Climate Change and the Insurance Industry

http://flickr.com/photos/that_chrysler_guy/8139133299/

As the entire eastern seaboard slowly recovers from its lashing by Sandy, insurance companies are bracing for the hurricane’s aftermath and the possibility of another Katrina-scale loss.  If there’s any major incumbent business with an incentive to publicly acknowledge the risks and costs of climate change, it’s the insurance industry, and especially the re-insurers — mega-corps that backstop individual insurance companies by pooling their risks globally.  These companies can do the math, and what they’ve seen over the last couple of decades is a steady upward trend in both the number of extreme weather events and the resulting insured losses that they’ve been on the hook to cover.  The situation is well summarized in a new report from Ceres, entitled Stormy Futures for U.S. Property/Casualty Insurers.  They suggest that insurers face an existential risk from climate change.

What is insurance, anyway?

Insurance companies are actually investing companies that use your premiums (or “float”) as their capital.  Most people imagine that insurance companies make money by charging more in premiums than they have to pay out in claims — this is called the “underwriting business” — and some insurers do end up making money this way, but historically they’ve been willing to take a small loss in the underwriting business, because they get to invest the float for their own benefit until it has to be paid out as claims.  If you want an entertaining, eloquent outline of the insurance industry, read pages 8-10 of Warren Buffet’s shareholder letter for 2011.  His company Berkshire Hathaway alone sits on a $70 billion mountain of float.

Insurance is a great investment financing model because the cost of capital in the form of insurance premiums is very reasonable — even negative in Buffet’s case.  Additionally, expected underwriting profits/losses are uncorrelated with market returns, so the two potential profit/loss streams diversify each other, reducing an insurance company’s risk exposure.  Or at least, this used to be how it worked.

If you imagine an insurance company that only sells insurance in Florida, and also only invests its float in Florida businesses.  What happens when Florida is demolished by an unusually destructive hurricane?  The company simultaneously has to pay out a lot of claims to its customers while big chunks of its investment portfolio are tanking.  This might make for a very unpleasant quarterly report to shareholders.  If this happened year after year, the company would quickly eat through its capital reserves (the float), eroding its ability to either pay claims or make investment returns in the future.

The Ceres report hints at a similar scenario on a global scale — where climate-driven insurance losses increase, and the global economy is stunted by the impacts of climate change and the resulting instability: an uninsured, and uninsurable world.  Despite what you might think from listening to the dialogue on climate in the US, this is not a fringe position within the insurance world.  Many large, historically conservative financial institutions have taken positions on this threat that might sound “extreme” in the context of US politics.

Insuring the Future

Just a couple of weeks ago German re-insurance giant Munich Re published a report entitled Severe Weather in North America, unfortunately the whole thing isn’t available online, but Climate Progress blogged about it recently, and the Ceres report mentioned above cites it extensively.  The company concludes:

Climate­-driven changes are already evident over the last few decades for severe thunderstorms, for heavy precipitation and flash flood­ing, for hurricane activity, and for heatwave, drought and wild­-fire dynamics in parts of North America.

Partly, they attribute the increase in natural disasters to climate change (and not some kind of reporting bias) because the frequency of geophysical events (earthquakes, volcanoes, tsunamis) has not increased since 1980, while the rate of disasters resulting from extreme weather in North America have roughly quintupled:

An upward trend also exists in the total economic impact of those events, though it’s not as obvious:

Another enormous re-insurance company, Swiss Re, collaborated with McKinsey consulting as part of the Economics of Climate Adaptation Working Group, in putting together a suite of case studies called Shaping Climate-Resilient Development (PDF) exploring adaptation costs and strategies in areas as diverse as the Mopti region of Mali, (which may be subsumed by the Sahara), to Maharashtra, India (where subsistence agriculture feeding tens of millions is susceptible to increased drought frequency), to southern Florida (which faces higher storm surges due to rising sea level and increased storm intensity).

The International Monetary Fund is a strong advocate for carbon taxes or some kind of cap-and-trade scheme, because their costs are both lower and more predictable than those of unmitigated climate change.  The World Bank published a white paper this fall entitled Investment decision making under deep uncertainty — the application to climate change, arguing that we should move away from seeking efficient, optimal investments that do very well in the environment they were designed for, but which are brittle and fail if our predictions are inaccurate, and toward robust, resilient investments that can succeed under many different scenarios.

These organizations and companies are confident that the financial risks associated with climate-driven natural disasters are increasing.  They recognize that in many cases their businesses and the financial solvency of nations will depend on accurate assessments of those risks, and that in this context history is no longer the most effective guide to the future.  Large insurers are developing in-house climate modeling groups (Swiss Re is a perennial presence at the American Geophysical Union’s meetings, snapping up newly minted PhDs).  Smaller insurers are outsourcing to companies like Risk Management Solutions, which recently predicted that annual average hurricane losses would increase by 40% on the Gulf Coast, Florida, and the Southeast, and by 25-30% on average for the mid-Atlantic and Northeast coastal regions, due to increased storm frequency and intensity, as well as political and economic dynamics that arise from super catastrophic events, like Katrina.

The obvious solution from the insurance industry’s point of view is to raise their premiums.  Assuming they can get a handle on the magnitude of expected future losses, they should be able to figure out how much they need to charge to remain solvent.  This is fundamentally a much more difficult task than the traditional approach of having actuaries examine historical data.  It also happens to be legally challenging in many states, where insurance industry regulators often require rates to be justified on the basis of past losses.  The actuarial difficulty of predicting future losses combined with the political difficulty of charging rates that would be sufficient to cover those losses has resulted in some regions being effectively abandoned by the insurance industry.  Many major insurers have stopped writing policies altogether for hurricane coverage in Florida, leaving the state government responsible for losses.  State subsidized insurance programs were once aimed primarily at low income populations, but today in Florida, they are the primary form of weather related property and casualty insurance.

Unfortunately, governments face exactly the same actuarial and political difficulties as insurance companies in charging appropriate rates.  The Insurance Information Institute reported in July that in the US alone such programs are now exposed to nearly $900 billion in potential losses, with roughly half of those assets being located in Florida.  Their rates do not generally reflect actual risk exposure.  As losses continue to increase, if these governments are unable or unwilling to charge the premiums required to cover the resulting claims, the result will be either bankruptcies or bailouts.  In the meantime, we will be subsidizing new construction and re-investment in vulnerable coastal areas.

Your Money or Your Life

Of course, the insurance industry is only impacted by insured losses, and even in a developed economy like the US, only about half of the losses that result from climate-driven catastrophes end up being covered by insurance.  There are also less tangible financial losses, such as decreased economic productivity in the areas hit by climate-driven catastrophes, and the entirely intangible human costs, measured in misery and tragedy.  Insurance is a way of spreading the financial costs over large pools of risk-takers, and long spans of time.  Often you can translate between the financial and the human tolls of disasters, if you plan ahead.  Building storm or earthquake resistant buildings might cost more up front, but you only have look at the death tolls from similar disasters in different countries to see what those investments end up being worth.

Broken bridge

Disasters in wealthy nations tend to kill few people, and be very expensive.  The same events in poorer places are cheap and deadly.  The 2011 Tōhoku earthquake and tsunami in Japan killed 16,000 people and cost $235 billion, while the 2004 Andaman earthquake and tsunami in Indonesia killed 250,000 people, but cost only $10 billion (according to RMS).  In 2005 Hurricane Katrina took fewer than 2000 lives, but cost more than $100 billion, while in 2008 Cyclone Nargis killed 140,000 in Burma, but the economic losses came in at a paltry $10 billion.  Last year’s floods in Queensland, Australia (pictures here) killed less than 50 people, but cost more than $30 billion, while the 2010 floods in Pakistan (pictures here) displaced almost as many people as the entire population of Australia, despite affecting about the same area and having a similar economic cost.

Crop failures in particular are likely to be one of the most serious climate-driven losses.  In the Climate Resilient Development report I mentioned above, the case studies done in Mali, north China, and India all found that drought based crop failures were likely to be the most serious economic losses, affecting the food security of more than 100 million people.  In developed countries we tend to insure crops through our governments at subsidized rates to encourage the stability of the food production system.  Most of the world’s one billion subsistence farmers have no such protection.

Whether you’d prefer to live in a cheap and dangerous world, or an expensive safe one is a question of how much you value human life, and whether you’ve got the wealth to invest in safety, stability, and resilience.

The Costs of Climate Adaptation

What’s happening broadly is that the insurance industry is starting to make some of the costs of climate change adaptation visible.  Their rates have to go up because the risks associated with living on a warmer world are higher — more of our existing real assets will be destroyed each year by climate-driven events, regardless of whether they’re insured or not.  A world with a higher overall loss rate that was stable wouldn’t necessarily be any less profitable for the insurance industry — the overall volume of insurance sold might even go up — but an unstable, changing climate will make it very difficult for companies to accurately assess their risk exposure.  Both the increased loss rates and the increased uncertainty about future loss rates will increase the cost of  insurance.  The ten trillion dollar question is: just how much are we willing to pay?

When the insurance industry is forced to withdraw from a market because nobody will pay the premiums required to cover expected losses, we are in effect saying that we’re not willing to pay those climate adaptation costs.  The rational response if we’re not willing to pay would be to reduce our exposure: to stop building beachfront condos in Florida, to re-locate New Orleans to somewhere that isn’t below sea level, and to tax the hell out of greenhouse gas emissions.  When we instead demand government backed insurance, we’re saying we want somebody else — the taxpayer at large — to pay, but who pays for adaptation doesn’t affect its overall cost.  That cost depends on how violent the weather gets, and how much we choose to expose ourselves to it.  As long as we remain unwilling to make serious mitigation efforts, or adjust our behavior to reduce our exposure, the cost won’t go away.  At best it gets shuffled around, and at worst we create subsidies that encourage investment in vulnerable areas and actually increase the long-term costs of adaptation.

New Orleans, The Day After

When somebody says that the cost of climate mitigation is too high, they’re either comparing it to a fantasy world in which we can continue doing what we’re doing, and not bear the consequences, or they’re imagining that the costs of adaptation are in the distant future, and can thus be discounted heavily.  If we want to adapt gracefully and cost-effectively, and not in a miserable, ad-hoc disaster-relief kind of way, we need to do it preemptively, and that means we need to start investing in adaptation now.  We need to start comparing the modest costs of intelligent adaptation and aggressive mitigation to the high costs of inaction.  We need to start attributing some fraction of the costs of present day climate-driven catastrophes to climate change.  What portion of this year’s wild fires in Colorado and New Mexico, or the crop failures in the US Midwest, Russia and Kazakhstan are due to climate change?  How much of the damage done last night to New York City by Hurricane Sandy is a climate cost?  How much of the $200 billion lost to Gulf Coast hurricanes in 2005 was a climate cost?  What about last year’s floods on the Mississippi?  Or the stoppage of barge trade due to record low water levels on the same river this year?

These costs are here today, they are growing, and they are in some measure attributable to human-caused climate change.  What are we going to do about it?

Uninsured or Uninsurable?

As I see it we have three somewhat independent choices, on whether or not to mitigate, to ruggedize/adapt, and to insure.  I’m an atheist but I gotta say, the serenity prayer seems apropos:

God, grant me the serenity to accept the things I cannot change,
The courage to change the things I can,
And the wisdom to know the difference.

Accepting the things we cannot change in this context is ruggedization/adaptation.  There has already been significant climate change.  There is guaranteed to be more, because the warming effects of the last 50 years worth of emissions have yet to fully show up.  We need to be planning and building with the likely consequences in mind — sea level rise, increased storm energy, more variable precipitation, etc.  As the World Bank paper on decision making under deep uncertainty points out, we need to shift away from optimality and toward resilience.  This shift is desperately needed in our coastal cities, water-stressed regions, and agricultural systems.

Changing what we can is mitigation.  An atmosphere with 1000ppm CO2, and a 6°C rise in surface temperatures are not inevitable, but they do appear to be the current plan.  We need predictably increasing taxes on carbon, cities for people, massive amounts of carbon-free electricity, a different food system, reforestation, closed-loop material resource systems, etc.

The cost of insurance will depend directly on how aggressively we pursue mitigation and adaptation.  Whether we’re willing to accept those insurance costs is a measure of whether we’ve accepted the future consequences of our mitigation and adaptation choices.

We can combine these three choices in various ways.

In the ideal scenario, we would invest heavily in mitigation beginning now, throwing whatever remaining political, financial, and even military weight we have behind the effort.  This is a front-loaded cost, but it would substantially reduce the long-term burdens of adaptation and ruggedization.  If you use a relatively low discount rate (as per the Stern Review) this is by far the most cost-effective option.  Unless we get creative in reducing our risk exposure, resulting insurance rates would still be higher than we were accustomed to in the 20th century, but at least they would eventually stabilize.  Currently nobody on Earth appears to be taking this option very seriously.  Mitigation efforts in the EU and Japan, while leaps and bounds ahead of the US, are still modest relative to what the atmosphere demands.

Oosterschelde Storm Surge Barrier

Effective mitigation is ultimately a global decision, while adaptation is much more localized, so it’s easy to imagine futures in which we fail to mitigate, but aggressively seek to ruggedize and adapt.  In this scenario, insurance costs might be relatively moderate, but the adaptation costs are very high, both financially and in terms of the magnitude of required behavioral change.  For example, Hurricane Sandy’s storm surge was a little less than 3 meters.  Sea level rises of as much as 2 meters by the year 2100 are predicted.  This would mean modest flooding in Lower Manhattan almost every time there’s a new or full moon.  Any storm that happened to coincide with a high tide would flood the streets and subways and tunnels as they are currently constructed.  Would the hundreds of billions (trillions?) of dollars worth of real assets in NYC be abandoned?  Or will we attempt to build a seawall around the city, and occupy it for another century?  After all, without mitigation there’s no reason to think sea level will stop rising in 2100.  The Netherlands has centuries of experience holding back the sea, and they’re investing many billions of euros in upgrading their seawall and dyke system to protect the 70% of their population and urban areas that are below sea level already, and while they emit only 50% as much GHG per capita as the US, they’ve still got nearly twice the emissions of developed-world champions Sweden and Switzerland.  But even those best rich nations still emit about 7 times more GHG than would be consistent with stabilizing the climate.

If we choose instead to neither mitigate nor adapt and ruggedize, then things seem grim.  This appears to be the track the US is currently on.  The result will be a large amount of climate change in combination with vulnerable infrastructure and systems.  Insurance costs — losses — in this scenario would be astronomical.  Our response in Florida suggests that they would be high enough relative to our level of wealth that we would be either unwilling or unable to pay them.  We might be able to fool ourselves for a while with under-capitalized government-backed insurance programs and emergency response agencies designed in the 20th century, but how many $100 billion loss events does it take before that becomes untenable?  Before abandoning coastal cities and expensive, vulnerable infrastructure becomes an imperative, and not a choice?

I’m worried that we in the so-called developed world seem intent on becoming much more like the developing world in terms of our exposure to climate-driven disasters.  We know what it looks like to under-invest in resilience against natural disasters while exposing ourselves to their effects and being unable or unwilling to buy insurance.  It’s cheap and deadly, like the developing world disasters I listed above, that have claimed hundreds of thousands of lives in just the last few years.  It’s a world in which we’ve decided it simply isn’t worth investing to protect human life and prosperity.

The insurance industry is just the messenger.  They’re telling us that our future looks like it’s going to be expensive.  Our choice is whether we’d rather measure the costs of climate change in blood, or treasure, or both.

Leave a Reply