In response to my earlier post on climate change and aggregate demand, Lance Taylor sends along his recent article “Economic Growth, Income Distribution, and Climate Change,” coauthored with Duncan Foley and Armon Rezai.
The article, which was published in Ecological Economics, lays out a structuralist growth model with various additions to represent the effects of climate change and possible responses to it. The bulk of the article works through the formal properties of the model; the last section shows the results of some simulations based on plausible values of the various parmaters. 1 I hadn’t seen the article before, but its conclusions are broadly parallel to my arguments in the previous two posts. It tells a story in which public spending on decarbonization not only avoids the costs and dangers of climate change itself, but leads to higher private output, income and employment – crowding in rather than crowding out.
Before you click through, a warning: There’s a lot of math there. We’ve got a short run where output and investment are determined via demand and distribution, a long run where the the investment rate from the short run dynamics is combined with exogenous population growth and endogenous productivity growth to yield a growth path, and an additional climate sector that interacts with the economic variables in various ways. How much the properties of a model like this change your views about the substantive question of climate change and economic growth, will depend on how you feel about exercises like this in general. How much should the fact that that one can write down a model where climate change mitigation more than pays for itself through higher output, change our beliefs about whether this is really the case?
For some people (like me) the specifics of the model may be less important that the fact that one of the world’s most important heterodox macroeconomists thinks the conclusion is plausible. At the least, we can say that there is a logically coherent story where climate change mitigation does not crowd out other spending, and that this represents an important segment of heterodox economics and not just an idiosyncratic personal view.
If you’re interested, the central conclusions of the calibrated model are shown below. The dotted red line shows the business-as-usual scenario with no public spending on climate change, while the other two lines show scenarios with more or less aggressive public programs to reduce and/or offset carbon emissions.
Here’s the paper’s summary of the outcomes along the business-as-usual trajectory:
Rapid growth generates high net emissions which translate into rising global mean temperature… As climate damages increase, the profit rate falls. Investment levels are insufficient to maintain aggregate demand and unemployment results. After this boom-bust cycle, output is back to its current level after 200 years but … employment relative to population falls from 40% to 15%. … Those lucky enough to find employment are paid almost three times the current wage rate, but the others have to rely on subsistence income or public transfers. Only in the very long run, as labor productivity falls in response to rampant unemployment, can employment levels recover.
In the other scenarios, with a peak of 3-6% of world GDP spent on mitigation, we see continued exponential output growth in line with historical trends. The paper doesn’t make a direct comparison between the mitigation cases and a world where there was no climate change problem to begin with. But the structure of the model at least allows for the possibility that output ends up higher in the former case.
The assumptions behind these results are: that the economy is demand constrained, so that public spending on climate mitigation boosts output and employment in the short run; that investment depends on demand conditions as well as distributional conflict, allowing the short-run dynamics to influence the long-run growth path; that productivity growth is endogenous, rising with output and with employment; and that climate change affects the growth rate and not just the level of output, via lower profits and faster depreciation of existing capital.2
This is all very interesting. But again, we might ask how much we learn from this sort of simulation. Certainly it shouldn’t be taken as a prediction! To me there is one clear lesson at least: A simple cost benefit framework is inadequate for thinking about the economic problem of climate change. Spending on decarbonization is not simply a cost. If we want to think seriously about its economic effects, we have to think about demand, investment, distribution and induced technological change. Whether you find this particular formalization convincing, these are the questions to ask.
- In the jargon, this is a calibrated model. To economists calibration means picking values for model parameters based on what seems reasonable. The alternative to a calibrated model is an estimated one, where the parameters are picked based on statistical tests.
- It would take quite a bit of work to figure out exactly which of these assumptions is driving the results.
Why would labor productivity reverse trend so dramatically? Your footnote indicates you don’t know either, but I think it’s worth asking explicitly given how everything else turns around it.
The causal links are straightforward enough, though a bit complicated. In their model, labor productivity growth is a function of growth in the capital stock and the employment-population ratio. Conceptually, the first of these effects includes both the effect of additional “capital” per worker and a pure return to scale a la Verdoorn’s law, but in the model these are combined in a single parameter, which is set at 0.5 in the calibrated model. (I.e. an additional 1 percentage point growth in the captial stock yields an additional 0.5 point growth in output per worker.) Climate change then has two relevant effects: It reduces the profit rate, discouraging investment; and it leads to faster depreciation of the existing capital stock. Because output is demand-determined and demand is profit-led, lower profits result in lower employment as well.
So we have four channels for climate change to reduce labor productivity: (1) lower investment –> slower growth in K; (2) higher depreciation –> slower growth in K; (3) lower investment –> lower short-run equilibrium employment rate –> less induced productivity growth; (4) slower output growth –> slower employment growth –> declining emplyoment rate –> less induced productivity growth.
That much the article lays out quite clearly. What is not clear is how important each of these channels are in driving the results. For example, one might suppose that faster depreciation would imply higher investment — as coastal structures are lost to rising seas, more new construction is needed, etc. That modification to the model would reduce the effect of the first two channels and weaken or reverse the third channel. But would that be enough to change the qualitative results? That’s what’s hard to tell.
The macroeconomics and modeling in this paper are way over my head, but its BAU picture of economic collapse under global warming doesn’t mesh well with the real-world effects of climate change. It’s unlikely that climate change will have more than a modest impact on production and output, even in worst-case scenarios.
Take agriculture. The IPCC consensus is that warming, up to a worst-case 5 degrees celsius, will have modest effects on yields that can be compensated for by adaptations. Meanwhile, advances in irrigation and desal bid fair to drought-proof agriculture (a process well under way in Israel). The Rezai paper posits a BAU temperature plateau at + 7 degrees after 300 years; that’s not likely to crush food production, given the dramatically advanced technology we will have by then.
Sea-level rise is the other big issue; it will drown coastal areas and displace hundreds of millions of people over several centuries. But that burden of reconstruction and resettlement is small compared to the frenzy of migration and city-building that the world is already undertaking as part of ordinary development. Population growth and demand for higher living standards will have a hugely greater impact on economies and resources than climate change will.
(See more on these points at http://progressandperil.com/2018/02/23/the-conquest-of-climate/ )
Given the modest impacts on output, it makes little sense to worry about the effects of climate change on productivity and employment. They will be dwarfed by the effects of automation and AI in the near term and by demographic shifts in the long term as the world’s population ages and then starts to decline in the mid-22nd century. Climate change will hardly register beside these much larger influences on labor markets.
It’s good to invest in clean energy to mitigate warming and to abate pollution and resource depletion. But the cost-benefit analysis shouldn’t be premised on far-fetched economic disaster scenarios. The world economy is a lot bigger than climate change, and its developing capacities will outrun the effects of warming.
It’s unlikely that climate change will have more than a modest impact on production and output, even in worst-case scenarios.
You may be right. I’m not convinced myself that lost GDP or equivalent is a useful way to think about the costs of climate change. If one buys the arguments in my last post — I know you don’t, but for the sake of argument — it might well be that the path of world GDP was higher as a result of World War II than it would have been otherwise. But this tells us nothing about the desirability of avoiding the war.
Sea-level rise is the other big issue; it will drown coastal areas and displace hundreds of millions of people over several centuries. But that burden of reconstruction and resettlement is small compared to the frenzy of migration and city-building that the world is already undertaking as part of ordinary development.
This I think is a better way to think about it. Even if we are interested in economic costs narrowly defined, it’s better to do the analysis in concrete, physical magnitudeds than monetary aggregates like output or consumption. What fraction of the world’s urban areas or housing will be lost each year, and how does this compare to current rates of new development? How much farmland will be lost, and how does this compare to increases in agricultural productivity? How many people will be displaced, and how does this compare to current and historical migrations? Etc. I haven’t seen anyone make these comparisons — maybe they’re at your link? — but they seem like the right comparisons. I do not think that comparing an aggregate cost measure to some aggregate measure of growth is meaningful.
The world economy is a lot bigger than climate change, and its developing capacities will outrun the effects of warming.
Well, that’s exactly the question. Surely we can agree that it’s possible in principle for the effects of climate change to outrun our capacity to adapt?
Anyway, I’m not endorsing this specific analysis — I hope my caveats made that clear. Just thought it was an interesting addition to the conversation.
OK, I read your piece. Makes a very strong case — will write a followup post encouraging people to read it.
A few quick reactions:
– Talking about the concrete effects of climate change and comparing them to similar historical developments is definitely the right approach, imo. Much better than the abstractions economists work with.
– Even if everything you say is right, I don’t see this as an argument against green politics, public investment in decarbonization, etc. As you acknowledge in the section “conquest of carbon,” this kind of politics helps drive the technological adaptations you’re talking about. I strongly suspect that the Israeli desalination program would be a lot less advanced if people had not been talking about the apocalyptic effects of future water shortages.
– Connected with this, there’s the balance of risks. If you are right, then green politics is still helpful on balance, or at most moderately costly. But if you are wrong and the effects of climate change are much greater, with various positive feedbacks, then decarbonization is absolutely critical. It’s perfectly rational to be concerned about unlikely but extremely bad outcomes.
– “The technology is banal” seems like the key point.
– Agree about nuclear. There is definitely an incoherent position on the left. The greater your estimate of dangers of climate change, the more you should favor nuclear. The fact that the people most concerned about climate change are also most opposed to nuclear power suggests something irrational at work.
– On the other hand, I wish you’d rethink your aversion to solar and wind power. The same kind of banal technologies that you talk about elsewhere can overcome problem of intermittency — more extensive grids, better batteries, pumped storage, etc.
–“But if you are wrong and the effects of climate change are much greater, with various positive feedbacks, then decarbonization is absolutely critical. It’s perfectly rational to be concerned about unlikely but extremely bad outcomes.”
Yes, and if there’s a 1 percent chance Saddam has WMDs, we have to invade Iraq. And if there’s a 1 percent chance a reactor might melt down, we have to ban nuclear power; that’s the heart of the “irrationality” in the left position on nuclear that you mentioned.
There are a lot of ways that a decarbonization panic can go wrong, like biofuels. And most of the 50 percent or 80 percent or 100 percent renewables by 2050 plans that we read about are predicated on drastic curtailments in energy usage. That’s unrealistic—and thank God for that, because on the whole the world needs to use a lot more energy. Constricting energy use to the extent envisioned in mainstream green agendas would itself likely be a developmental catastrophe far worse than climate change.
So down with the precautionary principle. Basing policy on the apocalyptic tails of the risk distribution is bad in just about any context. The tails are always going to be more about speculation and demagoguery than about science, and they are bound to yield irrational overreactions and incoherent policy that trips over itself trying to thread a minefield of competing doomsday scenarios. Better to plan from likeliest outcomes, and as much evidence as we can muster.
Yes, and if there’s a 1 percent chance Saddam has WMDs, we have to invade Iraq.
I thought you might say that. Here’s the thing, tho: When Cheney and co. said that an Iraqi nuke would be a catastrophe for the US, *they were wrong*.
The statement that it is rational to take precautions against unlikely but very bad outcomes doesn’t turn “one percent” into a magic formula. It requires us to make an affirmative judgement about the actual distribution of outcomes. It was evident to may people at the time — including me, and I’m pretty sure you — that the claim of a long tail of negative outcomes from nonintervention was not justified by the available evidence. It was bullshit. Whereas the evidence did support a long tail of bad outcomes from going to war, which, again, was widely recognized at the time.
In the case of climate, anyone who follows the science at all is aware of positive feedbacks — permafrost methane, seabed clathrates — that, one initiated, could raise the path of future temperatures far above current forecasts. And if you read the writings of climate scientists, it’s clear that they pretty much unanimously believe that the long tail is all in the direction of higher costs. It would take some chutzpah to argue that because the people with genuine expertise were almost all right about the Iraq war (it was a bad idea), we should now assume they are all wrong about climate change mitigation (it’s a good idea).
Of course we also have to consider the costs of action. If we were talking about, I don’t know, a proposal to drastically reduce the human population, or some Elon Musk geoengineering scheme, then I would agree that there are fat-tailed risks there too and the precautionary principle doesn’t obviously point one way or the other. But that is not what we are talking about. We are talking about public investment in renewable energy and conservation; and measures to encourage less energy-intensive forms of production and settlement. If we look at the distribution of costs here, there’s no plausible risk of them being far higher than expected. And there is a quite plausible case that the costs will turn out to be lower, or even negative, given both there non-climate-change-related positive externalities and the possibility of pervasive demand constraints (or “savings glut”) in much of the world. If we live in a Keynesian world where any kind of public spending will make us richer, much better that it goes to windmills and retrofits than to pyramids and digging hole sin the ground — or to the military, the actual most likely alternative site of public spending.
“In the case of climate, anyone who follows the science at all is aware of positive feedbacks — permafrost methane, seabed clathrates — that, one initiated, could raise the path of future temperatures far above current forecasts.”
There has been strong pushback against the methane apocalypse scenario. There is no consensus on how likely it is, when it might happen or how bad it might be if it happens. It is very much in the Cheneyite vein of imponderable doomsday scenarios that are outside the scientific consensus.
My piece was premised on the IPCC’s most extreme warming scenario (called RCP 8.5). That’s the current consensus on a worst case.
“It would take some chutzpah to argue that because the people with genuine expertise were almost all right about the Iraq war (it was a bad idea), we should now assume they are all wrong about climate change mitigation (it’s a good idea).”
Mitigation is a good idea, I’m just assessing its urgency based on the expert consensus about worst-case warming and feasible responses to it. I’m not arguing that the climate change experts are wrong, I’m arguing that they are right. You are arguing the Cheneyite line that an outlier methane doomsday scenario should trump the expert consensus.
Climate scientists lately have been tamping down the more extreme alarmism. Their response to the recent New York Magazine climate doomsday piece was very negative.
“If we were talking about, I don’t know, a proposal to drastically reduce the human population, or some Elon Musk geoengineering scheme, then I would agree that there are fat-tailed risks there too and the precautionary principle doesn’t obviously point one way or the other.”
If you really think a methane apocalypse is coming, why would you balk at geoengineering?
This is another case of greens tripping over their own tangled doomsday stories. Naomi Klein et al have been busily ginning up barely articulated doomsday fears about geo-engineering. And the green academic establishment has followed suit by getting governments and the UN to proscribe geoeng trials, ensuring that we make no progress on what might be a crucial tool for averting climate change if their other doomsday scenarios come true. The precautionary principle strikes again.
We should be experimenting with lots of geoengineering schemes, on a massive scale, right now, to see how well they work. I really do want to mitigate climate change as fast and cheap as possible and I don’t want green fear-mongering to slow that down—precisely because I don’t believe in doomsday prophecies.
In the 60s there were plans to build thousands of nuclear reactors. They were stopped by greens harping on fat risk tails; nuclear got banned or priced out of the market with suffocating regulations. Thanks to that exercise of the precautionary principle, we are now fretting over the methane apocalypse.
“We are talking about public investment in renewable energy and conservation; and measures to encourage less energy-intensive forms of production and settlement. If we look at the distribution of costs here, there’s no plausible risk of them being far higher than expected.”
Realistically, “conservation” and “less energy-intensive forms of production and settlement” spell energy austerity. The costs aren’t in the fat tails of alarmist scenarios, but in the persistence of poverty and squalor and death from lack of energy. Fear of the methane apocalypse may head off a new gas plant in India that could run reliable air conditioning and sewage treatment, but many people will die from that.
I know you don’t want that, but why not? We don’t know the tipping point; maybe your methane apocalypse will come at 3 degrees, maybe 1.5 degrees, maybe 1.1 degrees. Maybe we have just a few years to go completely carbon neutral before doomsday dawns. If you take it seriously, the methane apocalypse justifies immediate radical austerity policies that could kill people on a massive scale in order to save people on a massive scale. That’s the logic of doomsday prophecies, and that’s why people invoke them: to sweep aside opposition to extreme policies.
I get that you’re angling for moderate Keynesian stimulus, not harsh austerity. I’m not against that, and that’s what the world will accept. Planned energy austerity is not politically feasible anywhere and just won’t happen no matter the doomsday theorizing. (That’s one reason that fossil fuels will be around for a while). And I kind of agree with the argument that with current low borrowing costs governments can finance a lot more investment in clean energy. Realistically, given politics and competing needs and problems with RE, that program will be moderate: faster decarbonization than now, but not fast enough to make 2 degrees.
That’s still a good program, but it should be argued on the merits, not with methane doomsday scenarios. (And not with the World War II analogy, which doesn’t fit well.) First because the methane scenario is probably wrong. Second because the moderate program you will actually get is inconsistent with a World War III against Doomsday. If you really believed in the methane apocalypse you would go whole hog for austerity, geo-engineering, the works; you won’t do that, so you will be arguing in bad faith. Third because the prospect of apocalypse at any moment can induce fatalism—or denialism, to alleviate anxiety. Making things better is better politics than saving the world.
Finally because the precautionary principle is one of the worst principles ever invented. It’s already done enormous damage to the climate, and it’s visibly slowing decarbonization and other mitigation efforts. It’s seldom used rationally, or even coherently, even by you. The apocalyptic mindset must be challenged vigorously at every turn. If we live in fear that the world will end if we put a foot wrong, we won’t move forward.
“When Cheney and company said that an Iraqi nuke would be a catastrophe for the US, they were wrong. The statement that it is rational to take precautions against unlikely but very bad outcomes doesn’t turn one percent into a magic formula. It requires usto make an affirmative judgment about the actual distribution of outcomes. The cliam of a long tail of negative outcomes from nonintervention was not justified by the available evidence. It was bullshit. Whereas the evidence did support a long tail of bac outcomes from going to war, which, again, was widely recognized at the time.”
But we didn’t know for sure that Cheney was wrong until we invaded. And yes, these things were “widely recognized,” but not universally recognized.
Doomsday scenarios aren’t like actuarial tables where we can make a confident judgment on the distribution of outcomes. They are just guesstimates based on suppositions. It only take one man with dredentials to create a doomsday scenario, and unless it’s ruled out by the laws of physics it goes into the risk tail of expert opinion. If we then treat it respectfully instead of skepticallyit automatically win the probability X consequence calculation: by positing essentiallyinfinite consequences, it trumps any counterargument from improbability or any balancing against finite costs.Doomsday scenarios are both imponderable and unanswerable, which is why demagogues like Cheney and Nader love them.
“I wish you’d rethink your aversion to solar and wind power. The same kind of banal technologies that you talk about elsewhere can overcome problem of intermittency — more extensive grids, better batteries, pumped storage, etc.”
Right, I should check my own tech biases.
Solar and wind are definitely helping. They are not growing fast enough for the 2-degree benchmark but they are appreciably slowing the growth in fossil-fuel usage, and that’s important for knocking the emissions trajectory away from worst-case scenarios. A 3-4 degree warming, say, is less of a problem than 6-7 degrees.
But solar and wind won’t completely get rid of fossil fuels. Grids can’t be extensive enough because common-mode failure in wind and solar output occurs on continental and hemispheric scales. Pumped hydro is geographically limited. Batteries will help with very short-term storage and grid stabilization. But comprehensive decarb would require storage of RE on seasonal time-scales, storing summer solar to use on winter evenings; that’s out of the question with batteries. (See here for an analysis of exorbitant battery storage needs in Germany https://thebreakthrough.org/index.php/issues/renewables/the-grid-will-not-be-disrupted )
A better bet than batteries is carbon capture and storage for the residual fossil fuels required to firm up a mostly RE-powered grid. (CCS hasn’t worked too well yet, but it’s not a dead-end and it would be needed anyway to decarbonize cement and steel production, so it should be funded and developed.) It’s probably feasible to get to an 80 percent RE and 20 percent FF grid eventually. Even without CCS, that would be a cleaner grid that would head off worst-case scenarios and give us more time to adjust and figure things out.
A nuclear grid would be better for a lot of reasons, but building wind and solar, much as I hate them, is still a useful thing to do. They won’t fully decarbonize the grid, but I don’t think we need them to.
Pumped hydro is geographically limited. Batteries will help with very short-term storage and grid stabilization. But comprehensive decarb would require storage of RE on seasonal time-scales, storing summer solar to use on winter evenings; that’s out of the question with batteries.
How about with ammonia?
Ammonia for RE storage:
Maybe, but current P-A-P cycles lose about two thirds of the energy or more. So two thirds or more of the RE we would “store” in ammonia is actually just spilled. Maybe fuel cells can get that down to about one third wastage, but those are still just research programs. Cost estimates I’ve seen suggest ammonia storage would be two to three times as expensive as CCS gas-fired peaker electricity. We should fund all these technologies to see which work best and cheapest. Trump signed an expansion of tax credits for CCS, so it’s not all bad news.
Nukes do not need storage to run the grid, (though they may benefit from short-term battery storage of a few hours to arbitrage fluctuating electricity prices). That’s a huge point in their favor, because storage is still hella expensive and constrained.
There’s a complex of problems here – CO2 being only the most salient. For instance, increases in agricultural productivity owe a great deal to nitrogen fertilisers, insecticides and larger-scale farming, but nitrogen overload is a major problem, and large-scale farming is basically soil mining – you run out of soil if you persist (a good many places have found this out the hard way). Likewise species loss, especially of insects, is likely to have cascade effects at some point (when the insecticide level reaches the point, as it has in parts of the UK, where the load in worms is enough to poison foxes, then you are close to the limit).
Desalination will provide water for people, It won’t provide water for agriculture. It won’t keep forests alive.
This is not to say I have answers – just that this kind of thing typically has quite marked tipping points, past which you fall off the cliff.
“Desalination will provide water for people, It won’t provide water for agriculture.”
The Israelis are using it for irrigation.
There are very few major food-surplus regions in the world (US Midwest, Argentina/Brazil, Russia, Australia). All rely on broad-acre rain-fed agriculture. Desalination for Kansas, pumped over from the East Coast?
Or up from the Gulf. It would cost about $15 billion a year to replace the entire annual draw from the Ogalalla aquifer with desal–plus a distribution infrastructure, but much of that is already in place in the Midwest. We can move water a long way. China is building 700 mile-long canals.
I don’t expect this to happen in Kansas because raw water is still usually a lot cheaper than Israeli desal prices, (though in California during the recent drought auction prices for irrigation water were considerably higher). Regions that have abundant natural water won’t find it profitable to build out desal just for occasional droughts; cheaper agricultural improvements will take precedence.
But in semi-arid regions like the Middle East or Australia, desal-fed irrigation could be deployed on a massive scale if food prices rise a lot, and hugely increase production across wide areas.
To get an idea of the advantages of irrigation:
In Kansas last year about 40 percent of the corn crop was grown on irrigated acreage, and yields were 84 percent higher on irrigated land than on non-irrigated land.
For winter wheat, irrigated yields were 26 percent higher.
So if we face a food crunch in the future, more irrigation in Kansas may be a priority.
OK, you got me going.
In Texas irrigated acreage produced 58 percent of the corn crop last year with a 90 percent higher yield than non-irrigated land; and 26 percent of the winter wheat crop with an 88 percent higher yield than non-irrigated land.
In Nebraska irrigated land produced 65 percent of the corn with a 44 percent higher yield; and 15 percent of the winter wheat crop with a 73 percent higher yield. (see https://www.nass.usda.gov/Quick_Stats/Ag_Overview/stateOverview.php?state=NEBRASKA etc.)
I’ll stop now, but it’s clear that irrigation is hugely important in the US grain belt, and we should be trying to increase it.