. As a result, no action was taken, and the MCAA overcharges — of 450 percent on Michigan businesses — have continued to the present day.[20]
As noted, the MCCA assessment is now significantly higher than in 2002 and has been so for all the intervening years except 2003, when it was $69.00.[21] The negative impact on Michigan business investment has therefore probably been even greater than in the above example. The MCCA's board should initiate a study to determine that economic damage. Lawmakers should change the law to require that commercial vehicles be assessed by the MCCA based upon their actual claims experience.
Michigan's insurance industry provides a good example of the unintended consequences of government intervention in the market, and of Mises's interventionist dynamic, in which government intervention begets more intervention.
In an attempt to keep any particular person from suffering economic losses due to an auto accident, the legislature enacted the requirement of unlimited personal injury protection. This created a situation where insurance companies face great uncertainty as to their possible liability in auto accident claims, and has led to overuse of expensive medical treatments. This in turn has led to high insurance premiums.
In order to address the reluctance of insurance companies to insure against unlimited losses, the legislature created the Michigan Catastrophic Claims Association. This then led to assessments that have been inefficient, volatile, and have reached levels that are in excess of $100 per vehicle. In response to the high insurance premiums that resulted from this, the legislature restricted the use by insurance companies of variables that are correlated with certain socioeconomic characteristics, such as income and location. This, in turn, increased the cost of providing insurance for the majority of drivers. This led to political pressure to limit premiums and further government interference in the market, with additional unintended consequences, as Mises discussed.
Thankfully, that cycle can be reversed. There are some things Michigan lawmakers can do. First, Michigan legislators should replace the unlimited personal injury protection purchase requirement with a system that is more aligned with that in other states. Second, insurance companies should be free to use the risk-management and underwriting tools they deem most fit to determine the probability and costs of accidents. Finally, the Michigan Catastrophic Claims Association should be wound down as reinsurance becomes more affordable for insurers. Combined, these measures would reduce the pressure on premiums and the attendant political pressure to attempt to centrally plan the auto insurance industry in Michigan. The biggest winners would be consumers, who would benefit from lower premiums and greater choice.
