I could be wrong about a number of things, but I try --in this forum --to be correct in applying economics to situations involving risk and insurance. My opinions are based upon economic theory which includes the theory of how insurance is priced. In a comment to a previous post on free insurance someone said,
"No one is asking for free insurance. Compare the expected losses to the premiums. I dare you."
So, I take up the dare. Except I punt a little. I do not have access to expected losses. So I looked at past performance.
One of the problems with this incredibly simplistic assertion is that it is based on the assumption that everything works with in a one year period. Firm's are in business for the long run, thus they expect to make profits in the long run. In fact, they do not have to make a profit every year, just in the long run. Otherwise they will leave the market.
In the figure below I plot the loss ratio (losses incurred in a given year to premiums earned). Thus this is actual losses rather than expected. Florida has the big swing in 2004 and 2005 due to the hurricanes in those years. If we look at the average of these ratios over time Florida's is 72.77. That means that for every dollar of insurance premium paid, the state's customers received 72.77 cents in loss payments. Kansas (one of those middle states without any hurricane risk) has a relatively stable loss ratio and an average ratio of 62.64. That means for each dollar of premium paid, the state's customers received 62.24 cents back in loss payments. Finally, in the U.S. overall, the average loss rat is just a bit above Kansas'.
So, Florida's customers already get a better deal than the average customer in the U.S. given that they receive more back in terms of loss payments. They would like an even better deal (as would I). To get this better deal, they would like Kansas and the other states to chip in.
The data come from the NAIC and are available from them if the commentator wants to check my work. To be fair, maybe no one wants free insurance. Maybe they just want less expensive insurance. Either way, someone else who doesn't share that risk has got to chip in to make it so.
Insurance is not socialized risk sharing. It is sharing of risks where everyone pays their risk based price. And for the record, I was (and still am) against federal programs like TRIA. I also believe that the private market could cover flood insurance.
(Note this post was updated slightly from the previous version because I meant to save it and I inadvertatnly published it before it was finished).
Thanks for the information. Of course, the key question about your data is whether the 2004/2005 hurricane losses in Florida are a 1 in 10 event. If they are 1 in 15 or 20 or 25, the numbers change. At least you didn't start your Florida data in 1992 and end it in 2005 as your buddy Hartwig always does to rig the sample in the industry's favor. I suppose I could pick 1994 to 2003 and Florida would beat Kansas, but that is not my point.
My point is that the private insurance model will not work for hurricane high-risk areas for the same reason it doesn't work for floods or earthquakes.
Insurance for low frequency/high severity events requires insurers to collect much more in premiums than they expect to pay in claims so they can build up reserves and buy reinsurance each year in case this is the year the big one hits. You guys taught me that the premiums have to be high enough to guarantee reinsurance investors very high rates of return so they will invest in reinsurance instead of some other investment bubble. But you only see this from the industry point of view as justification for more premium increases whereas I see it as the reason for unavoidable market failure. It really doesn't matter how much states regulate rates or the structure of state risk pools, those are just your scapegoats. The system does not work and will not work. You can't expect people to pay 5 to 7 times more in premiums than their expected claims. That is what reinsurance costs for high-risk areas according to the pro-industry research that you and the Wharton guys and others have published. I thank you for showing me precisely why only the federal government can provide hurricane insurance in a large enough pool that spreads the risk geographically so that a major hurricane affects only a small portion of it, and that the government can set rates closer to expected losses, with a cushion for timing risk and to build reserves, and keep them there without the deliberate "hard market" manipulation and profiteering of the market by the reinsurance industry after every major disaster.
Posted by: Brian Martin | May 15, 2009 at 03:35 PM