Why Do we Have Mandatory Auto Insurance?
Mike the Actuary (here and here) is musing about how to fix the mandatory auto insurance system. Most states mandate coverage, penalize those without insurance, and then lean on insurers to keep the rates low. This causes all sorts of pressure to refrain from rating based on risk and, in general, turns the automobile insurance system into a socialized risk program.
After musing along with Mike, I started to question why we even have this mandatory coverage. One could make the argument that the automobile insurance system should be socialized due to the negative externality caused by those #$%@ uninsured motorists. I recall seeing an Insurance Research Council study of uninsured motorists and there was approximately 15 percent unisured motorists across the states. I wonder (no data, just wondering) if the costs to the auto liability system would be lower if we had no mandatory coverage. We could examine this, presumably, if we had the data on participation across states and across time. This sounds like a job for a hungry Ph.D. student!
I was actually musing about this same thing. What I am curious about is the moral hazard aspect of compulsory auto insurance. Do states that have more uninsured drivers actually have lower fatality rates and/or lower accident rates since these uninsured drivers would presumably drive more cautious? Controlling for different regulatory regimes, liability systems, ect. that would be an interesting finding. But the key thing to remember is how would it impact uninsured motorists coverage. Would the net effect be lower cost auto liability insurance? Sounds like a job for a hungry assistant professor vying for tenure!
Posted by: tyleverty | November 30, 2005 at 09:16 PM
I almost put "this sounds like a job for Ty", but I know you are working on a good number of projects.
I, too, was wondering what the equilibrium unininsured motorist coverage would be like. I think I am going to get someone to see if they can track down this data!
Posted by: Martin | November 30, 2005 at 09:22 PM
This much is clear: In most states it's pretty easy to avert the current mandates. Most of the people who do not comply with the law do so because they cannot afford the additional cost, and it doesn't seem that the public will exists to take away people's means of transportation -- and often their means to earning a living.
So instead of being fined or having their vehicle taken away, motorists are given a ticket and the fee is waived when they show up in court with proof of insurance. Naturally, they can then cancel the coverage or cease making payments once the court date has passed.
State efforts at enforcing mandatory insurance, then, have often focused on imposing paperwork burdens on insurers rather than targeting the motorists who don't comply with the law.
My point is that aside from looking at the total cost of UIM (which is still necessary as a coverage because people don't comply with the law)versus mandatory coverage, which ever hungry PhD student or prof examines this issue should also take into account the total costs of enforcing these laws. I suspect they're greater than the supposed benefit.
Posted by: Kevin O'Reilly | December 01, 2005 at 11:44 AM
Not clear that uninsured drivers' incentives are really structured to drive more carefully. As Kevin notes, there's no political will to enforce mandatory insurance laws, and most people who don't buy insurance make that decision because they're outside the system as an illegal alien or too poor to worry about the prospect of civil liability.
Posted by: Ted | December 01, 2005 at 04:49 PM
Kevin and Ted... I agree there is hardly any will to enforce the mandatory coverage laws and that people will only voluntarily buy insurance when they have something to protect. It may be enough of a cost to lose a license for those who do not purchase insurance (the cost of driving without a license is probably high enough in terms of hassles) to provide enough discipline to keep the price of UMI at a reasonable level. However, this is an empirical question and it may differ in each states as the mixture of drivers is quite different.
Posted by: Martin | December 02, 2005 at 04:18 PM