Social discount rates?

July 31, 2012 9:22 PM

Help me understand the argument that the social discount rate sets the amount we should invest to mitigate climate change.

After listening to this Planet Money podcast and reading this article on the Economist (amongst others), I am confused.

I think I understand what the social discount rate means, but I do not understand how it sets the current cost of investment to mitigate the effects of climate change in the future (as per the Stern report). Instead, it seems like it indicates how likely we are to make this investment. Surely the investment required will depend on the scope of future climate change, and not our likelihood to invest.

Can someone make this clearer to me please!

posted by piyushnz to work & money (8 answers total)

A dollar today is worth more than a dollar tomorrow. The future value of that dollar is discounted at some rate to account for the time value of money.
posted by dfriedman at 9:49 PM on July 31, 2012

"Discount rate" is used in several different fields to refer to how much more you value a dollar now, than later. If you're paying $100 now, in order to get $100 in a year, that's not a good deal because you're having to go without the $100 for a year and not getting anything back. How about if it's $101 in a year? Etc. However much money it needs to be, that determines your discount rate.

This often comes up in psychology, as a sort of "what is wrong with you that you want a lollipop now rather than two in fifteen minutes?". But it's also true from an economic perspective that, having money now, you can always hold it until later; but you can ALSO do a lot of other things with it, and that increases the value of today-money over future-money.
posted by Lady Li at 10:33 PM on July 31, 2012

The social discount rate determines investments we make today in order to mitigate damage in the future. The scope of future climate change is unknown, but there is some best educated guess as to the damage it will cause, and there is some valuation scheme for how this damage can be priced. People's estimates of these will differ but let's assume a world panel of experts agrees that damage will be $1 trillion in 100 years. Let's assume we could totally prevent these damages by spending $10 billion now. Should we do it or not?

There are discount rates that make sense when considering the time value of money -- financial people have developed these through experience about the way people weigh risk. A common one is 6%. If we apply that here, then the present value of the damages is about $3 billion. Therefore, we would not make the investment. This is generally what's happening with climate change.

However, the 'social discount rate' argument says that something critical happened when we converted the damage into a monetary figure, $1 trillion, using a valuation scheme. That conversion changed environmental damage into money, and there is reason to believe that people care about the environment differently than they do about money when it comes to long term risks. If we assume a 6% discount rate, then that means the severity of an environmental damage decreases by 6% for each year further into the future that it occurs, i.e. is cut in half every twelve years. This makes some sense for short time frames; if given a choice between destroying the planet now or in 12 years, we would want the 12years, and hope we can fix the problem in time. However, if given a choice between destroying the planet in 100 or in 112 years, we'd pick 112, but we probably wouldn't think one option was twice as good as the other. They're both almost equally terrible from this vantage point. If this corresponds with how you feel, then a 6% discount rate does not accurately describe the way you feel about the risk of environmental destruction. Some have argued that the social cost is different depending on how far into the future you look, a quality not captured by traditional financial discount rates.

Stern says you can model this effect with a lower discount rate, namely 1.4%. Now the $1 trillion in 100 years is worth $250 billion today. Paying the $10 billion to eliminate this risk will save us $240 billion. It's a slam dunk, no-brainer investment. Whereas with 6%, it was a slam dunk in the other direction.

The crux of the matter is that the discount rate is mathematically encapsulating how society values environmental destruction. Sticking dollar signs on environmental damage is methodologically hairy, but it can be done, and the pretext under which these investment decisions are made (or more likely, not made) is that society, the taxpaying population, will be willing to direct financial resources to prevent this damage from occurring (at the time it occurs), and that the amount they will be willing to direct can be thought of as the cost of this damage, and that once expressed as a cost we can infer how society should prioritize the use of funds by comparing against other costs. Which is valid. Except that if the discount rate used does not match with reality, and in fact far underestimates the seriousness with which the public views environmental damage that occurs in the distant future (and by extension, far underestimates the amount they would be willing to pay to eliminate this damage), then the analysis is totally bogus. So, the battle over how we should handle climate change hinges on this one little number, which carries the world within it.
posted by PercussivePaul at 11:35 PM on July 31, 2012 [4 favorites]

This is a really hard, really important question, and I hesitate to wade into it because I will not do it close to justice, both because I don't know the subject nearly well enough and because it would take much time and effor than I am willing to give.

However, I have some nits to pick with the discussion above, so I'll start there and try to then give some pointers to additional reading.

"There are discount rates that make sense when considering the time value of money -- financial people have developed these through experience about the way people weigh risk." Although risk is a crucial issue both in finance and in social planning, such as thinking about global warming, discount rates are really about time preference, not risk. I can give you one marshmallow today or two marshmallows tomorrow. Which do you want? The answer depends on both how much you hate waiting and how much you trust me to actually give the two marshmellows tomorrow, but those are to a great extent separable issues.

"...there is reason to believe that people care about the environment differently than they do about money when it comes to long term risks." I'm not sure what that means, as environmental damage is a long-term risk, perhaps the earth's major one. (There is a part of environmental economics that measures people's appreciation for simply the existence of the enviroment - e.g., how much would you pay to keep polar bears alive, even if you will never even see a polar bear in your life - but that's not really an issue in global warming.)

"The crux of the matter is that the discount rate is mathematically encapsulating how society values environmental destruction." I'd say that the crux is how we weigh today's costs against the benefits of giving future generations a better place to live. We (or at least I) generally don't care about the "environment," we care about people, just as I don't care about my house, I care about having a nice place to live.

There are two other major related issues that haven't been mentioned:

This is a vast oversimplification, but assume for the purpose of exposition that it would cost $100 billion now to prevent global warming, and that the present value of the benefits would be $1 trillion. Sounds like a no-brainer, right? But what if we could invest $10 billion in some technology that would allow future generations to deal with global warming? In effect, the return on that investment to deal with global warming would be higher than the return on stopping global warming. In that case, the social discount rate should equal the return on that investment.

Future generations will likely benefit from improved technology and therefore be richer than current generations - at least in the developed world. Just as there's usually no justification for a poor man to give money to a rich man, one can ask why we should give up some of our standard of living to help future generations. (The answer is that the benefits to them will be much greater than the cost to us. A poor man should rip up his shirt to make a tourniquet to a sick rich man, even if he will gain no direct benefit.)

OK, that wasn't very good. These should be better:

A Battle Over the Costs of Global Warming
Recalculating the Costs of Global Climate Change
If you are feeling a bit ambitious, Discounting Inside the Washington D.C.
Beltway
posted by Mr.Know-it-some at 7:09 AM on August 1, 2012

Perhaps this is a sidenote, but even in purely commercial decisions, I believe that the apparent mathematical and logical elegance of discounting leads to poor decisions. I am actually involved with decisions that are much more natural and obvious applications of discounting than climate policy. I'm talking straight-forward decisions like should we spend X millions of dollars in order to gain cash flows of Y millions of dollars over Z years, given a discount rate R. Discounting is such an integral and unquestioned fundamental axiom that we even compare the hypothetical discount rate for which a given investment will break even (the so-called DCFR, discounted cash flow rate). In a purely commercial setting, the presence of competing investment alternatives and their DCFRs as well as the test of 'is this better than just taking the X millions of dollars and buying 'risk free' government bonds yielding T, makes discounting seem a very logical and compelling approach.

However, there are problems, in my mind, that arise when looking somewhat far into the future, or considering more 'strategic' investments, or assessing something like climate change. And these problems don't just come from the inherent uncertainty of choosing an appropriate discount rate or from the fact that your ability to accurately forecast costs and revenues tends to sharply diminish the farther these are into the future. Most fundamentally, adopting a 'discounting' approach to decision making inherently and by definition makes you value the future less than the present. My thinking is that if you are in the business or role of trying to make long-range, strategic decisions, the last thing you should adopt is a methodology that by definition lowers the value of future costs and benefits. Even with very low discount rates, the present value of something that you are planning for even as little as 50 years in the future (that's what? One or two generations?) becomes almost negligible.

A lot of the back and forth on discussions of climate change policy has hinged on trying to set a discount rate that gets you the answer that you are advocating for. Apparently, Stern adopted something like 1.5%, which is extremely low, compared to historic rates of return or inflation or comparable guides. But this makes him vulnerable to strong argument and disagreement, not least from economist more used to typical discount rates used, say, in the corporate world. Someone used to oil company valuations, for instance, would find discount rates less than 5% absurdly low (few large energy companies are willing to commit large investments that don't break even at discount rates lower than 10%).

My own opinion, and its an early and still forming one, is that this is just the wrong kind of game to play for long range planning. Again, even at 1.5%, the present value of something 100 years out is basically zero. But I'm sure you can think of lots of things you benefit from today that are the result of good or fortuitous decisions made 100 years ago. The 100 year old house I live in is certainly worth quite a bit today (though its true that it has yielded value over much of the time period). I think of the 250 year old oak tree shading my parents house today, and I don't think the day my brother and I planted 500 oak seedlings on his land in Quebec was pointlessly spent.
posted by bumpkin at 4:11 PM on August 1, 2012 [1 favorite]

Thanks for your thoughtful answers so far.

So, if I have understood correctly thus far, I can think of at least one example of discounting from my own dealings: when my car needs something repaired, I'll take it to the mechanic, who in addition to fixing the original fault will point out a number of other things that need maintenance or repair (we have an old car). Usually I will only get what is urgently necessary done immediately with the idea of leaving the rest till later. However, I know that those repairs I have left for later may end up being bigger bills in the future. I suppose one could determine a number for this. (if the mechanic told me fixing something today would cost $100 now, but $120 a year from now I would probably keep the money).

So, if I am correct about this being an example of discounting, the next bit is the really the root of my confusion regarding the notion of using discounting to set my current investment level in my auto repairs.

It is this: I know I'm being irrational. For example, if I then went to a financial advisor and said, look I have all these repairs and they will cost x dollars now and y dollars a year from now (y > x), my financial advisor might say "pay x now" if the difference was sufficiently large, irrespective of my discount rate. In fact I assume they would base this decision on the fact that I could get a return of z dollars on an investment of x now, but that z is less than y-x. Or they might suggest the reverse (invest the money) if z is greater than y-x.

So, I believe a rational decision on whether to invest now or later should be based on what the real value of that money is a year from now, and not some other number (the discount rate) which is my perceived value. Now, I suspect what everyone is trying to tell me is that z is actually the discount rate, but then why call it that instead of the rate of return on an investment. I know I am oversimplifying things here. Maybe more generally we could call z an opportunity cost which is some sort of average over all the other things I could do with x dollars instead of spending them on my car today… the problem for me is that the discount rate is usually introduced as some behavioural artefact and not as an unbiased estimate.

I'll stop now, since it's late and I've probably written a bunch of gibberish...
posted by piyushnz at 9:25 PM on August 1, 2012

I know I'm being irrational.

That's just it--a lot of financial and economic theory uses as one of its assumptions the notion that man is rational. Yet, as a cursory understanding of people's nature shows, people are not rational.
posted by dfriedman at 2:26 PM on August 2, 2012

I think the more precise way to say it is that you are likely being inconsistent. Let's say you could either pay $100 now for car repairs or $200 in a year. It's not irrational to say that you'd prefer to pay $200 later; it's just a preference. But if, at the same time, you have $100 in the bank that you plan to leave there for a year to earn $2 in interest, you are essentially saying that $102 next year is preferably to $100 now (in the bank account), but that $100 now is preferable to $200 next year (for the car repairs). That is clearly inconsistent. (Again, this is an oversimplification because there's much more uncertainty about the car repairs, but you clearly get the point.)
posted by Mr.Know-it-some at 4:00 PM on August 2, 2012