Jun 17, 2017

Trash Arguments from Lukewarmer Oren Cass

For those just tuning in, let's have a quick recap. Here's the argument in favor of strong climate policy: Unchecked climate change looks bad, potentially very bad, therefore we should cut the greenhouse gas emissions which cause it.

In his "lukewarmer" manifesto articles, Oren Cass disputes the "potentially very bad" clause of the argument, asserting that there is no scientific consensus behind predictions of extreme devastation from climate change, and that continued economic growth would allow us to buy our way out of any problems we might have.

In my response, I argued this was improperly conflating predictions about possibilities with ones about what is most probable. It is true that the IPCC summary of what is most likely to happen does not generally track with the most alarming predictions (though as Michael Mann — an actual climate scientist, unlike either of us — points out, Cass rather understates the actual level of alarm in the IPCC and also its track record of overly cautious predictions).

But there is generally a scientific consensus that those really terrible predictions are at least possible, even if they aren't most likely (for example, a recent paper predicted a sea level rise of several meters over the next century, along with many other disasters). And if they are possible, then they must be factored into our risk management thinking. Even if the chance of a civilization-threatening outcome is only, say, 1 percent, it's still worth a great cost to avoid that risk. Would you bet your life on a 1 in 100 chance? Or a 1 in 20 chance? Even the IPCC says there is a 17 percent chance sea level rise will exceed 98 centimeters by 2100, on a bad emissions trajectory, perhaps by a lot. There are over 30 million people in Bangladesh alone living below that level.

Cass, by contrast, implicitly treats a lack of consensus about whether worst-case scenarios will happen as a consensus saying they definitely will not happen. Which is dumb.

I further argued that predictions of future economic growth over the next century were a thin reed to hang on, in two ways. First, Cass leans heavily on a simulation by William Nordhaus predicting only minor damage to GDP a century out from climate change. To get this result, Nordhaus simply assumes that growth rates will not be harmed by climate change. A survey of 1,103 experts on the economics of climate, by contrast, found that 78 percent thought it would harm growth rates — making for vastly larger economic damage.

Second, I argued that economic predictions of any kind over a century were at least as uncertain as the most wild-assed climate change predictions. Not only is there no consensus theory explaining why there will be steady and continual increase in productivity and growth over the next century, there is not even a consensus theory about why productivity growth happens in the first place. (People argue, necessarily rather hazily, that it's probably technology and efficiency, or something.)

Economics, being a social science, is simply not as predictively rigorous as physics and chemistry. Cass's credulous approach towards highly uncertain and non-consensus economic predictions illustrated the overall unsoundness of his argument. He is cherry-picking his evidence and using an epistemic double standard.

Now Cass is back with a response, and unsurprisingly he has completely failed to grasp the argument. Here's how he recapitulates the above:
Let’s try to untangle this.  
The flow of the argument is: A: "Climate change will be a catastrophe that kills billions."  
B: "Actually, the costs look manageable given the expected rate of human progress."  
A: "That’s only true if human progress continues; what if it halts? I don’t think it will. But it could. So much uncertainty."
Everything about this is wrong. First, climate change might be a catastrophe that kills billions, and therefore we must treat that danger seriously. Second, the economic point above does not rest on zero growth. In my article, I used a total flatlining of growth as an extreme example, both because economists can't rule it out and because US productivity growth was negative in 2016 (also note the entire Eurozone just had zero GDP growth for eight years), but not the only one.

It does not take a total cessation of growth to blow apart Cass's Panglossian economic future. Because it relies on piling up a huge GDP by steady growth, it would be called into question merely by coming in consistently slower than the Nordhaus prediction (as productivity growth has been extraordinarily slow since 2009). Or, as referenced previously, if climate change harms growth rates instead of levels, the cumulative economic damage will be larger by "many orders of magnitude." So actually, there's a pretty good chance the costs will not be manageable even if growth continues, and something like a carbon tax (which incidentally even Nordhaus supports, unlike Cass), looks a lot more responsible.

So the whole response is epically botched. But later in his article Cass makes a couple more erroneous arguments that are worth addressing. Still obsessed with the zero growth canard, he argues that this would be a catastrophe worse than climate change:
And in that case, climate change would be really bad. Of course, another thing that would be really bad is the halting of human progressWorrying about climate change in a zero-growth world is like worrying about the difficulty of achieving universal health insurance coverage in the midst of a second U.S. civil war. Sure, that would be a problem. Just maybe not the one to focus on.
This does not follow at all. If growth does stop, then that increases the necessity of climate policy, precisely because we won't have the future wealth to buy our way out of trouble, as Cass assumes we will. If we do define human progress as continued economic growth (highly contestable, but never mind), then that's a problem that might be fixed at a later date, as human society existed for thousands of years with zero growth in the past. That means the key task for people stuck in a zero growth rut would be to preserve society in as good a condition as possible, so future generations might figure it out. That means strong climate policy.

Second, Cass makes a serious factual error:
There’s also the awkward detail that forecasts of rising fossil-fuel emissions, and thus rising climate risks, themselves rely on continued growth. Cooper is describing a contorted scenario where the growth exists for purposes of creating the climate change but not for purposes of coping with it.
This is wrong. It's true that the Representative Concentration Pathways (RCPs) used by the IPCC to model the entire world climate system generally predict continued economic growth, but they also contain confounding assumptions about increased efficiency and other factors. All but the worst have emissions peaking sometime this century — the aggressive one actually has negative emissions starting towards the end of this century.

What Cass gets wrong here is that rising climate risks track rising concentrations of greenhouse gases, not increased emissions, because emissions are already high enough to cause high warming, relatively quickly. Suppose carbon dioxide emissions are frozen at their current level (about 40 billion tonnes as of 2015), as part of the zero growth scenario. The folks over at Carbon Brief created this handy graphic, based on the IPCC carbon budget analysis:

A neat thing such a steady-state assumption allows, due to the linearity (hopefully) of climate sensitivity, is that warming will also happen steadily, so we can make a 4 degree estimate. At the 66 percent chance level, each degree takes an additional roughly 1,300 billion tonnes of carbon dioxide, and so at 40 billion tonnes per year, that will be accumulated after about 33 years of emissions.
Therefore, by 2103 — without any increase in emissions at all — the chance of keeping warming below 4 degrees will only be 2 in 3, and fading fast. That would be among the worse RCP trajectories. Here's what the IPCC says about that level of warming:
Global climate change risks are high to very high with global mean temperature increase of 4°C or more above preindustrial levels in all reasons for concern...and include severe and widespread impacts on unique and threatened systems, substantial species extinction, large risks to global and regional food security, and the combination of high temperature and humidity compromising normal human activities, including growing food or working outdoors in some areas for parts of the year (high confidence)...precise levels of climate change sufficient to trigger tipping points (thresholds for abrupt and irreversible change) remain uncertain, but the risk associated with crossing multiple tipping points in the earth system or in interlinked human and natural systems increases with rising temperature (medium confidence).
So, something like a 1 in 3 chance of very serious impacts on all sorts of critical social systems (growing food and being able to go outside are generally considered important activities for people), and an uncertain but nontrivial chance of having hit some tipping point that will cause warming to spiral out of control, with risks rising fast. And after that, every 3-4 decades we'll have racked up another degree of warming.

Let me emphasize again, at no point does any part of my argument actually rely on zeroing out growth, as he bizarrely suggests over and over. I addressed these latter points only because they illustrate Cass's slippery reasoning.

Back in 2014 I described Ross Douthat's thinking about climate change as "vague handwaving that reads very much like he has cherry-picked a bunch of disconnected fluff to justify doing nothing." I can see why he's drawn to this sort of garbage.


  1. Like your comment re 1% probability
    Note that USA response to 9/11 predicated on this --
    In November, 2001, Suskind writes, Vice-President Dick Cheney announced that if there was “a one percent chance” that a threat was real “we have to treat it as a certainty in terms of our response.”

  2. So much of climate change denial is based in either premillennial (God/Jesus will step in at the end of all things to save the righteous) or postmillennial (how dare we interfere with "human progress"?) thinking. I come at this from a Christian amillennial perspective: how we respond to the danger here and now is the test of our righteousness, and we are perfectly capable of creating Hell on Earth if we ignore God's command to be good stewards of the earth.

    Your numeracy is unimpeachable, Ryan, and I enjoyed reading your takedown of Cass, but I fear this type of argument will avail us very little in changing policy.

  3. You just published this article on The Week:

    It's nonsense. As The Week does not allow comments I'll put my thoughts here.

    'we can say with very strong confidence that increased warming will cause more extreme weather overall'
    Actually, you cannot. Some weather events are increasing in frequency or severity, others are declining. Most obviously, though never mentioned by people hyperventilating about extreme weather, the temperature itself is getting slightly less extreme - minima are rising faster than maxima. Even if minima and maxima were rising at the same pace it would be good for humanity, as far more people die from cold than from heat.

    Your argument about the uncertainty of continued economic growth was debunked already by Cass: if the economy does not grow above a certain rate, then *necessarily* emissions will decline. This is simple arithmetic. The historical rate of decarbonization of global GDP is 1-1.5%. If GDP grows less than that, then emissions will decline. So any scenario that assumes continued growth in emissions is in fact assuming GDP growth of at least 1.5% a year, and therefore per capita GDP growth of at least 0.5% a year.

    (As the economy was already decarbonizing 1-1.5% a year before politicians started holding climate conferences, one would expect this rate to at least remain steady, you know, with all the efforts around 'mitigation' of climate change. In fact the decarbonization rate declined in 1997-2016 versus previous decades, but one should be kind to the emission-cutters and assume that's just noise and not the result of their failed policies).

    Either GDP grows below 1.5%, in which case emissions decline but we still get much wealthier over time, or it grows above 1.5%, in which case emissions keep increasing slightly but we become much much wealthier over time. Make up your mind. You cannot have economic stagnation with stagnant emissions - if the economy stagnates then emissions decline. All the historical evidence is that no matter what GDP does, emissions grow 1-1.5% more slowly.

    There may not be a physical basis for such a rule, but it's real. As real as the observation that no matter what the rate of population growth is, food supply increases at least as fast.

    'That brings me back to Bret Stephens. He cites a 2014 chart from Roger Pielke, Jr., showing relatively steady natural disaster damages over time at 0.3 percent of world GDP per year.'
    It actually shows damages *based on insurance claims* declining from 0.25% to about 0.1% in 1980-2016 or a similar final year.

  4. (continued as previous comment exceeded character limit)

    'What Stephens failed to anticipate (as his column was from last week) was that Harvey will cost an estimated $190 billion in damages — or 1 percent of U.S. GDP, from one single storm, with an even more powerful one landing mere days later.'

    A complete non-sequitur. You pick a single country instead of global data, and then botch the comparison by linking to speculation about Harvey's total damages - again, the figure Pielke uses is for insured damage. (Which should be more accurate, but still have an upward bias over time because rich people are more likely to have insurance - how many homes in India are insured? How many homes in China were insured 30 years ago and how many are insured now?).

    When the numbers for 2017 are out, do the comparison again - with global figures, of course. One can always cherry pick a single country, a single class of event that appears to be going up or down. That means nothing. You have to look at all the data.

    'That finally brings me to my second point. So far, the world has experienced one degree Celsius of warming, and America has been suffering an unusual spate of natural disasters. But it has always been obvious that developing, poorer countries will be hit with by far the worst effects of climate change. Indeed, massive, highly unseasonal flooding over the past few weeks in India, Bangladesh, and Nepal has killed something like 1,200 people.'

    So... ?

    By juxtaposing the 1ÂșC increase in temperature with these floods, are you saying these disasters are more likely to have happened because of global warming? Well, hydrologists are not so sure.

    'Stephens' advice to developing countries is basically "become rich." That is happening in India and China (though growth in the latter country has slowed markedly over the last few years). But it is not happening in much of Africa, the Middle East, and central Asia.'

    Please specify which are the countries that make up the 'much of Africa, the Middle East, and Central Asia' which isn't getting richer. Economic growth in virtually all countries in these regions continues, save for war zones like Yemen and Syria. And it continues at rather unprecedented rates (not seen before 1970 or 1980).

    'Climate disasters in such places will not appear very "expensive," because the countries are already so poor and likely to remain that way. But that is where by far the most death and suffering is going to happen. If places like the Democratic Republic of the Congo remain poor in 2100, how likely do we think it is that rich countries will pony up billions or trillions to rescue them from climate disaster — or accept tens of millions of climate refugees?'
    Not sure how likely these things are - or why you'd expect 'climate disaster' in these countries. Even if one believes emissions will bring a given level of warming, which in turn will cause a given level of 'disasterness', then the best option is to have as much economic growth as possible for that level of emissions - in other words, to have a very high CO2 efficiency of GDP, or a very low CO2 intensity of GDP if you want to express is that way.

    'All things considered, it's far better to start ratcheting down emissions now and not put all our chips on imagined future growth magically rescuing everyone.'

    Wrong. Reducing emissions per se cannot be a solution - if it was that simple, you'd conclude that Greece, Syria and Venezuela have got this climate mitigation thing figured out.

    The challenge is reducing emissions per unit of GDP at a faster pace than was already happening, i.e. faster than 1-1.5% a year. The US is doing that as fast as Europe, and was already doing it as fast as Europe back when Bush II was president. So I'm not sure who your advice is for.