Whether and to what extent we are adapting to climate change is a critical question for determining how we should respond to future climate change. At a macro level the more we can adapt to reduce the damages of climate change the more time we may have to reduce emissions. At a micro level the extent to which we have already successfully adapted in some locations may help guide decisions in places that remain less well adapted. In particular, whether existing adaptation has been driven by public or private decisions is especially important in guiding the role of policy in this process.
A new paper by researchers at Stanford takes on both the question of whether we are adapting to climate change and, indirectly, how we should measure such adaptation. Their conclusions with respect to the first question are less optimistic than what many economists might expect. But I’d like to discuss the methodological issues they raise in the discussion of how we measure adaptation first.
What is adaptation to climate change?
This is an apparently straightforward question with a complicated answer. The IPCC, the authors of the new paper, and economists broadly speaking tend to think of climate adaptation “as any response that improves an outcome or reduces damage” from climate change. This clearly differentiates adaptation from mitigation, which is generally considered to be actions that reduce emissions that contribute to climate change.
That definition contrasts slightly the definition preferred by many who work on disaster response, which distinguishes between adaptation and resilience. Here adaptation is (generally) seen as actions that minimize the consequences of experiencing a shock while resilience is the set of actions that increase the capacity to recover after a shock occurs.
This may seem like a distinction without a difference. After all, the point is to minimize the adverse consequences of climate change. Does it matter if that occurs because the damages of a particular shock are minimized or the recovery after those damages occur is easier? If the goal is minimizing welfare consequences to those that suffer the shock these distinctions may not matter. But from the social planner’s perspective there may be reasons to differentiate between these responses.
Consider insurance, something often given as an example of climate adaptation. If the ‘outcome’ under consideration is welfare then insurance would satisfy the economist’s definition of adaptation: paid out after a shock it can make those that suffer the shock whole and improves the outcome. In the adaptation/resilience framework insurance would be considered resilience building because it facilitates recovery but does not directly reduce damages.
Because it facilitates recovery, insurance has also been shown to reduce investments in actions that could reduce damages from shocks before they occur. That illustrates an important point: encouraging resilience or adaptation is not always complementary and choosing one may work at cross-purposes to achieving the other. It may, therefore, be useful to distinguish between these concepts to help clarify when it is preferable to pursue policies that reduce damages before shocks versus help facilitate recovery afterwards.
Regardless of whether a distinction is made between adaptation and resilience, the authors make the important point that ‘adaptive’ actions need not be explicitly related to climate change. Increasing access to healthcare can significantly reduce the mortality effects of extreme heat while improvements in the security of property rights can facilitate investment in irrigation that reduces crop sensitivity despite neither of these being directly related to climate change.
How should we measure adaptation?
Whatever set of actions is included in your preferred definition of adaptation, any assessment of whether it is occurring requires the choice of how to measure it. The authors choose to measure adaptation as the change over time in the sensitivity of a variety of economically relevant outcomes to a standard climate shock. Reductions in these sensitivities are taken as evidence that adaptation is occurring.
They do not take a stand on the adaptations that might be driving the reduction in the sensitivity of an outcome to any particular climate shock. But the approach they take – of directly measuring changes in sensitivities to a given shock – offers a blueprint for studying specific adaptations. It highlights a point that has often been overlooked in past studies that purport to study adaptations: the assessment of adaptations should assess how they change the climate sensitivity of an outcome, not just the outcome itself.
Take an example from an area in which I’ve worked, the impact of heat on student performance. Air conditioning (AC) is an obvious example of an adaptation that might help reduce the adverse effects of heat for students. But an attempt to empirically assess the scope for AC to adapt to these adverse effects should do two things: (1) it should estimate the sensitivity of student performance to heat and (2) it should estimate how AC changes this sensitivity. To assess adaptation it is not enough to show the impact of AC on student performance directly (although that might be independently interesting).
The authors discuss in the detail the (sometimes complicated) mechanics of estimating whether the sensitivity of a particular outcome has changed over time. For anyone interested in empirical examinations of adaptation that section is worth reading closely. But there is one point I would like to highlight. For many outcomes we expect that there is a non-linear relationship between climate change and the outcome. In these cases the change in sensitivity should be assessed at multiple points in the distribution, not only (for example) at mean temperatures. Because of the non-linearities, and the nature of adaptation, it may be that we are adapting at particular points in the distribution but not others. Adaptive responses might not be readily observable if only a single point of the distribution is examined.
Are we adapting to climate change?
The evidence on whether we are adapting is mixed. The authors find that the climate sensitivity of some of the examined outcomes – primarily mortality and agricultural yields – is declining over time. That suggests that adaptation is occurring. But this is true in a minority of the outcomes, fewer than 30%, they study. In a nearly equal number it appears the sensitivity to climate change is increasing. And for nearly half the outcomes they study the evidence is too unclear to be conclusive.
An optimistic read of this evidence is that we are adapting in some areas and could be adapting in many more. That suggests that the technology or policies exist to facilitate adaptation in at least some places for some risks and could be applied in others. A pessimistic view is that climate change is occurring, appears to be causing increasing amounts of damage, and we are not adapting to the majority of hazards studied.
Either view underlines the vast scope for more research in this area. Understanding what has worked in the areas in which we have adapted and how that can be applied in other settings would help continue the progress already made. Examining the obstacles to adaptation in areas in which we appear not to have adapted is equally important for facilitating progress in those areas or motivating faster mitigation efforts if progress on adaptation cannot be made.
Join the Conversation