Time discounting and sustainability

Ecological Economics has just published an interesting paper (Climate change economics and discounted utilitarianism) by Ulrich Hampicke. He explains the controversy over time discounting in the context of sustainability and climate change - what sacrifices should we make for the good of future generations? Different choices of time discounting parameters lead to wildly different policy prescriptions. These choices are actually moral choices or perhaps ex-post rationalizations of policy preferences. Hampicke criticizes the moral and empirical soundness of these choices, but his alternative proposal does not yield a practical, quantitative criterion for making decisions. My take is even more disheartening.

Two parameters quantify time discounting, δ and η. δ is the pure time discount rate, which expresses the lower importance we assign to future events or individuals just because they happen to be in the future. The larger δ the less we save.

η (the elasticity of marginal utility relative to consumption) expresses, in Hampicke's words, how quickly people get satisfied as consumption increases. It compares how much the rich and the poor enjoy a little more consumption, and thus measures our willingness to take from the rich to give to the poor, or vice versa. As we will probably earn and consume more in the future than we do now even if we don't save, a large η justifies saving little. Saving would be akin to taking from the poor, making them substantially worse off, to give to the already satisfied rich. η also measures our aversion to risk - why take risks to earn more if more consumption adds little to our satisfaction.

For climate policy, larger δ or η imply fewer sacrifices in the present, and thus a warmer future climate.

The Stern–Nordhaus controversy shows that the “adjustment screws” δ and η have such overwhelming influence on the result of the optimization (and for ensuing policy recommendations) that a new agreement on which numbers to plug in for them would at once invalidate previous results. [...] By tweaking the adjustment screws of the model, researchers can arbitrarily lend support to any climate policy recommendation.

One can take a descriptive view of δ and η, set to empirically find their actual values in the human population, and enter those values in the equations. Or one can take a normative view of δ and η, reckoning that they are moral choices ("we ought to give due importance to future generations; we ought to satisfy ourselves with a frugal life"), and enter the morally good values in the equations. Both finding the values of δ and η that currently guide actual human behavior, and finding the morally superior ones are difficult tasks.

Given the empirical and theoretical difficulties of finding δ and η Hampicke advocates giving up entirely on the undertaking. But the alternative he tentatively proposes (doing everything in order to avoid serious evil for future generations) does not specify a quantitative criterion for making choices:

Under intergenerational fairness, each generation strives for its own happiness while recognizing that its actions and omissions imply consequences for future generations, and takes these into account. And each generation accepts that fairness is reciprocal: We must not burden our successors with morally unjustifiable disadvantages, while, conversely, our successors cannot demand undue sacrifices from us.

There is probably no good solution to the problem of time discounting. This, of course, is no excuse for collective inaction on climate change. Nor is it for collective action.