Insurability of damage caused by climate change: a commentary.

• Author: Faure, Michael G.
• Position: Response to article by Howard C. Kunreuther and Erwann O. Michel-Kerjan in this issue, p. 1797

INTRODUCTION

The article by Howard Kunreuther and Erwann Michel-Kerjan provides a powerful overview of changes in extreme weather-related events. Supported by empirical evidence, they show that in recent decades insurers have, more than before, been confronted with catastrophic losses resulting from weather-related events (e.g., hurricanes and flooding). Moreover, they also predict that it is very likely that these types of high-damage events will occur even more in the future. The authors' contention can hardly be debated; one could easily cite other studies that equally predict that as a result of climate change it is likely that in the (near) future more of these extreme weather-related events will lead to high losses. Much evidence in this respect is gathered in the studies prepared for the Third Assessment Report (TAR) of the Intergovernmental Panel on Climate Change (IPCC), which holds that it is very likely that in almost all land areas humans will experience higher maximum temperatures and that various extreme weather events can be the expected result. (1) The IPCC predicts heavy rainfalls in particular areas, and other extreme weather events like droughts and cyclones in other areas. (2) Even during the week of this Symposium, the Conference of the Parties to the United Nations Framework Convention on Climate Change (UNFCCC) gathered in Nairobi, (3) where a report was presented that argued that by 2040 it is likely that the damage resulting from climate change might be as high as one trillion dollars annually. (4)

This study merits two comments. First, even though one could argue that these estimates often relate to total losses (which are by definition much higher than the insured losses), it is likely that (as the authors rightly argue) insurers will be confronted with more losses from weather-related events. Moreover, the amount of losses per incident will be higher as well. Second, although there is of course (as many contributors to the Symposium mentioned) still a lot of debate among scientists as to whether the predicted increase in these extreme weather events is actually the result of anthropogenic emissions, this does not matter from an insurance perspective. As Kunreuther and Michel-Kerjan rightly mention, for an insurer it is not directly relevant whether this increase in extreme weather-related events is the result of climate change. (5) All that matters for the insurer is that empirical evidence suggests that an increase in extreme weather events is likely to occur in the future. Insurers and reinsurers have long understood the importance of climate change to their businesses and have devoted various studies to this topic. (6)

The central question discussed by Kunreuther and Michel-Kerjan is under what conditions insurers can play a role in covering the damage caused by these extreme weather-related events (whether or not they are caused by climate change). (7) To answer this question, the authors focus on the central notion of insurability. They provide an overview of the classic criteria that are advanced in the literature to judge the insurability of certain risks. In my comments below, a few of these insurability criteria will be discussed in further detail, a few will be added, and Kunreuther and Michel-Kerjan's analysis will be extended by addressing some of its policy implications. Indeed, in their concluding remarks, the authors briefly address (mainly as a point for further research) the question of how policy should react to these threats to the insurability of damage caused by climate change--more particularly, they suggest a role for the government.

In this Commentary, I will suggest that it is difficult to discuss the insurability of the catastrophic losses that potentially will be caused by climate change without addressing these policy issues. The nature of the damage caused by climate change is indeed such that it poses serious problems both on the demand side as well on the supply side, as I argue in Parts IV and V. Hence, it may not be possible to judge the insurability of those risks without addressing the question of whether some intervention at the policy level is necessary to facilitate this insurability. Before addressing these more fundamental issues, I would first like to focus on two classic insurability conditions, which are of particularly more importance for judging the insurability of liability for climate change:" (1) retroactive liability and (2) causal uncertainty.

1. RETROACTIVE LIABILITY

   1. Efficiency

      A classic source of worry for insurers, particularly in the area of environmental liability, has been the retroactive character of liability rules. In some cases, legislators explicitly opted for a retroactive liability regime; (9) in other cases, judges held that a particular behavior (like dumping toxic waste) was already considered wrongful at the moment when the act happened (e.g., twenty-five years ago) even though it may be doubtful that this actually was the case. (10) A retroactive liability rule has been heavily opposed in the law and economics literature for the simple reason that a finding of liability can only affect future behavior. (11) Thus, retroactive liability can never have any beneficial effect (in the sense of providing incentives for prevention) for a behavior that already occurred in the past. (12) If one were to consistently reject retroactive liability, however, and hold an injurer to comply merely with an "old" standard of care, the problem would arise that standards could never be adapted to changes in technology. Therefore, it has been argued that the fact that there may be liability ex post may give incentives to obtain information about risk to industrial operators. (13) In order to have a dynamic adaptation of liability standards to evolutions in technology, a system of so-called "prospective overruling" has been suggested. This means that a new liability standard would only apply to future cases, but not to the current case that raises the issue before a court. This would still allow for a dynamic process of learning by courts whereby the standard of care is not static, but changes over time. (14)

      In the case of liability for climate change damage, retroactivity may be a serious problem. (15) If one is already willing to accept a causal relationship between anthropogenic emissions and damage caused by climate change, the problem will arise that many of these emissions may have taken place in the distant past, when information about such a causal link was lacking. Of course, it is debated whether information about (or suspicion of) such a causal link was available. Some reports suggest that even as early as the beginning of the nineteenth century there was already scientific evidence available on the effect of emissions by industry on the climate. (16) From a legal perspective, however, the relevant question is whether at that time (even assuming that information on the causal link was indeed available) emissions would already have been wrongful in order to justify an action in tort. l; The precise time of that moment is debated. Roda Verheyen argues that since the IPCC published its first report in 1990, states have foreseen that greenhouse gas (GHG) emissions create a risk of damage. She argues that one could even hold that governments should have been aware of such a risk much earlier--in 1950 or even before. (18)

      Whether one also wishes to accept liability for historic emissions will to a large extent depend upon the goals one wishes to achieve with tort law. (19) If deterrence were the only goal, a retroactive liability rule may make little sense. If, however, one accepts other goals of tort law, (20) including distributional justice and compensation, liability even for historic emissions could become meaningful again.

   2. Insurability

      The assertion that climate change damage (at least partially) may have begun in the distant past, when C[O.sub.2] emissions originated, may have some bearing on the liability issue, but may not necessarily prevent climate change liability. The issue that Kunreuther and Michel-Kerjan address in their article is the insurability of climate change liability. No matter how one views the desirability of retroactive liability for climate change damage itself, we have to recall that retroactive liability is often considered to endanger the insurability of particular risks. (21) If insurers are confronted ex post with their insured's liability for an act that was not wrongful at the time when it was committed, insurability is endangered for the following reasons: (1) no premium could be asked for this risk, (2) no reserves could be set aside, and (3) no preventive measures could be required. (22)

      Therefore, the conclusion from this literature is straightforward: if a type of liability for climate change damage (23) were to be accepted, it would be desirable from an insurability perspective to limit this liability. The liability should then be limited more particularly (assuming

      that this is technically possible) to the damage caused from the moment that the emitters knew or should have known that their emissions could have caused climate change, and thus that there was wrongfulness in tort. (24) From that moment insurers could also have required a premium, set aside a reserve, and instituted preventive measures to reduce the risk. Limiting the liability to the moment that emitters (25) knowingly contributed to climate change allows liability to help prevent climate change and increases the insurability of climate change damage.

2. CAUSAL UNCERTAINTY

   1. Possible Solutions

      Another classic condition for insurability of (environmental) liability is that the insured injurer is only held to compensate for the damage caused by her activity. Kunreuther and Michel-Kerjan rightly argue that this problem of causation will be a major issue when assessing climate change liability. Uncertainty...