Ten years ago, an economics professor named Randall Wright resigned from his job at Cornell and drove his Dodge Daytona Turbo down to Philadelphia to begin teaching at the University of Pennsylvania. When Professor Wright found out how much Philadelphians pay to insure their cars, he gave up driving.
If you live in Philadelphia, your auto insurance probably costs about three times what it would in Milwaukee--and more than twice what it would in Seattle. Philadelphians have traditionally paid more for insurance than their counterparts in Baltimore, Chicago, and Cleveland, despite much higher theft rates in those other cities. This led Wright to ask a question that ultimately became the provocative title for an article in the prestigious American Economic Review: "Why is automobile insurance in Philadelphia so damn expensive?"
Areasonable first guess is that the answer has little to do with economics and much to do with the behavior of state regulatory agencies. But the facts don't support that guess. Pittsburgh is in the same state as Philadelphia, and Wright could have insured his car in Pittsburgh for less than half the Philadelphia price, even though Pittsburgh's theft rate was then more than double Philadelphia's rate. Other states provide equally striking contrasts: San Jose, Calif., is much cheaper than neighboring San Francisco; Jacksonville, Fla., is much cheaper than Miami; Kansas City, Mo., is much cheaper than St. Louis, Mo.
While Wright was puzzling over these discrepancies, a Penn graduate student named Eric Smith was involved in an auto accident. The other driver was at fault, but he had few assets and no insurance, so Smith had to collect from his own insurer. That unpleasant experience gave Smith and Wright the insight that led to a new theory of insurance pricing.
In brief, the theory is that uninsured drivers cause high premiums, and high premiums cause uninsured drivers. In somewhat more detail, a plethora of uninsured drivers increases the chance that, like Smith, you'll have to collect from your own insurer even when you're not at fault. To compensate for that risk, insurers charge higher premiums. But when premiums are high, more people opt against buying insurance, thereby creating the plethora of uninsured drivers and completing the vicious circle. Once a city enters that vicious circle, it can't escape.
In other words, insurance rates are driven by self-fulfilling prophecies. If everyone expects a lot of uninsured drivers, insurers charge high premiums and then many drivers choose to be uninsured. Conversely, if everyone expects most drivers to be insured, insurers charge low premiums and then most drivers choose to be insured. Either outcome is self-reinforcing. A city that falls into either category (for whatever random reasons) remains there indefinitely.
So it's possible that modern Philadelphians are paying an exorbitant price for a brief outbreak of pessimism among their grandparents. If, for just one brief moment--and contrary to all past evidence--Philadelphians could believe that insurance rates will fall and their neighbors will become insured, that belief alone could cause insurance rates to fall and the neighbors to become insured. And then forever after, Philadelphia's insurance market might look like Milwaukee's.
It's not certain that a burst of optimism would be so richly rewarded; the Milwaukee-style outcome will be undermined if Philadelphia is home to enough of the "hard-core uninsured," who are unwilling to insure themselves even at Milwaukee prices. The Smith-Wright theory predicts that some cities, but not all cities, have the potential to maintain low insurance premiums in the long run.
But in cases where that potential exists, it would be nice to see it realized. One way to accomplish that is by enforcing mandatory-insurance laws. (Smith and Wright point out that enacting a mandatory-insurance law, which a majority of the states have already done, is not the same as enforcing a mandatory-insurance law, which is nearly unheard of. Moreover, even where the laws are enforced, minimum liability limits are typically very low, and probably too low to make much difference.)
I n theory, mandatory insurance could make life better for everyone, including those who currently prefer to be uninsured. Philadelphians who are unwilling to buy insurance for $2,000 might welcome the opportunity to buy insurance for $500. So if mandatory insurance yields a dramatic drop in premiums, then both the previously insured and the newly insured can benefit. (In practice, there will probably be a small segment of the population-- presumably at the low end of the income distribution--who will be unhappy about having to buy insurance even at $500. But income-based insurance subsidies would allow even the poorest of the poor to share the benefits of lower premiums.)