Historical Social Discount Rates

Last week, with a day to spare, I got my submission into the 2016 Joint Statistical Meetings (JSM). My submission is something of a side project that grew out of my dissertation. When doing benefit-cost analysis, including a look into the future of more than a year, it is typical to discount future benefits and costs. This discount is based on the idea that a dollar in the future is worth less than a dollar now.

This is not an effect of inflation, either. Even with a zero-inflation expectation, future dollars are worth less because a dollar I can spend right now has greater utility. Benefit-cost analysis takes this into account by discounting future cash flows. The only question is how much. And that’s where the fighting begins. Every couple of years there’s a new estimate given in the pages of the Journal of Benefit-Cost Analysis, and everyone has an opinion.

Without going through the entire history of the world, I generally like a simpler approach. I firmly believe the appropriate discount rate is the cost of borrowing to finance the project. Generally, we don’t know that ahead of time. We can guess, especially in the case of Federal borrowing. But more than a few nows out, we have to go with instinct and call it.

This leads to an interesting question in the case of retrospective benefit-cost analysis. In retrospective analysis, if we follow the logic through, we should “reverse discount” long ago expenditures because of the increasing value we have received from them. But selecting a rate is difficult. That’s where my view of using the government borrowing rate becomes an advantage. We know exactly what it is for all time.

That’s what I did for my analysis of the NFIP. When I brought this up with Scott Farrow, my advisor, I don’t even recall him thinking twice about it. But then, two years later, my committee got ahold of it…and they were confused. They let me pass provided I eventually wrote up a note explaining this thing I did with the discounting. So here I am following up on that, finally!

My proposal for JSM is a method to calculate that historical rate. Here’s the abstract:

Benefit-cost analysis assumes that a future dollar is worth something less than a dollar today. The difference between the future value and the current value is called the discount rate. Economists, statisticians, and politicians routinely disagree over how to select the rate. Further, when performing a retrospective benefit-cost analysis, there is no clear method to apply social discounting in reverse. In this presentation, we will first develop the case for discounting retrospectively. Second, we will discuss potential rate selection regimes and develop a method based on historical borrowing. Third, we will give proposed values for the social discount rate for each year from 1961 through 2015. Finally, we will discuss the implications using this method and social discount rate on historical analyses.

See you in Chicago!

Image by Kurt Bauschardt / Flickr…not that kind of discount.