Insurance has been raised as an institutional response to the threats of climate change. The Framework Convention on Climate Change (FCCC) reserves it as one of the options the parties are to consider. Yet, one finds that "insurance" has been used in the greenhouse debates to connote distinct strategies. In its most robust sense, "insurance" is applied to any precautionary undertaking. For example, there is a popular sense in which one who installs a home security system is "insuring" against robbery. If the cost of the system exceeds the expected loss, the difference might be referred to as a "premium." Some commentators on the greenhouse effect seek to justify precautionary investments in these terms. The investments are designated social "premiums" warranted to prevent low probability, highly undesirable outcomes. In its more technical usage, however, insurance applies to arrangements designed to avoid  outcomes, so much as to distribute  the costs of the losses that other measures fail to avoid. Both meanings of "insurance" merit consideration, although the focus of Stone's research has been on the more technical sense of the term of spreading the burden of possible losses among risk-holders.

Stone does not tout technical insurance as a first-line response. He outlined carbon-specific and general policies which merited priority. In terms of carbon-specific or greenhouse gas-specific policies, Stone mentioned "no regrets" strategies, such as improving energy efficiency, which make economic sense from many points of view, but which are often not adopted due to a variety of institutional barriers to their implementation. General policies are those that seek to "insure" future persons against a whole range of perils that may befall them, whether climate-driven or otherwise. The best "insurance" flexibility is added global wealth. In addition, there are promising non-monetary forms of insurance, such as investment in portfolios of real assets, such as germ plasm banks. Even more important, we should provide the insurance of institutional flexibility, for example, education, and international institutions that provide relief and assure free movement of capital, people, and goods.

Conventionally, techniques for managing risk are divided into five categories: avoidance, reduction, control, transfer, and retention, all of which may be relevant to dealing with climate change. A risk managers aim is to combine them into the lowest cost combination.

Turning this model to global warming, Stone reminded that it is not "climate change" one would be insuring against, but particular local manifestations, such as floods, drought, crop loss, etc. In law, he reminded, insurance in this technical sense plays a distinctly secondary role. Ex post,  that is, after the fact strategies are basically harm-based liability rules which provide compensation for victims in the event of an accident. This has the effect of reminding the actor that he is acting under the shadow of potential future loss. Ex ante,  or before the fact strategies usually depend on standards, as in the nuclear safety field. Instead of waiting for damage to occur and then having the actor compensate the victims, society decides that there are certain things we must seek to prevent before they happen. This displaces the judgments of managers with societal judgments. Both ex post  and ex ante  strategies are precautionary and subject to over-regulation.

Stone recited certain benefits of exploring the insurance option, which has been relatively dormant. First, insurance fits naturally with the most common expectations about future climate-change: that there will be some hard-to-predict winners and some hard-to-predict losers from climate-driven phenomenon. These are circumstances that invite commercial insurance. Indeed, insurance companies and reinsurers are already implicated (and concerned) just from the continuation of their calamity underwriting. Insurance promises to compensate risk bearers for risks that cannot be efficiently reduced further. But there is also an element of efficient deterrence. Insurance sensitizes risk bearers to the social costs of their own conduct via the premiums they bear, just as proposed carbon taxes sensitize risk creators. The behavior of both is nudged in the right direction. Moreover, the insurance industry has already assembled considerable expertise in the assessment and distribution of many of the anticipated climate-driven losses.

But there are serious questions about how strongly we can rely on commercial insurance (and other market mechanisms) to deal with problems as potentially vast as climate change may bring. The insurance industry has limited resources, and calamity underwriters appear to have been taking a beating in recent years. Even its statisticians are uneasy about their ability to assess climate change-driven losses. Their exposure has increased as more people have sought calamity insurance, and more wealth has moved into dangerously situated regions coastal zones in particular. Over the long periods we are considering, they face appreciable problems of adverse selection and moral hazard. Moreover, there are many "gaps" in commercial coverage. Normally, insurance covers sudden catastrophes, not gradual erosion of land or gradual loss of farmland. Also, public goods like global "commons" areas, are almost certainly beyond the range of coverage that can be made commercially available. How will future courts react to insurers who, after collecting premiums for 30 years, see which way the wind is blowing and want to cut off coverage? Public policy will resist and in anticipation of this backlash, the industry will be all the more leery about underwriting the losses.

One of the most difficult aspects of applying the insurance theme to climate change is the subject of attribution. Consider an event comparable to the recent major flood in the Philippines. Is it caused or exacerbated by human-caused climate change? Or was the "real" cause human deforestation? The difficulty of separating out a variety of possible causes and determining the proximate cause will always be challenging. This, Stone says, is one reason insurance may make more sense for climate cataclysmic events than familiar legal compensation. A contract can be written to provide that the holder of a first party insurance contract has less of a burden that a plaintiff in a civil suit who or what caused the loss. One doesn't insure for "floods caused by climate change;" one simply insures against floods, so collecting is not dependent on establishing causality.

Considering, however, the limits of the coverage the private industry is likely to provide, the question arises of government participation in climate insurance. Such participation could take a number of forms. The government could provide a subsidy for the premium payers. Alternatively, it could provide backstop coverage to fill coverage gaps (like the Price-Anderson Act does for the nuclear power industry). A tax on greenhouse gas emissions could supply the money for this backstop insurance fund. Government can also participate in more direct ways, including managing a pool to cover losses (as contemplated by the Association of Small Island States), or by insuring rejected applicants.