He discusses it with Russ Roberts.
I claim we should use an intergenerational discount rate of zero. That is, the distant future we should not discount at all. There’s positive time preference within a life, but over the course of generations no one is sitting around impatiently waiting to be born. And once you adopt that move, the further-out future becomes very important for our deliberations. And then the gains from getting this higher compound rate of economic growth, they really do just overwhelm anything else in the calculation.
It is an argument for thinking about long time horizons when making economic policy. That is easier said than done, of course.
If you pretend that, beyond your own lifetime, the discount rate should be zero, then you should be completely indifferent between actions that benefit humans a 100 years from now and those to be living 1,000 in the future. Despite the fact that, at even very small rates of growth, the latter will be much richer than the former.
Right – isn’t compound economic growth the reason why we should discount the future?
Growth will not stay above an “even very small” constant in perpetuity. Normal, well-adjusted people know this instinctively.
Exponential discounting means that having thoughts about eternity is irrational. If your discount rate is on the order of 1%, this eternity starts pretty soon. After a few thousand years at most.
If I were god I wouldn’t let exponential discounters in heaven. I’d save the spots for those who can appreciate it.
Utility (i.e., happiness) should not be discounted. Perhaps *economic value* should be discounted slightly, on the assumption that it has declining marginal *utility*.
Unfortunately, future people don’t get to vote on current economic policy.
This is extremely dangerous thinking, especially in the hands of people who don’t think economic growth is as important as other priorities.
William Nordhaus’s critique (PDF) of the Stern Review provides a great example:
The fundamental philosophical question is, if one is going to optimize policy in favor of some kind of population metric, how many people should be encompassed by the circle of ones considerations?
If one expands on circle to everyone in the world in the present, then the typical universal utilitarian logic leads to conclusions that are not very good for economic development, and often repugnant and completely infeasible. That’s a dead end.
If one expands the circle even further to include everyone in the future too, then that is much more friendly to free-ish market policies and even subsidies that tend to aceelerate innovation and savings and projects with long time horizens.
Of course it’s possible to take a mixed position, which is to keep ones circle narrower than “everybody in the world” and restricted to some identifiable and perpetually-definable subset, but then to take into account the interets of all the future members of that subset when making ones policy calculations. One can argue that this is what many firms already are supposed to do vis a vis their shareholders, and there is also a obvious paralllel for the governance of nation-states focusing on both the welfare of current citizens as well as that of their posterity.
Tell me if I’m off base here, but I thought that the reason you apply a discount rate to some future cost is that — instead of taking some action to reduce that cost now — you could put the money towards some project that has an actual rate of return equal to x%. In other words, if we’re deciding between lowering the temperature by 0.1 degree in the year 3000 and fixing a bridge, the latter project has some return associated with it, and we have to take that into account when we decide if we’d rather spend the resources on lowering the temperature.
Bottom line: This has nothing to do with whether you cross generations, how many lifetimes you span, etc. There’s a market discount rate and it should guide you in how to use resources.