Climate change and the decision to delay
I went to a debate about climate change a few days ago and, uncharacteristically, decided to take notes of my thoughts throughout the talk. In order that they not be wasted I’ve decided to do a series of posts on some of the interesting points that came up in the course of the seminar. Today’s topic is whether it would be less costly to delay doing something about climate change.
The argument is made that, if we delayed doing something for a while, then it would be cheaper to do it. The reason it would be cheaper is twofold: first, technology would be better and, secondly, the discounted cost of starting in the future is lower than the discounted cost of starting now even though reductions will have to be more dramatic if we start later. Obviously uncertain technological change is a risky thing to bank on; however, the key issue I want to address here is the discount rate.
The first thing to recognise is that people don’t discount in an exponential fashion, but rather in a nearly hyperbolic fashion. What that means is that they will suffer from time inconsistency in their decision making: they’ll plan to do something but then, when the time comes, they won’t want to do it any more. This situation is particularly important for climate change policy. If we delay now because the sacrifice isn’t worth it then we need to make greater sacrifices in the future. When we get to the future it is unlikely to seem optimal to act straight away. Absent of any precommitment mechanism it is unlikely that any plan to radically decrease emissions starting in the future will be followed. Unfortunately, governments tend not to have access to such commitment mechanisms.
The second issue with discount rates concerns inter generational equity. By saving money now and letting the environment deteriorate we can compensate the future generations in cash for the loss of environmental services. However, by allowing the environment to deteriorate we haven’t only cost them in terms of environmental services: we’ve also restricted their choice set. It isn’t possible for them to buy back the lost biodiversity or destroyed ecosystems. Those changes are essentially irreversible. So, when making the decision to delay, we need to ensure that we compensate future generations for both the degraded environment and the reduced choice set that they are left with.
I’ll finish by noting that I have omitted to mention the vast issue of animal welfare and the costs that we impose upon them by delaying. That is a topic I’ll leave for another time.
Also known as the “first mover disadvantage”.
“Also known as the “first mover disadvantage””
Which occurs when we have strategic complements, which implies that if there is no intervention there will be a sub-optimal value of “environmental restraint” produced 😛
Good post Rauparaha. Is it true that humans definitely use hyperbolic discounting? I haven’t read more than a couple of summary papers on it, so I’m intrigued – do you have some empirical literature on it that I may have a look at?
Look on the bright side:
http://stuff.co.nz/4486484a34.html
Ah, Matt, you know it warms my heart when you say things like that 😉
The seminal psychological work on hyperbolic discounting is George Ainslie’s 1992 book, ‘Picoeconomics’, I believe. Wikipedia’s ‘Hyperbolic discounting’ article has a pretty good bibliography so you might want to check that out, cos I don’t think Picoeconomics is available online.
It became cool to use it in economics after David Laibson’s 1997 paper, ‘Golden Eggs and Hyperbolic Discounting’:
http://harbaugh.uoregon.edu/Readings/Time/Laibson%201997%20QJE,%20Golden%20eggs%20and%20hyperbolic%20discounting.pdf
For applications to procrastination check out Akerlof’s 1991 paper or the more recent work by O’Donoghue and Rabin. There’s also some application to cigarette smoking and addiction as recently written about by Gruber and Koszegi.
Most people these days don’t use hyperbolic discounting cos it’s mathematically more complex. Instead they use quasi-hyperbolic discounting, which is basically exponential discounting with an extra bit of discounting of the (t+1)th period.
1) I’m unhappy with the implication you seem to make with hyperbolic discounting, namely that interventions to prevent costs being delayed are welfare enhancing most of the time.
2) I don’t know if the concept is applicable to government decisions. A lot of policy decisions involve at least a pretense of cost-benefit analysis and pre-defined discount rates.
3) Even without hyperbolic discounting, one can claim that government’s lack a long-term focus because of electoral cycles. But as everyone claims that their policies have long-term benefits, I’m not certain which specific policies this claim supports.
4) Existing climate change policy already has a long lead-in time for meaningful change, with most of the sacrifices likely to occur in the future. Relative to the current policy, delaying the starting period hardly seems like much of a change.
5) Governments have commitment mechanisms – presumably it would be costly to our international reputation to exit Kyoto, and governments also care about their domestic reputation.
6) There is another argument for delay to do with the prospect of diminishing uncertainty.
7) At the end of the day, I have no idea whether delay is optimal, but with some assumptions about rate of technological change versus rate of climate change, it would be.
8) Is it valid to expect future generations to compensate us for the expanded choice set that will come from higher incomes and newer technologies?
hmm, who knew that an “8” and a “)” was a little dude in sunnies?
I think the key implication of hyperbolic discounting isn’t that it isn’t optimal to delay now: it’s that, if it’s optimal to delay now, it is likely to remain optimal to delay at the future point when you plan to act. If governments had plausible precommitment mechanisms then this wouldn’t really be a problem. However, I’m not convinced that rejecting an international treaty that costs a lot of money to implement would really hurt them at the polls.