Slightly less sternly bemused – a more considered reflection on the discounting question
Regular visitors to Uncharted Territory will be aware of my willingness to totally deride the economists’ concept of moral hazard. Well, in case anyone hasn’t guessed from my first, second and third diatribes on the subject so far, I’m starting to feel the same way about the idea of discounting when trying to assess how much to spend on averting global warming. What the two approaches – moral hazard and GW damage discounting – have in common, of course, is that they both represent an attempt to imbue the grey science with some life by taking another discipline – psychology – into account. Look, guys, sure, a bit of Undercover Economist type stuff livens things up a bit. But it’s probably best to resist the temptation to try to add a personal touch to macro-level problems. Just a suggestion.
Now let’s try to keep argument very, very simple.
What we’re trying to do is prevent a catastrophe. Remember this. I’ll say it again. What we’re trying to do is prevent a catastrophe.
Now, in my previous post I referred to a a presentation [pdf] by Mike Hanemann referred to in a Realclimate post which noted that:
“Factors underlying the interest rate
• Pure rate of social time preference (ρ)
• How much richer the future generation will be (g)
• If they are richer, how much that lowers the value of money to them (η)
• Interest rate is: r = ρ + ηg
– Nordhaus & Stern both use same η (=1) and similar g (= 1.3%)
– Nordhaus uses ρ = 2.3-3%; Stern ρ = 0.1%
– Result is that for Stern r = 1.4%, while for Nordhaus r = 3.6 – 4.3%. “
Stern’s 0.1% discount rate
What I want to take issue with first is Stern’s ρ = 0.1%. Stern argues that we should spend less to deal with global warming than the benefits will be worth in the future, because the human race might not survive long enough to reap those benefits. A self-fulfilling assumption if there ever was one, of course, but let’s try to look a little deeper.
I started to wonder how the chance of the human race dying out could be 1 in 1000. Surely, if there were possible extinction scenarios it would be worth doing something about them, thereby reducing the probability below 1 in 1000. Unless, perhaps, we don’t do enough about them, because we’re worried we might not be around to reap the benefits…
It seems to me, though, that even the K-T event – the asteroid that killed the dinosaurs 65 million years ago – would be avoidable or at least survivable. We’re already pretty clever little creatures, we humans. So perhaps the probability of extinction in a given year is less than the probability of another K-T meteorite – say 1 in 100 million…
But what sort of extinction scenario are we powerless to do anything to prevent, avoid or survive? A nearby supernova or gamma-ray burster? Well, it might be worth trying to develop a survival plan even for such eventualities. There are only really the unknown unknowns – and we might want to fund research aimed at identifying these. I’d generously put the probability of out-of-the-blue extinction at 1 in a billion, that’s 0.0000001%, which is quite a bit less that 0.1%, especially when compounded!
But if we spend less on averting (say) an asteroid impact catastrophe than necessary to actually succeed in doing so, because of the possibility we might not be around to enjoy the benefit, then we make the catastrophe more likely. And surely GW itself counts as this sort of catastrophe type problem. Now, I can think of several problems (epidemics, nuclear terrorism etc.) that it’s worth doing something about preventing. What if there are really 100s of them and the overall probability of disaster is actually higher than 1 in 1000? If we assume that we don’t do anything about the other problems, then that suggests that we would do nothing about the problem in question, this case, GW. But the fact of the matter is that we want to minimise all the risks.
I suppose the reason 0.1% has been chosen is to make some kind of sense over long timescales. Let’s take an example. Say we had only one chance to deflect a comet that will otherwise return in 10,000 years and crash into the Earth. We have to spend now to avoid certain disaster in 10000 years. 1.001 raised to the power 10,000 is over 20,000. By Stern’s logic, it would not be worth doing anything even if the cost was 1/20,000th of the damage to be caused by the disaster. Would we take that chance? If it cost $5 billion wouldn’t we deflect the comet, even if the damage it would otherwise cause was going to be less than $100 trillion? (And remember all this is at present-day prices). I think we’d launch the mission, but I concede that it’s debatable.
But what if the discount factor is not 0.1% per annum but 0.5%? Then, after 10,000 years the total discount factor is not 20,000 but has 22 digits. It would hardly be worth getting out of bed now to avoid certain destruction of the human race in 10,000 years. Stern has chosen his 0.1% to be low enough to seem plausible, but in fact it is the principle that is bizarre.
Including Stern’s 0.1% – or any other value – for the possibility of the human race not being around to enjoy the benefits of averting global warming is popular psychology economics gone mad.
Nordhaus’ Pure Rate of Social Time Preference
Words almost fail me. But not quite!
What we’re trying to do, remember is avoid a catastrophe. And it’s a social problem, not an individual one.
If I go to my bank manager and explain that I’m only just managing to pay my mortgage, my credit cards are maxed out, and I won’t be able to start paying him back for a few years, but I’d like some free credit please to buy a sports car, because I have a high pure rate of time preference, then, well, he’s going to ask whether I’m pulling a Comic Relief stunt.
If we’re to avoid catastrophe, we have to be the bank manager, not the consumer. The pure rate of time preference must be zero.
The choice is presented as an “ethical” one. Maybe it is, but it’s a choice between fixing the problem and ignoring it and just letting GW happen. Let’s not let the simple ethical choice be confused by smoke-and -mirror economic formulae. Let’s not confuse the way we expect individuals to behave with the constraints we expect institutions to impose on us.
But we’ll be richer in the future…
This is the aspect I’ve dealt with most in my previous posts and I remain very confused.
I’ve had a good look at Stern and he expresses his conclusions (e.g. in the very first conclusion on p.xv) in terms of GDP now (we spend 1% pa) and GDP in the future (we lose 5% to 20%). He assumes a 1.3% GDP growth rate, but this is internal to his models. Basically it’s almost irrelevant.
So I’m forced to conclude that the original Scientific American article, Hanemann’s PPT and much of the discussion at Realclimate is seriously misleading.
Why does all this matter?
The crucial point is that it might be a good idea to determine the maximum price we can get away with putting on carbon. We are seriously underestimating the price needed to force less use of fossil fuels. A few $10s a tonne will make no difference.
Consider the oil price graph in Stern p.288. There, he shows that an oil price of $70 a barrel represents an increase since 2003, when the price of a barrel of oil was $26.30 on average, of £300 per tonne of carbon (i.e. £300/tC or around $500/tC when Stern was writing). Oil today is trading at around $126 a barrel – or over £600/tC compared to 2003 – and we’re now just starting to reduce consumption, though “global demand” may conceivably increase even at present prices.
So, to prompt a switch away from fossil fuels, we’re going to need to impose a carbon cost of something around the external, or social, cost of carbon (which, incidentally, Stern does not derive, only noting e.g. on p.322 that estimates range from “£0/tC to over £1000/tC”!). Individual consumers of fossil fuels won’t be able to get away with spending $1 now to avoid $5-20 damage in the future, though the cost to society as a whole in terms of overall reduction in wealth may well be only $1.
Anyone who thinks that imposing a cost of $30-100/tC is going to make more than a minimal difference to carbon emissions is sadly mistaken. If we go with tradeable quotas, the carbon price will be bid up to $100s/tC, maybe $1000/tC (well, it won’t, because the politicians will wimp out long before that point). And if we go the carbon tax route, the tax will have to be of a similar amount to have any effect.
As usual, we’re all in denial.
Or rather, we’re confusing the “cost” of reducing fossil fuel use in one isolated case, with the (much higher) cost that needs to be imposed to cause a global reduction in consumption. We’re thwarted, of course, by the Displacement Fallacy.
So, as I’ve been arguing, we’re going to need to determine a carbon price by considering emissions to be a debt owed by individuals, institutions or countries to the global community as a whole.