Felix Salmon

Greg Mankiw has a

simple equation:

Cap-and-trade = Carbon tax + Corporate welfare

There’s a kernel of truth here, but it ignores two big issues. The first is

that cap-and-trade caps emissions, while a carbon tax doesn’t. In other words,

even if the fiscal sides of the equation balance, the carbon output might not.

The second issue is that cap-and-trade both can and should allocate quotas

not by right but by auction. Most cap-and-trade schemes, indeed, include some

kind of partial auction mechanism. And it’s here where I’m much more optimistic

than Mankiw. He would agree, on the level of his "fundamental theorem",

that a fully auctioned cap-and-trade system has no corporate-welfare component

and is tantamount to a carbon tax. So then the only question is how do we get

there from here.

One option is to try to impose a carbon tax, which is politically very difficult,

and which in any case would not be high enough, certainly at the beginning,

to significantly reduce carbon emissions.

The other option is to build a cap-and-trade system, which would,

by definition, cap carbon emissions. You could then, by accident or by design,

slowly increase the amount of quota which was auctioned each year. (This would

happen almost automatically, in that it’s an obvious place for any government

to increase its revenue without raising taxes.)

You end up with a de facto large carbon tax, but without ever passing

a bill where your opponents can accuse you of raising taxes. Smart, no?

Paul Klemperer has three

unanswered questions on climate change. I can’t answer them fully, of course,

but it might be useful to at least make a first-order approximation, or an attempt

at one.

The first question is how likely the occurrences are against which we should

be paying an insurance premium: what’s the chance that global warming is going

to get disastrously bad?

This is actually two questions. You can insure your house against fire, but

you can’t insure your house against global thermonuclear war. There are two

types of disastrous climate change: the insurable type – which can be

prevented if we "pay the premium" in terms of reducing our carbon

emissions – and the uninsurable type. Some disastrous outcomes, such as

the aforementioned nuclear conflagration, are going to be possible no matter

how much we reduce our carbon emissions.

That said, the "business as usual" forecasts in the IPCC reports

and elsewhere are unremittingly grim. If we don’t reduce our carbon

emissions, then the chances of a disastrous outcome are very high indeed –

somewhere between 50% and 100%. The ice sheets in Antarctica and Greenland will

continue to melt, the atmosphere and the oceans will continue to get warmer,

and sea levels will continue to rise. Eventually – and this is mostly

a question of when, not of whether, if we don’t act now – most of sub-Saharan

Africa and the Indo-Gangetic plain will become uninhabitably hot, while most

people living in low-lying cities will find their homes flooded. Both are indubitable

global disasters.

So the probabilities Klemperer is worried about are very high indeed –

much higher than most insurance companies would ever be comfortable with. The

much more difficult question is the size of the insurable probabilities. Let’s

say we do spend 1% of global GDP reducing our carbon emissions: then

what would the disaster probability become? The difference between that answer

and our first answer is very important, and it’s that number which it’s very

hard indeed to get a good bead on.

Klemperer’s second question deals with other types of catastrophe-as-seen-by-future-generations,

such as the extinction of millions of species, especially in the oceans. This

one’s easier, I think: we know is that after you add them in, the chances of

a global disaster can only go up, and the chances of an insurable disaster can

only go up as well. So if a course of action makes sense to our eyes, it only

makes more sense after answering this question.

The third question is about the moral standing of future generations. Klemperer

concentrates on discount rates here, but the really big effects of Nick

Stern’s calculations come not only from discount rates but also from

the fact that if you push forward 200 years, the sheer number of future humans

so vastly outweighs the number of present humans that even if they’re given

only a fraction of our moral weight each, they still easily outweigh us in aggregate.

The journalistic shorthand, which Klemperer uses, is to talk about "our

great-grandchildren" – which I think provides one interesting hint

as to how we might tweak our approaches here. It’s well known that individuals

care much more, in terms of how much they are prepared to spend, on their own

family as opposed to others’ children; on their own neighborhood; on their own

state; and on their own country. A certain percentage of today’s population

will die childless, and therefore have no great-grandchildren at all; other

families, too, will die out within a generation or two. It is reasonable to

assume that those families might not be prepared to spend quite as much, in

terms of insurance premiums for their great-grandchildren, than those families

which will be vastly larger in 100 years’ time.

It’s also reasonable to assume that most of the population growth over the

next 100 years will come from countries which punch well below their population

weight right now in terms of global GDP. Essentially, Northern Europeans are

paying the insurance premium for Indian families. Which is quite right, in that

it’s the Northern Europeans and the North Americans who caused the problem in

the first place. But even so, the number of families today who would want to

pay the insurance premium might be a little bit lower than assumed in Nick Stern’s

calculations. So maybe the answer here is to keep Stern’s discount rate within

families, but to do the calculations giving each person alive today –

along with all their descendants – an equal weight. That would have the

effect of raising the effective discount rate, but probably not enormously.