Why Should the Government Subsidize Terrorism Insurance?

Just before Christmas President Bush signed into law a second extension of the Terrorism Risk Insurance Act of 2002 (TRIA). This Act hasn’t gotten a lot of press, but it provides a pretty substantial subsidy for the insurance industry. In non-technical terms (because the technical terms are so complicated that one needs a slide rule, a pocket protector, and several actuarial tables to understand the terms completely), under this Act the government requires all commercial (not residential) property/casualty insurers to offer insurance for terrorism risks. In exchange, the government agrees to pay 90% of all losses over a certain amount (determined by a complex formula) that result from certified acts of terrorism. In short, the government provides reinsurance for commercial insurers in the case of a terrorist attack. But unlike normal reinsurance agreements, commercial insurers pay nothing up front for the reinsurance. They get to distribute their risk free of charge. Rather, if an act of terrorism occurs, and the insurance industry pays out over their limit, and the government is forced to pony up some cash, then the government has the right to recover its costs from future premiums received by the insurance companies.

Two things puzzle me about this Act: why we need the government to provide this service and why the government is providing this service for free.

The common justification for this act is as follows: Without insurance for terrorism, there would be significant disruptions in certain sectors of the economy and this would have an adverse macroeconomic impact. Insurance companies won’t provide terrorism coverage because terrorist acts are not probabilistic and are thus uninsurable. That is, insurance companies claim that they cannot provide terrorism insurance because they cannot predict how often terrorist attacks are going to occur or how large the impact will be. Thus, the government has to step in to force insurers into the market.

But this doesn’t make sense to me. Certainly we have enough data to make some guess about the costs and frequency of terrorism, and insurance companies can price their products based on that information. Moreover, although it is certainly a possibility that an attack will be so large as to make insurance and reinsurance impossible (e.g., the entire country is leveled in fell swoop), most insurance and reinsurance companies should be able to create a portfolio large enough to eliminate most of the concerns about the size of the attack. Please take a look at this article by Dwight Jaffe and Thomas Russell for a more thorough explanation.

Even if one believes that government should step in to this market, it is unclear why the government is providing this service for free. In any other insurance context, the insured has to pay a premium to receive the security that someone else will step in to pay some of the insured’s losses. Here, the insurance companies are paying nothing up front. The government is taking the entire risk. This is quite a subsidy for insurance companies. I assume that this is being factored into the price of the insurance products being offered, but that just seems like another subsidy for business and industry. Remember, there is no requirement under TRIA for insurance companies to provide residential terrorism insurance.

Yet this bill has created a bit of a topsy turvy world in D.C.: The Dems are on the side of the insurance industry and big business and the Republicans are against. What am I missing here?

You may also like...

5 Responses

  1. Carroll says:

    First, the government doesn’t pay till about $40 billion dollars in insured losses (after all the slide-rule and calculator business is taken into account). That’s about more than the size of 9/11 but less than, say, Katrina. For attacks smaller than that, the government doesn’t pay anything.

    Second, it is also remember that between $40 billion and about $100 billion, all of the government payments will be recouped so it is really more like a loan than a pure giveaway.

    You correctly point out that one of the justifications used by insurers+government for TRIA is that terrorism is uninsurable because it cant be predicted well enough to price insurance. However, it is a mistake to think that the historical data on terrorism is good enough to do the job.

    Terrorism data is incomplete, politically manipulated in some cases, and of lower quality/quantity than for natural disasters, fires, etc. Terrorist attacks just don’t happen as often and on as big of a scale as all that other stuff to produce enough data.

    Even if the quantity and quality of terrorism data were good enough, terrorism changes faster than the data. The “best data” is held by the government as classified information which insurers do NOT have access to. Terrorism changes – terrorists adapt and change and new terrorist movements pop up. All of this happens without producing data that insurers really have access to. Al Quaeda isn’t like the islamist terrorists of the 80’s and isn’t like domestic terrorists and isn’t like all the other terrorists in the world. None of these groups can be modeled and predicted like a hurricane in the Gulf of Mexico. Historical data is therefore kinda useless.

    You mention that it would take leveling the country “in one fell swoop” to wipe out insurers financially, but thats really not the case. A nuclear weapon in one city would certainly do the trick. Remember that the program (broadly speaking) only kicks in for attacks that are about as damaging as Katrina. Somewhere in between a Katrina scale terrorist attack and nuking NYC, there is the actual break point for the insurance industry. Estimates vary widely.

    I agree with you that letting the insurers pay for the government coverage after the fact is weird and bad. I wish Congress would have passed and Bush would have signed something less dysfunctional in that regard.

    I wouldn’t read too much into this bill as one where democrats and republicans differ. TRIA was passed in 2002 and renewed in 2005 both under Republicans. The current extension that just got signed isn’t really anything new, but I do wish the Dems had used this as a chance to reform this program.

  2. Rick Swedlof says:

    Carroll, thanks for the thoughtful response. Here are some further thoughts…

    Although the government doesn’t pay anything for attacks less than 9/11, private insurance and businesses that partake are still getting a free ride. They pay nothing up front for the right to be protected if a major attack occurs. Why isn’t this like me telling my insurance company, “I won’t pay any premiums for insurance on my house, but if a major fire occurs, I’ll pay you back in installments over time after the loss.”?

    As for the probability of attack, I understand that history may not be a good predictor, but it can offer some guidance in pricing. If insurers are worried that they don’t have enough data, why can’t they just include a risk premium in their pricing? If the price of third party insurance is too high, businesses may choose to go without insurance or self insure. But that’s not what happened. Private insurers didn’t charge too much, they said they didn’t want to be in the market at all. Perhaps this was a public relations move (better to be out of the market than be seen as gouging businesses fearful of terrorist acts), but we don’t know for sure.

    Insurers may be concerned about catastrophically high losses (although I believe that nuclear attacks are excepted out of the policies), but as you note, they lost more in Katrina than they did in 9/11. Major insurers have not gone out of business as a result of these losses. Why? Because they have a portfolio of insurance and other investments. Why can’t they do the same with terrorism?

    Even if the government has to step in to provide temporary help after a major attack. The two renewals of TRIA are beginning to feel like permanent subsidies.

  3. “private insurance and businesses that partake are still getting a free ride. They pay nothing up front for the right to be protected if a major attack occurs.”

    Is this really true? I read your original post to say that insurers are required to offer the coverage, but that doesn’t mean it will be free. It will either be in a rider (separate fee) or in the base insurance (which will surely raise insurance costs for all).

    I see this law as creating something similar to earthquake insurance – as state backed insurance for something that insurers won’t touch. Having owned a house in California for 10 years, I can assure you that it isn’t free.

  4. BDG says:

    Hi, Rick. Interesting post. In Florida we have similar issues with hurricane and flood insurance. I suppose one principled defense of subsidies for all forms of disaster insurance (whether on the insurer or insured side) is that individuals irrationally underestimate the risk of disastrous losses, so that, at actuarially fair prices, they will underinsure.

    On the other, the public choice explanation for large, hidden subsidies for insurers is pretty powerful, which leads one to be suspicious of efforts at principled explanations. For instance, before endorsing the underinsurance story, I’d want to know just how much, if any, of the subsidy is passed on to consumers.

  5. Rick Swedloff says:

    Michael and BDG,

    Thanks for the input. I don’t know that I have staked out a particularly firm position on this, but it is an interesting piece of legislation. Your posts, taken together, illustrate why.

    First, it seems relatively clear that the insurance industry is getting a subsidy here. Insurers don’t pay a premium for the reinsurance they receive from the government. The subsidy is most obvious if an attack never occurs. In that case, the insurers have received premiums from their insureds for as long as the policy is in force, have carried only portion of the risk because of government reinsurance, and have never had to pay a premium for the reinsurance they receive. In other words, the government is carrying part of the insurers’ risk, but the insurer isn’t paying the government. If any of this subsidy is passed on to businesses in the form of lower premiums, the businesses also get a subsidy, even if, as in the case of Michael’s home owner’s insurance, the rates are quite high.

    Second, there may be solid policy reasons for government to enter into this market or not. I’m not sure. Having lived a good chunk of my life in Hurricane Alley in the Carolinas, I am sympathetic to the cost of disaster insurance and I like BDG’s defense of some subsidy for individuals. But I also agree with his concerns. And further when we talk about TRIA we aren’t talking about individuals, we are talking about corporations. That, to me, is a different animal altogether.

    Lastly, I find it striking that the government passed this enormous subsidy for the insurance industry and possibly businesses, but in the same month failed to expand the SCHIP program. If we are concerned about the macroeconomic impact of no terrorism insurance on businesses why aren’t we as concerned with the macroeconomic impact of medical care for children. This last thought is, of course, vastly underdeveloped, but it did give me some pause.