Pindyck on Two Kinds of Discount Rates

I recently finished Tyler Cowen's new book Stubborn Attachments. I enjoyed it and I appreciated the message. The book is mostly about how important economic growth is. As societies get wealthier, other problems tend to solve themselves. Richer populations tend to demand cleaner environments, lower rates of crime, and solutions to other social problems. Expanded wealth gives these populations the means to tackle these problems. So, the thinking goes, if we can keep our eye on the growth ball, other problems simply dissolve. Given the logic of compounding growth, it becomes extremely important over long timelines to maximize the rate of growth. Adding a mere 1% to the rate of economic growth means the economy will be 2.5x larger in 100 years. If we care about future generations, it seems like this should be a serious consideration.

It reminded me of a topic that was discussed in this episode of Econtalk with Robert Pindyck. About 10 minutes in, there is a discussion of what interest rate should be used to discount the future. A thousand dollars today is worth more than a thousand dollars a year from now, or ten years from now, or 100 years from now. So if we're talking about solving problems today versus passing those problems on to our children or grandchildren, we need to think about how we compare costs today versus costs in the future.

In a distinction I have never heard made anywhere else (and I've read quite a lot on economics), Pindyck makes a distinction between two kinds of interest rates: a utility discount and a financial discount. The "utility" discount is the "I'd rather have it now" effect. Given pleasure today or pleasure 100 years from now, I'd rather have it now. If you want someone to forego pleasure they could have today and instead take it decades from now, you'll have to compensate them (even assuming it's the same quantity of "pleasure", however you'd measure that, and even assuming there is a 100% guarantee they'll get it later). Some people argue, I think reasonably enough, that this discount rate should be zero. There's no moral reason that we should prefer ourselves to other people, whether they live in another country or another century. As Russ Roberts puts it in the podcast, their blood is just as red as mine.

The other component of the interest rate is the financial discount rate. This rate should not be set to zero, even if we care for future generations just as much as we care for ourselves. The reason isn't too hard to understand. A future generation values $1,000 in cash the same as $1,000 worth of climate abatement (by definition). We should pass to them whichever option is cheaper. Resources we spend today on climate abatement (or environmental cleanup or violence reduction) are resources we can't pass on to our grandchildren. A meager-sounding 3% interest rate causes savings to grow by a factor of 19 in 100 years. At 4% it's a factor of 50. At 5% it's a factor of 130. This is a big deal. Trying to solve a future generation's problem incurs a serious opportunity cost.

(I'm not sure whether to use the market rate of return here or the rate of economic growth. I can stow away money today in financial assets today and it might grow at a stellar 7%, which means growing by a factor of 870 in 100 years. But maybe social investments are different from financial investments. The social rate of return might be smaller than the private rate of return. Foregone consumption today may only compound at 2% or 3%, something more like the rate of economic growth. These are very different propositions: "Society as a whole foregoes consumption on $1,000 worth of economic resources so that future generations can use those resources." versus "I personally set aside $1,000 in an investment account to make my great-great-grandchildren filthy rich." Society's investment in the future probably grows more slowly than my personal investment. But if future generations have better technology and better institutions, the resources we bequeath to them will be used more efficiently. So the foregone $1,000 of consumption today compounds at something like the economic growth rate.)

Pindyck claims that when economists say that "the rate of discounting should be zero" they are only talking about the utility discount rate, not the financial discount rate. I think he may be giving a little too much credit to his fellow policy analysts, some of whom aren't quite so nuanced, some of whom fail to communicate this point clearly. But I take his point.

I kept expecting Stubborn Attachments to make this point explicitly. The book spends a lot of time discussing the "utility discount" (though not using Pindyck's terminology). It basically lands on: we shouldn't discount the future at all. We should treat the interests of future generations as equivalent to the interests of people living today. Maybe I missed it, or maybe it's so implicit to the books message that it didn't need to be said, but I had this expectation that Cowen would spell out the two kinds of discount rates and explain why only one is relevant for calculating costs or benefits to future generations. This comes out clearly in the Pindyck and Roberts discussion. They say that if we can grow the economy in a way such that future generations enjoy the fruits of that growth, those future generations would like us to do that. Anything else we'd like to give them has to be traded off against the economic growth we could otherwise bequeath to them.

There's another reason not to use a discount rate of zero, even if we care dearly about future generations. Existential risk. To take an extreme example, if there were a 1% chance each year of a civilization-ending asteroid strike or nuclear war or super-volcano, then it would make very little sense to care much about climate-related harms that take 100 years to manifest. Giving next century's generation, say, $1 trillion worth of climate change abatement isn't very useful if they only have a 37% chance of enjoying it (0.99 raised to the power 100). Even supposing that climate change itself is an existential risk, it may need to be traded off against other existential risks. Maybe shutting down fossil fuels slows the economic rate of growth such that future societies aren't prosperous enough to build an asteroid defense system or space travel. Maybe the wrong emissions reduction policy costs us a century of economic growth, which is the margin between being able to solve some other existential risk and not being able to do so. Like any other set of costs, existential risks need to be traded off against each other. There's no simple answer here, and there's everything at stake. It's a hard problem. But I think Cowen's "maximize growth" principle in Stubborn Attachments is a good place to start.

______________________________________________________

Listen to the Econtalk episode, all the way to the end. It's kind of funny. Russ Roberts is something of a luke-warmer, and Pindyck is the "warmist" in this discussion (that's not to say an alarmist). Roberts brings up the worst-case-scenario for global warming, and Pindyck has to cool him back down. Roberts discusses the possibility of civilization ending and human extinction. Pindyck responds with:

I think you are going a little overboard. There's the worst case scenario; and there's the worst worst-case scenario. And I think you're on the worst, worst, worst case scenario. I don't think that--when we talked about a 5% chance or even a 1% chance, I don't think it's that bad. I think we're looking at something catastrophic, but not that terrible. It would be something that would impose pretty big costs on the world economy, maybe reduce the effective capital stock by 10% or even 20%. That's pretty bad.

A 10-20% reduction in the capital stock would be pretty terrible. It's a bland way of stating things; maybe "reduction in the capital stock" is a euphemism for building falling down or random cities falling into the ocean. But it makes me think some of the worst scare-stories over global warming are oversold.