What economists know about climate change

With a mean value of huh, and a standard error of who knows

Back in 2019, Binyamin Appelbaum wrote a really excellent book called “The Economists' Hour: False Prophets, Free Markets, and the Fracture of Society”. In it, Appelbaum explains how economic ideas came to insinuate themselves into thinking about public policy over the past half century or so, often to quite detrimental effect. Whatever the policy question—whether there ought to be a military draft, for example, or how to regulate corporations—events tended to unfold in a fairly predictable way. Economists applied a mode of analysis to a phenomenon which seemed to cut to the heart of the problem, often because the economists abstracted away from important aspects of that problem. Then, partly because of the balance of political interests, and partly because it’s hard to beat a number with no number, the economic way of looking at things won out over alternate perspectives. And then over time the consequences of abstracting away from important aspects of the problem became clear, but it didn’t matter because economic thinking had established a legitimacy in policy circles that isn’t easily shaken.

Then, because economic thinking has established a legitimacy in policy circles that isn’t easily shaken, Appelbaum’s book was quite well received but did not lead to a sudden rejection in economic thinking about policy problems. Which is a shame, but I nonetheless admired the effort. I particularly appreciated his inclusion, as an epigraph to the book, of this quote from “The Elementary Particles” by Michel Houellebecq:

When modern science appeared, medieval Christianity was a complete, comprehensive system which explained both man and the universe; it was the basis for government, the inspiration for knowledge and art, the arbiter of war as of peace and the power behind the production and distribution of wealth—none of which was sufficient to prevent its downfall.

When economics falls from favor, let no one say that Binya didn’t call it.

Now, an economist might say that the reason Appelbaum’s book didn’t lead to the downfall of economics was because it didn’t provide the modern science to economics’s Christianity. Numbers beat no numbers, and until critics of the profession can out-forecast economists, they’re unlikely to make much headway. (Whether the quality of the numbers is really what matters, as opposed to the role economics may play, in the cheeky wording of Fed economist Jeremy Rudd, providing “an apologetics for a criminally oppressive, unsustainable, and unjust social order”, is best left for another post.)

But that argument absolves economists of the responsibility to improve their own scholarship. The truth is that the real human world is ridiculously complex. Even the most clever of models can only provide so much insight into how people make choices, or why things end up one way rather than another. To be actually useful and robust, an inquiry into most any economic question ought to rely on many sorts of analysis: including the use of models made tractable by simplifying assumptions and econometrics, but also simulation and more qualitative sorts of work—drawing on other social science fields, and even the humanities, as Dierdre McCloskey argues. If economics wildly overrates the utility of quantitative analyses built on absurd oversimplifications, then that’s a problem for all of us (given the influence economists wield) but it’s surely not a problem which non-economists alone ought to have to fix. 

And besides that: sometimes bad numbers are worse than no numbers. I’ll give you an example. Tyler Cowen recently wrote a column explaining that economic thinking has its advantages—and that, in particular, it is useful to think about quantities relative to the size of GDP. This is correct! Is $3.5 trillion a big number or a small one? Well, it’s pretty big relative to a lot of things, but as a share of US annual output it’s only about 16%.

But then Cowen continues:

One recent estimate suggests that climate change is likely to destroy about 10% of global welfare (a GDP effect plus an amenities effect) by the year 2200. To the economist, that is a truly significant quantity of resources. Furthermore, the distribution of those losses may be unfair — and just how unfair is more easily judged if one has a sense of the magnitudes involved.

I appreciate the point that he’s trying to make here, but it strikes me as something of an indictment of economic thinking that he treats the 10% of GDP estimate as being at all meaningful. This is a loss as of the year 2200. And if you go look at the paper he’s citing, you’ll see that they run the analysis through the year 2400. A lot of important things have happened over the past 180 years, and even more things have happened over the past 380. Two centuries from now, it’s quite possible that a substantial share of human output will be produced off the planet and a substantial share of humankind will have genes that have been intentionally modified, perhaps substantially. It’s also possible that we will have nuked each other back to the Stone Age—perhaps as a consequence of conflicts that were caused or exacerbated by climate change.

It is my personal view that a scholar with an appropriate degree of intellectual humility would not attempt to publish an analysis like this, because the uncertainty is simply too enormous and there is a chance that the headline figures will be misused. But if I were going to set those things aside and plow ahead nonetheless, I would definitely want to make some attempt to assess how climate change might affect prevailing institutions, systems of government, norms, beliefs, etc. If that seemed too hard, I might at least present, as part of the lit review, a survey of historical work on how climate variability may have affected society’s critical institutions in the past. At the very least, I would raise a giant red flag throughout the paper acknowledging that the analysis takes as given the uninterrupted continuation of prevailing institutions, values, etc in their current form. Instead, the authors write:

Inevitably, our model abstracts from some important aspects. First, the model we propose does not introduce multiple sectors and the effect of temperature on relative sectoral productivity. [Other researchers] show how this can be done in a related framework. Second, we have abstracted from purposeful innovations in green, fossil, and abatement technologies. In our model these technologies only evolve through spillovers from other innovations. Third, the model we develop gains tractability from assuming an economic structure in which anticipatory effects from future shocks or policy only affect land rents, but do not affect allocations.

Most work estimating the long-run costs of climate change does slightly better than this acknowledging the limitations of the analysis. But not much better, and whatever caveats end up included in the work, the findings which end up being cited to make all sorts of policy arguments are those which reckon that as a consequence of x amount of climate change, output will be y percent smaller than would otherwise be the case. This is the approach that informs sophisticated policy debate about global warming: that what we’re looking at, more or less, is a menu of options, and we can decide how much to do about the problem depending on how much of a GDP loss we’re prepared to accept.

And when we do that we are absolutely, 100% kidding ourselves. Reflect for a moment. We are pushing the atmosphere and the global climate well outside the historical range in which complex civilization developed. Though we cannot know the extent to which different causal factors were responsible for this or that major historical event, there is at least some evidence suggesting that changes in climate may have contributed to past episodes of civilizational collapse: of late Bronze Age civilizations, or indeed the Western Roman Empire. We are facing a future in which, across densely populated parts of the tropics, extreme heat waves may mean that it is unsafe to leave climate-controlled environments for days on end; in which extreme weather events occasionally deal massive shocks to agricultural output; in which migrations unfold which dwarf the flows of unfortunates which have roiled rich-world politics in recent years.

It is quite possible that those of us living now, even the younger ones, will see global temperatures rise, higher and higher, for the rest of our lives. Even after we stabilise the level of carbon in the atmosphere it will get hotter; even after we stabilise global temperature the seas will keep rising. For decades after the world achieves net zero, the effects of climate change will continue to mount. Who knows how this will affect people? Who knows what values or belief systems will emerge in response to these trials? Who knows to what desperate measures governments might turn when their citizens demand a peace and stability they are fundamentally unable to provide?

None of us is who. But if we are going to try to develop some understanding of what we might be facing in the years ahead, we are going to need to deepen our understanding of what allows societies to function, of what keeps some stable and causes others to collapse. We are going to need so much more than what economists are prepared to offer.

I suppose that an economist might respond by pointing to one of those charts which shows real output per person in America marching steadily upward (give or take a Depression) from the early 19th century to the early 21st. But this is what’s so boggling! We ought to be absolutely captivated by those charts, and absolutely obsessed (as some economic historians are, bless them) with discovering: what extraordinary forces kept growth on that path, where they came from, why they aren’t present in other countries, and under what circumstances—or climates—they might erode. That kind of trend line doesn’t just happen. Why it looks like it does is among the most fascinating questions in all of science, nevermind economics. And yet economists are, to a remarkable degree, uninterested. Many of them are prepared to extrapolate that line, with scarcely a detectable shift in slope, right through what is likely to be the greatest challenge complex human society has ever had to confront. 

I get it. It is a comfort to be able to point to a quantitative analysis and cite a number. As someone working in economics, it is empowering; as a human, it is reassuring. It is something we can grasp, and write into our policy briefs, and deploy in our online arguments. But we haven’t earned it. We haven’t done the work to be able to say, with any confidence, that a given amount of warming above pre-industrial levels will reduce output in 2011 by a particular percentage.

But we could do better. We could, if we put the work in, say more about the nature of the challenges we are likely to face, and perhaps also something about how human societies tend to respond when placed under (not exactly but at least somewhat) comparable strains. Economists could certainly do better, and it is important to note that some of them are trying.

But I feel sure that those who are trying would understand if the rest of us resolve to be sceptical of meaningless numbers. Not everything that looks like knowledge is knowledge, and laypeople shouldn’t allow themselves to be intimidated out of expressing sensibly sceptical thoughts by fancy looking equations. The stakes, where climate change is concerned, are far too high for matters to be left to the economists.