The Insurance Industry’s Response to Anticipated Climate Change

My mom often sends me articles to read.  She just sent me this piece  about new disclosure regulation intended to nudge insurers into revealing how they are updating insurance pricing and expected probabilities of nasty future events (i.e severe floods) caused by climate change.  This is a nice example of the small ball of climate change adaptation but must government be involved here?   In this cross-post, I discuss why anticipated ex-post moral hazard leads me to say “yes”.  In a nutshell,  the optimistic insurers will charge lower premiums and attract more premium buyers than insurers who are more pessimistic about climate risk.  When Mother Nature strikes, the optimistic insurers will face huge claims bills and will go broke and the Federal Government will bail them out.  This moral hazard is anticipated and rewards the optimists for being too optimistic.

Author: Matthew E. Kahn

Professor of Economics at UCLA.

23 thoughts on “The Insurance Industry’s Response to Anticipated Climate Change”

  1. I don’t understand the logic here. The Feds will not bail out the optimistic insurers; it will bail out their customers. The customers have no freakin’ idea whether the optimistic insurers have decent actuaries; they just look to price and maybe claims-handling. (Some customers look to credit rating in life insurance, but I doubt that consumers ever look to credit for P&C.)

    More complications: insurers have something like FDIC insurance, funded by retrospective assessments on survivors. Insurers also issue bonds, purchase reinsurance, and live and die by their credit rating. I’m never sure about bondholders, but reinsurers care tremendously about pricing cat risk right.

    I’m not arguing that regulation is unnecessary, although many people in the insurance business view credit raters as if they were the true regulators. I’m very skeptical, however, about economists who start economizing before they learn institutional details. That’s why I prefer lawyers in a public policy role–maybe they can’t model worth shit, but they’re very seldom tempted to model complex systems as uniform spheres.

    1. The customers have no freakin’ idea whether the optimistic insurers have decent actuaries; they just look to price and maybe claims-handling.

      Well, yes. And customers choosing based upon price without regard to a insurer’s actuarial capabilities is the problem. Given that it is very difficult for most customers to evaluate actuarial models even if they had the raw data, this is a problem that is both bad and inevitable. It takes regulation from somewhere to deal with it.

      And I can’t share your enthusiasm for putting a bunch of people who have no concept of modeling in charge. Yes, they aren’t likely to model something as a uniform sphere, but that’s mostly because they’d prefer to take a random, uninformed guess. Economists may be the problem, but lawyers aren’t much of a solution.

      1. 1. “It takes regulation from somewhere to deal with it.” If you define “regulation” loosely enough, I concur completely with this statement. But there are a number of for-profit gatekeepers in the insurance business, especially the ratings agencies and the reinsurers. Are they more effective than the insurance departments? Maybe. Is the prudential regulation provided by the insurance departments necessary? I’m not sure. (This is not an argument against insurance regulators–they have plenty of work to do with sleazy marketing and claims-handling practices, fraudulent insurance, and the like. It is only mild skepticism about insurance prudential regulation, at least as practiced by the governmental sector.)
        2. I’m not sure you have the same mental image of lawyers as I do. I’m not talking about silver-tongued dolts. I’m talking about reasonably bright guys who immerse themselves in institutional detail, and then rely on their experience and judgment in patterning the details. I’m contrasting them with differently bright people who skip the institutional detail, and prefer a priori modelling. The best of both worlds would be a bright person immersed in institutional detail whose elegant models are consistent with deep practical understanding. But Herbert Simon is dead. Given a choice among the living, I’d rather go with the lawyerly type.

        1. Most of the lawyers I’ve known might look at institutional details, but have no grasp of math. You absolutely, positively, cannot evaluate actuarial policies without quantitative models. It’s a requirement. Maybe you want someone other than economists to build them (they aren’t the only quantitative guys around, after all), but someone has to. Skip the modeling, and you’ll get disaster. Again.

          And speaking of disaster, I can’t believe that anyone in 2012 is holding out the credit ratings agencies as organizations we should rely on for anything of value. They may have a profit motive, but there is zero evidence that they have changed their business models in a way that provides incentives that would be helpful.

  2. I’m very sceptical that the government is better at underwriting than insurance companies. Sounds, quite frankly, like the idea here is to force insurance companies to price as though they believed this global warming stuff, and then later point to the resulting rates, and say, “Look, insurance rates on the coast are rising! More proof you should take global warming seriously!”

    I suppose you need to do things like that, as the ‘pause’ in global warming reaches year 15…

    1. Ebenezer, I point to Brett as an example of what happens if you let someone who doesn’t understand statistics be in charge of this kind of thing.

    2. Part of the problem here is that state-level regulators really hate rate increases, regardless of how obviously the current rates are too low. (Florida is the most notorious state for homeowners insurance.) It’s probably most accurate to think of this as a trying to get regulations to cancel out problem.

      1. I would say that the problem is a bit different than a refusal to raise premiums. It’s more that we don’t charge more to people who build homes in riskier areas. The general public has trouble accepting the idea that they should pay more to live somewhere that is prone to floods/wildfires/other natural disasters.

        The question we need to answer is whether or not we, as a society, think that the cost to insure a home should be commensurate with the risks, or whether we wish to subsidize the ability to live somewhere riskier. Strictly from a risk management perspective, it’s exactly the same question as whether or not people with pre-existing medical conditions should have to pay higher rates for health insurance. They are riskier to insure, and so everyone else has to subsidize their health insurance if we want it to be affordable for those individuals.

        I come down on the opposite of the homeowners’ insurance question as I do on health insurance. Not only do I not think that we should subsidize living in higher risk areas, I think that there are generally reasons why we ought to discourage living in most of those places. Southern California is really a semi-desert with serious water issues. The Gulf coastline is a fragile ecosystem that would benefit from less human interference. I’m sure there are exceptions, but I think charging market premiums for flood and wildfire insurance is not only a good idea, but it’s more of a good start in providing incentives rather than an end state.

    3. This is, of course, a lie. The most charitable I can be is that Brett is as usual getting his news from right-wing nutball sources that he doesn’t understand are lying to him, but it’s a lie nonetheless.

  3. I know enough about statistics to know that “Don’t you understand statistics?” was a lot more reasonable question to be asking when the models had been wrong for five years, than it is when they’ve been wrong for fifteen. Give it another five years, and if you ask that question, expect the answer to be, “Yes. Do you?”

    1. Brett, you’re projecting again. The people who’ve been wrong about global warming are right-wingers like you.

    2. Brett, you’re playing Fun With Endpoints, and it exposes you as someone who is either dishonest or clueless about statistics. You can’t deliberately start your data series with an obvious outlier and then perform regression on it. End of story.

  4. I’m not sure I buy this.

    Many types of insurance have the characteristic that most of the time, the premiums collected by the insurance company are enough to pay out all claims, all operating expenses, and still have enough money left over to provide the insurance company with a healthy profit. Every once in a while, though, something happens causing a lot of claims to come in at once, and the company loses money.

    There are two strategies which insurance companies can use to deal with this. One is to set up a reserve fund which is large enough to cover the losses. Setting up a reserve fund requires raising a bunch of money from investors, who will want a healthy return in exchange for putting their money at risk, so much of the profits that the insurance company makes in normal times (when the premiums collected exceed the cost of claims plus the cost of running the business) will go to the investors who funded the reserve fund.

    The second strategy is to not do a reserve fund, make a lot of profits for a while, and then go bankrupt. This is the strategy adopted by AIG when it got into the credit default swap business. (A credit default swap is an insurance contract that pays out if a bond issuer defaults.) As a result, for a while AIG showed large profits on its credit default swap business, and paid out huge bonuses to the people working in the credit default swap unit. Then there was an economic downturn, meaning more bond issuers defaulted, and AIG went bankrupt (or would have if the government hadn’t bailed the company out).

    It is not clear to me that the prospect of being bailed out by the government played a major role in the decision making at AIG. It seems more likely that the prospect of large bonuses was the major factor in employee decision making. AIG could sell a handful of credit default swaps without setting up a reserve fund specificly for that purpose, because the general assets of the corporation would be sufficient to cover any losses. At some point, the business grew to the point where AIG should have set up a reserve fund, but who is going to suggest that? If you set up a corporate culture which centers around huge bonuses, ideas that would make everybody’s bonuses smaller don’t gain traction. I don’t think anybody at AIG set out to but taxpayers on the hook; they were simply chasing after bonuses, and managed to convince themselves that the bonuses were deserved, or if they didn’t quite believe that, at least they weren’t inclined to think too hard about why the bonuses might not be deserved.

    Therefore, I think that Matthew somewhat misses the mark when he talks about the moral hazard caused by the government bailing out insurance companies. I think it is likely that the bigger problem is that employees don’t necessarily act in the long term interest of the company both because (1) people tend to think short term, at least until they’ve been working in the business long enough to have personally been bitten by ignoring the long term, and (2) the long term interests of employees don’t conincide with the long term interests of the company because the employees may be long gone before the bad things happen and even if they are still there they won’t have to pay back their bonuses from earlier years. Either way, Matthew is correct that government has a role to play to prevent problems.

    The problems become less if we are talking about short term insurance contracts. Global warming makes certain types of disasters more likely, but does it require insurance companies to keep larger reserve funds? If an insurance company writes a policy which covers all hurricane damage over the next ten years, then global warming places the insurance company at risk because there are likely to be more hurricanes than in the past. But if the company writes one year policies, then the increased frequency doesn’t matter as much. In years when there is a major hurricane the company taps its reserve fund, just like it would without global warming. As the frequency of hurricanes increases the company must raise rates to remain profitable. If the insurance company falls behind the curve in raising rates, that’s bad for its stockholders, but it shouldn’t cause problems for policy holders as long as the insurance company maintains an adequate reserve fund. So in my view, the proposed disclosure regulations would make sense if they were being proposed by the SEC to protect shareholders, but I don’t see the case for insurance regulators to issue them.

    1. Ken,
      What you are talking about is known as “catastrophic risk” (or “cat risk”) in the insurance industry. These risks are pretty much those for which reserving is inadequate, because of the structure of capital markets. So reserving, pretty much by definition, cannot handle cat risk. However, many two strategies remain:
      1. Reinsurance, which is the most widely-used strategy
      2. Subsidiary compartmentalization of cat risk–let the cat risky part of the insurer go broke, while not touching the rest of the company.
      3. Cat bonds (which are to reinsurance as CDS is to selling bonds–the risk might happen without triggering the cat bond)
      4. Underwriting (and you wondered why your life insurance policy excludes war risks!)
      5. The government.
      In the end, there is no way to avoid #5, even with ideal regulation. Some risks swamp all capital markets, such as war risk. This is the insight behind Price-Anderson, for example, or FEMA.

      A ten-year hurricane policy probably leads to *less* cat risk than a one-year policy, because you have a ten-year income stream over which to average your payouts–a way of deepening the capital markets. It leads to greater underwriting risk, for the reasons you set forth. But this isn’t cat risk.

      1. That makes sense. What I’m saying is that global warming is shifting the frequency of events enough to cause underwriting risk, but not enough to create new types of catastrophic risk in the near future.

      2. Re 1: The biggest reinsurers – Munich Re and Swiss Re – are solidly in the climate realist camp. While insurance companies, may, as per Matthew’s scenario, deliberately underprice climate change risks, they are unlikely to do so out of pie-eyed denialism when their reinsurance options are correctly priced.
        I’m surprised at the (American?) assumption that insurers will in the end be bailed out. SFIK the underwriting syndicates that make up Lloyd’s of London have never in 323 years been bailed out by the British government, nor have they defaulted on valid claims, though a good number of Names have been ruined to achieve this. The syndicates were historically unlimited partnerships, though limited liability membership replaced it after the 1970s-1990s asbestos crisis and Lloyds is less important than it once was.

      3. Reinsurance is of limited value when dealing with a problem as large scale as climate change. Reinsurance works when it is one, or a small fraction of those in existence, insurance company that’s exposed to a catastrophic event. This allows highly concentrated risk to be spread around so that general actuarial principles can work.

        When it’s the entire insurance system that faces high levels of catastrophic risk, the value of reinsurance drops. It doesn’t dilute the risk; it just moves it around. An event like Katrina strains the system. Positing a significant chance of a string of such events would trash it. Reinsurance, per se, does nothing. At that point, it’s value stems solely from increasing the amount of capital invested in insurance. Whether it’s in a primary insurer or a reinsurer doesn’t make any difference. If you can come up with a strategy that would involve more people putting more money into insurance, that is what is needed. That doesn’t involve reinsurance; it involves higher premiums.

    2. ” then global warming places the insurance company at risk because there are likely to be more hurricanes than in the past.”

      It is my understanding that, according to hurricane experts, there isn’t actually any basis for linking global warming to increased numbers of hurricanes. It’s just a convenient boogy man to wave around.

      1. Yes, my mistake. Models indicate that global warming will cause hurricanes to become more intense, but the effect on frequency is less clear. I should have written that there are projected to be more category 4 and 5 hurricanes.

  5. I think that examining the insurance industry behaviors after Hurricanes Andrew and Katrina would be enlightening to y’all. In Florida, prior to Andrew, virtually all of the insurance companies chose to underprice their risks in order to get sales. In the aftermath, many smaller insurance companies went out of business, and the larger insurance companies wanted to abandon the business of insuring buildings from hurricane damage. Fortunately for Floridians, the insurance commissioner had some balls and said that any insurance company that abandons the home insurance line would lose their license to sell auto insurance in the state. So the home owner insurance companies stopped renewing the maximum allowed by regulation, and within a few years it became almost impossible to get private insurance east of Military Trail (one of the important N-S roads just west of I-95), so homeowners were forced into buying from the state insurance pool. This is just another reason I’m glad I don’t live down there anymore.

    My prediction is that insurance companies will fail to charge enough to customers, but just enough for huge bonuses to be paid to mismanagement. We’ll see more insurance companies that won’t cover wind or water damage, or basically anything you want to buy insurance to protect yourself from.

    1. This is the trick. It would all be symmetrical if insurance companies didn’t bay dividends or bonuses based on the size of their book. But as long as the shareholders and managers of the optimists get to extract a bunch more money from the cycle than do the shareholders and managers of the realists and pessimists, you have an unfortunate incentive.

      There’s another side of things that we saw with Katrina: policies sold and billed as covering some customer-attractive set of damages may be administered to cover some much-less-attractive subset that just happens to fit within the companies’ financial model.

  6. The Florida case is a case-in-point of what I mentioned above. The insurance commissioner insisted that insurance companies sell at prices that didn’t come close to covering the expected losses. Requiring them to publicize that “these prices are supported by no evidence whatsoever” would at least make him look even more culpable.

Comments are closed.