Cost/benefit analysis of government policy requires policy economists who evaluate the policy to take strong stands on how they quantify the costs and benefits of a proposed policy. Both the benefits and the costs of a policy that have not been implemented yet are random variables. Consider the case of judging the merits of policies that intend to reduce greenhouse greenhouse gas emissions. Such policies impose costs (such as higher fossil fuel prices) but they offer future benefits. Economists have tried to quantify these benefits using the ugly name of the "social cost of carbon". Adaptation efforts will mean that this social cost of carbon could actually decline over time.
Consider two case:
Case #1: Global GHG concentrations (which now stand at roughly 400 ppm) continue to rise and we make no progress in adapting to climate change impacts. In this case, the social cost of an extra ton of carbon is rising over time because the global economic damage imposed by rising GHG concentrations is a convex function.
Case #2: Global GHG concentrations rise at a dampening rate because of the substitution from coal to natural gas and renewables and the rise of electric vehicle use and at the same time our cities around the world make progress in adapting to climate change and endogenous technological change further helps us to adapt (i.e more efficient cheaper air conditioning). In this case the social cost of an extra ton of carbon is declining over time.
How will the researchers who wrote this 2014 Science piece incorporate Case #2 into their analysis?
By "linearizing" the problem and declaring a single number (i.e a constant marginal damage of $35 per ton), the authors are making many implicit assumptions that I would hope they will be explicit about.
The key point here is that the Social Cost of Carbon hinges on what are the marginal benefits of reducing carbon by a ton. Given that the benefits function is a highly non-linear function of many random variables; a universal "constant" (i.e some physics parameter) is not going to emerge. Yet, policy analysts are eager to collapse this problem down to a single number (such as $35 a ton). Successful adaptation would mean that the marginal benefits from reducing a ton of carbon decline.