Climate Change: Extinction or Adaptation?

On October 31st 2018, over one thousand protesters blocked the roads surrounding the UK's Parliament, in the first public action by the newly formed Extinction Rebellion. Its key demand was that the UK should reduce its net CO2 emissions to zero by 2025, a target which—if replicated across the globe—could limit Global Warming to no more than 1.5°C above pre-industrial levels. One year later, Extinction Rebellion has groups in almost every country on the planet, and its message that "we are facing an unprecedented global emergency" is everywhere.

But is this message correct? At much the same time as Extinction Rebellion proclaimed a global emergency, the academic discipline of economics awarded its highest prize to William Nordhaus "for integrating climate change into long-run macroeconomic analysis". In his Nobel lecture (Nordhaus 2018), Nordhaus described a 4°C increase over pre-industrial levels by the year 2140 as "optimal", in that it minimized the sum of the damages from Global Warming and the costs of abatement policies needed to contain it to 4°C (see Figure 2).

Figure 1: Slide 6 in Nordhaus's 2018 Nobel Prize Lecture, annotated

Nordhaus argued that, in the absence of policies to mitigate Global Warming, the temperature would rise to more than 6 degrees above pre-industrial levels over the next 150 years, and that this would impose costs on the global economy of roughly US$23 trillion in today's dollars. He estimated that abatement policies, which would cost roughly $3 trillion, could restrict Global Warming to a peak of 4 degrees Celsius over pre-industrial levels by 2140. This would reduce the economic damages from Global Warming to roughly $16 trillion, yielding a combined cost from both climate change and abatement policies of $19 trillion.

Any more abatement than this, he claimed, would not be worth it: the combined costs of climate change and abatement would exceed $19 trillion. For example, Nordhaus asserted that the policy Extinction Rebellion recommends, of restricting Global Warming to 1.5 degrees (even if this were achieved over the next century, rather than by roughly 2050, as would occur with Extinction Rebellion's 2025 target for ceasing net CO2 emissions) would limit the damage from Global Warming to about US$3 trillion, but would impose abatement costs of more than $50 trillion (see Figure 2).

Figure 2: Slide 7 in Nordhaus's 2018 Nobel Prize Lecture

That much abatement, Nordhaus concluded, simply wasn't worth the cost. We would be better off to only slightly limit the increase in global temperature, and instead learn "to love the altered landscape of the warmer world" (Nordhaus 2007, p. 693).

Nordhaus's conclusion was in line with that of the IPCC, the "Intergovernmental Panel on Climate Change", the United Nations body coordinating the global response to climate change. Its latest full report, published in 2014 (Field, Barros et al. 2014), stated in its Chapter 10 on "Key Economic Sectors and Services" that "For most economic sectors, the impact of climate change will be small":

For most economic sectors, the impact of climate change will be small relative to the impacts of other drivers (medium evidence, high agreement). Changes in population, age, income, technology, relative prices, lifestyle, regulation, governance, and many other aspects of socioeconomic development will have an impact on the supply and demand of economic goods and services that is large relative to the impact of climate change. (Arent, Tol et al. 2014, p. 662)
 

The Report noted that "Global economic impacts from climate change are difficult to estimate", and that many attempts to do so "do not account for catastrophic changes, tipping points, and many other factors". Nonetheless, it concluded that "With these recognized limitations, the incomplete estimates of global annual economic losses for additional temperature increases of ~2°C are between 0.2 and 2.0% of income" (Arent, Tol et al. 2014, p. 663).

Figure 3 shows the data points on which this conclusion was based, with the vast majority of them predicting a reduction in GDP of less than 3% for temperature increases of up to 3°C over pre-industrial levels. Notably, two of these data points predicted an increase in GDP caused by Global Warming—an almost 3% increase in GDP for 1°C warming, and a 0.1% increase for 3°C.

Figure 3: IPCC WGIIAR5 Figure 10-1 | Estimates of the total impact of climate change plotted against the assumed climate change (proxied by the increase in the global mean surface air temperature); studies published since IPCC AR4 are highlighted as diamonds; see Table SM10-1.

Also, as the recently deceased climate change economist Martin Weitzman emphasized in 2011, these are not estimates of enormous damages in some far future time, which have then been reduced to relatively small levels today by a high discount rate. They are instead estimates of how much lower GDP would be at some future date when global temperatures are several degrees higher than today, relative to a world in which Global Warming did not occur at all.

Weitzman illustrated how trivial these estimates of damage from Global Warming are using the example of a 10°C increase in temperature, which is well off the chart shown in Figure 3, but can be calculated using Nordhaus's DICE model circa 2011, which was fitted to the data in an earlier version of this table (see Figure 7). DICE predicted that this a 10°C increase in global temperature would reduce GDP by 19%, relative to what GDP would have been in the complete absence of Global Warming.

Figure 4: Weitzman's Table 3, which uses the "Damage Function" from Nordhaus's DICE model from 2011

Weitzman noted that even this relatively large fall in GDP translates into a trivial fall in the expected rate of economic growth, from a hypothetical 2% per year to 1.9% per year:

If the annual growth rate is, say, 2 percent and the time of impact is, say, two centuries from now, then the welfare difference between no temperature change in two hundred years and a temperature change of 10°C in two hundred years … is equivalent to a reduction in projected annual growth rates from 2 percent to 1.9 percent. (Weitzman 2011, pp. 280-81)
 

This is little more than rounding error in estimating the rate of economic growth: a fall in the rate of economic growth of 0.1% per year is something no-one would march in the streets about, let alone the several million who took part in the "Climate Strike" of September 20-27 2019. So are the kids like Greta Thunberg deluded?

The self-proclaimed "sceptical environmentalist" Bjorn Lomborg certainly thinks so, as he has made abundantly clear on Twitter (see Figure 5).

Figure 5: A Lomborg tweet, citing the 2014 IPCC report to support the argument that climate change is a trivial issue

While not denying that Global Warming is happening, and that it is caused by humanity's CO2 emissions, Lomborg argues that it is a far less important issue than, for example, directly reducing poverty by increasing the number of coal-fired power stations in India. He frequently cites IPCC reports to support his position. Commenting specifically on the Climate Strike movement initiated by Greta Thunberg, Lomborg stated that:

Yes, global warming is real and human-caused, but her vision of climate change as the end of the world is unsupported. The UN's Intergovernmental Panel on Climate Change estimates that by the 2070s, the total effects of climate change, including on ecosystems, will be equivalent to a reduction in average income of 0.2 to 2 per cent. By then, each person on the planet will be 300 to 500 percent richer. (Lomborg 2019)
 

So are the kids all wrong? To state my biases clearly before I answer this question, I initially believed that the kids were correct, and the economists were wrong. However, I expected that establishing this would be an arduous task: I would need to explain why the modelling framework they used was inappropriate for the topic of climate change, why this framework ignored the out-of-equilibrium dynamics that apply while the climate is changing, and many other technically challenging topics.

I assumed that economists had taken estimates of the impact of Global Warming developed by physical scientists working in climate change, and trivialized these results by putting them into equilibrium-based models (when climate change is anything but an equilibrium process) and by discounting these future damages so that catastrophe in, say, 100 years, was reduced to next to nothing in "Net Present Value" terms today.

Economists have done both these things, but not to estimates of the impact of climate change devised by physical scientists. They have instead made up their own estimates of the impact of climate change, using assumptions and methods that are transparently absurd. It's not the kids who are deluded, but the economists.

For Whom the Bell Trolls

You might expect, as I did before reading this literature, that the studies summarised in Figure 3 collated the work of physical scientists—meteorologists, chemists, engineers, geologists, physicists, agronomists, physicians and the like—whose training would equip them to make estimates of the physical consequences of such temperature changes on the industries whose outputs constitute GDP, and on the humans who will have to live in the altered climate. The role of economists in such exercises would be to translate these physical and physiological consequences into estimates of their impact on GDP expressed, in monetary terms.

Sadly, that is not the case: none of the studies summarized in Figure 3 were done by physical scientists. Instead, they are all estimates made by economists, using 4 different techniques, which they characterized as "Statistical", "Enumeration", "Expert elicitation" and "CGE" ("Computable General Equilibrium"): see Figure 6.

Figure 6 also lists the studies and their primary authors. The first thing that should be obvious is the paucity of both researchers and studies: there are just 15 researchers (counting all co-authors) and 18 studies, yielding 19 data points for 8 different increases in global temperature—all but two of which are below 3°C (recall that Nordhaus's optimum temperature rise was 4°C). This is both a trivial number of data points for such an important subject, and a ludicrously small number from which to derive functions that purport to predict the level of economic damage for any given level of temperature increase—and yet that is why these economists generated these data points in the first place.

Figure 6: IPCC WGIIAR5 Table SM10-1 | Estimates of the welfare impact of climate change (expressed as an equivalent change in income, calculated as a percent of global aggregate GDP); estimates

Secondly, there is certain repetitiveness in the authors. Nordhaus alone is the source of almost 1/3rd of the studies (5 out of 18), Maddison of 3, and Tol, Mendelsohn, Hope and Rehdanz are in 2 each. These 18 studies thus emanate from just 8 superficially independent sets of researchers—but even that is an exaggeration. As Tol documents in the paper that was the precursor to this table (see Figure 7), most of these researchers are academically related to each other:

Nordhaus and Mendelsohn are colleagues and collaborators at Yale University; at University College of London, Fankhauser, Maddison, and I all worked with David Pearce and one another, while Rehdanz was a student of Maddison and mine. (Tol 2009, p. 30)
 

Furthermore, Tol notes in the footnotes to his 2009 Table that "The numbers used by Hope (2006) are averages of previous estimates by Fankhauser and Tol" , while even the other well-known climate economist Nicholas Stern (whose results are not included in this tabulation) used Hope's paper as the basis of his research (Tol 2009, p. 31, Table 1, footnote f). Finally, Bosello and Roson have both co-authored numerous papers with Tol (Bosello, Roson et al. 2006; Bosello, Roson et al. 2007; Berrittella, Rehdanz et al. 2008; Bosello, Roson et al. 2008; Ronneberger, Berrittella et al. 2009).

Eight ostensibly separate research groups are thus just two: Boyer, Mendelsohn, Nordhaus, Schlesinger, Williams and Yang in one group; and Bosello, Fankhauser, Hope, Maddison, Plambeck, Rehdanz, Roson, Tol and van der Mensbrugghe in the other. The two groups are hardly strangers to each other either: this is a tiny set of like-minded individuals clearly subject to "intellectual inbreeding", a potential that Tol acknowledges:

Moreover, it is quite possible that the estimates are not independent, as there are only a relatively small number of studies, based on similar data, by authors who know each other well. (Tol 2009, p. 37)
 

Figure 7: Tol's 2009 summary of economic estimates of damages from climate change

A larger group of researchers have produced estimates of what economists call the "social cost of carbon", but Tol notes that these all used the numbers generated in the studies shown in Figure 7. Even though this group is more numerous, Tol remarked that "it is still a reasonably small and close-knit community who may be subject to group-think, peer pressure, and self-censoring" (Tol 2009, p. 43).

Indeed. Mainstream economists—"Neoclassical economists" in the tribal language of the discipline—pride themselves on modelling human beings as hyper-rational individuals, and see their discipline as a rational cut above all other social sciences as well (Lazear 2000). But these "irrational", atavistic, socially conformist behaviours of "group-think, peer pressure, and self-censoring" are the only factors that can make sense of the failure of otherwise highly intelligent people to realise that their much-vaunted climate change data is based on arrant nonsense.

The data you use when you don't have data

A famous Australian 1970s commercial, for a non-alcoholic beverage called Claytons, described it as "the drink you have when you're not having a drink". The product is long gone, but the expression of "a Claytons" lingers on in Australian vernacular as a tag, normally derogatory, for something that is not what it pretends to be.

"Claytons Data" is a polite description of the numbers shown in Figure 6 and Figure 7.

Figure 6 classifies the methods used to generate these numbers as "Enumeration", "Expert elicitation", "Statistical" and "CGE". The one of these which superficially implies the most objectivity—"Statistical"—is the best starting point for showing how invalid all this "data" is.

Lies, Damned Lies, and Statistics

There are some events whose significance is so great that when and where you were when it happened becomes hard-wired into your brain. This includes collective events, from good ones like the end of WWII or the Moon Landing (for those old enough to remember), to bad ones like 9/11. And it includes personal events, from the good (like when and where you met your life partner), to the bad (like, for many Americans, when and where they were when they realised they were bankrupt during the Subprime Crisis).

As an intellectual, I'm also blessed—and cursed—with remembering when and where I was when major intellectual revelations occurred to me.

Some of the time, it's a good, "Eureka!" moment. One example (chosen just in case you think I'm simply a typical leftie intellectual) is when and where I realised the key fallacy in Marxian economics—the false argument that all "surplus" (the source of profit) comes from labour (Keen 1993; Keen 1993).

But all too frequently, given the rotten foundations of conventional economic theory, it's a bad experience: it's a moment of sheer disbelief that anything as stupid as what I had just read had in fact not only been written by another sentient being, but was taken seriously by the economics discipline, and would therefore affect policy in the real world; or the realisation that some data that I knew was significant was being ignored by policymakers, because an errant economic theory told them it wasn't important.

The moment when I realised what these "climate change economists" meant by "the statistical method" is seared into my brain as the worst of those latter moments. I knew then that mainstream economists had drastically underestimated the dangers of Climate Change, simply on the basis of an insanely stupid assumption, and that their delusion has helped delay humanity taking significant action until when there was virtually no chance of avoiding a climate crisis. It was one of the most shocking and depressing moments of my life.

I was reading Richard Tol's 2009 paper "The Economic Effects of Climate Change", and the section that triggered my reaction of shocked disbelief is highlighted in italics:

An alternative approach, exemplified in Mendelsohn's work (Mendelsohn, Morrison, Schlesinger, and Andronova, 2000; Mendelsohn, Schlesinger, and Williams, 2000) can be called the statistical approach. It is based on direct estimates of the welfare impacts, using observed variations (across space within a single country) in prices and expenditures to discern the effect of climate. Mendelsohn assumes that the observed variation of economic activity with climate over space holds over time as well; and uses climate models to estimate the future effect of climate change. (Tol 2009, p. 32)
 

Bear with me, if it isn't also instantly obvious to you why this is an insanely stupid assumption.

First, just think what it would mean if this assumption were true.

Within a country like the United States, it is generally true that both very hot States and very cold states have a lower level of per capita income than median temperature States: Florida (average temperature 22.1°C) and North Dakota (average temperature 4.9°C), for example, have lower per capita incomes than New York (average temperature 9.0°C). But the difference in average temperatures is far from the only reason the differences in income, and in the greater scheme of things, the differences are trivial anyway: as American States, at the global level they are all in the high per capita income range (respectively $43,000, $67,000 and $73,500 per annum).

So if you do a statistical study of the relationship between "Gross State Product" (GSP) per capita and temperature using GSP, population and average temperature for every State in the continental USA, you're going to find a weak, nonlinear relationship, with GSP per capita rising from low temperatures, peaking at medium ones, and falling at higher temperatures. If you then assume that this same relationship between GDP and temperature will apply as global temperatures rise with Global Warming, you're going to conclude that Global Warming will have a trivial impact on global GDP.

I've done this in Figure 8, using the data assembled in Table 2 on page 25. The dots show the deviations by State of per capita income and temperature from the USA averages of US$56,593 and 11.67°C respectively. The curve is a quadratic—the same function that Nordhaus used to estimate the damage from Global Warming, the results of which Weitzman reproduced in Figure 4. This curve can be used to predict how much GDP will fall given an increase in temperature over existing levels, assuming, as Tol put it, "that the observed variation of economic activity with climate over space holds over time as well" (Tol 2009, p. 32).

The function predicts that per capita GDP will fall by $131 times the change in temperature squared. A 9°C increase in temperature over 2018 levels—which is roughly a 10°C increase in temperature over pre-industrial levels—would therefore reduce GDP per capita by $10,630. Given that the US average is US$55,593, this is a 19.1% fall in GDP—which is almost precisely the estimate given by Nordhaus's DICE model back in 2011, as reproduced by Weitzman in Figure 4 (see also Table 1).

Figure 8: GDP and Temperature deviations from USA averages, fitted by a quadratic

Table 1 shows that the all the "predictions" of this regression are in the same ballpark as those from Nordhaus's DICE model circa 2011, and from the 2014 IPCC Report itself, where "estimates of global annual economic losses for additional temperature increases of ~2°C are between 0.2 and 2.0% of income" (Arent, Tol et al. 2014, p. 663).

Table 1: Damage estimates from Nordhaus's DICE in 2011 versus regression on State deviations from USA GDP per capita & Temperature averages

Word's HTML export stuffed up this table, but it's readable in the attached PDF

°C increase over pre-industrial


2


4


6


8


10


12

 

Nordhaus Function


1%


4%


8%


13%


19%


26%

 

Figure 8 Function


-0.2%


-2.1%


-5.9%


-11.6%


-19.1%


-28.6%

This ballpark is distinctly Minor League. If a 4°C increase in global temperature over pre-industrial levels is only going to reduce GDP per capita by between 2% and 4%, and this damage will occur in the distant future, why would you even bother worrying about Global Warming today?

Because, in a word, this analysis is nuts! It is crazy to think that today temperatureàGDP relationship tells you anything at all about what will happen to the economy via Global Warming. Though the proposition is dressed up in superficially academic language like "Mendelsohn assumes that the observed variation of economic activity with climate over space holds over time as well" (Tol 2009, p. 32), it's equivalent to saying that it doesn't matter how temperature is changed—whether by changing location, or by changing the amount of energy in the biosphere—it's all just a change in temperature, and that's all that matters.

This, to repeat myself, is nuts. There are at least three major factors that determine temperature on Planet Earth: Earth's location relative to the Sun; the gases in the atmosphere that retain some of the Sun's heat; and the latitude of a location on the planet (plus other locational details such as altitude, distance from the ocean, etc.). The assumption that "variation of economic activity with climate over space holds over time as well" ignores the first two factors, considered only changes in the third factor, and then assumed that the impact of changes in the other two factors would be the same.

This is nonsense. These three factors have dramatically different ranges, and dramatically different effects.

The first effect is that the Earth's orbit around the Sun places it in the habitable or "Goldilocks" zone where H20 can exist in its liquid form—otherwise known as water—and this is essential for the existence of life. No lifeform on the planet (and, scientists believe, no lifeform anywhere else in the Universe) can function without liquid H20.

Given the reflectivity of the Earth itself, this "Goldilocks Effect" gives the planet a base average temperature of minus 18°C in the absence of Greenhouse Gases. The Earth's location in this zone isn't fixed, since there are factors that cause perturbations in its orbit, and over time, the habitable zone is moving out, as the Sun ages and expands. The latter effect is not significant on a timescale of less than hundreds of millions of years, but orbital variations do cause changes in global temperature.

Greenhouse gases firstly absorb and then re-emit some of the radiation that would otherwise be reflected back from the surface of the planet into space. Naturally occurring Greenhouse gases fluctuate over time, but immediately before the Industrial Age began, these gases made the average global temperature about +15°C, which is 33°C higher than it would be in the absence of the Greenhouse Effect.

These two factors together have caused the variations in the biosphere's average temperature over time. Over the last 800,000 years for which we have reliable data (from actual measurements since 1880, and from derived data before then), the range has been from 5°C below the 1951-1980 average to just under 3°C above it. We can call the sum of these two effects "Global Temperature", since these two factors together affect the entire planet's temperature.

Figure 9: Fluctuations in global average temperature over the last 800,000 years

Finally, the closer a location is to the Equator, the hotter it is likely to be, since Equatorial regions are more aligned to the Sun's rays than Polar regions. This is the one factor whose variations are captured in the data on temperature that Mendelsohn, Nordhaus and several others used to "predict" the economic consequences of climate change. They have since generalized this technique to the global economy, but the initial analysis was done just using US average temperature data, where the range of temperatures is from 5°C in North Dakota to 22°C in Florida—a 17°C range in what we can call "Local Temperature", compared to the historic 8°C range in Global Temperature.

It's easy to illustrate why the temperatureàGDP relationship Nordhaus derived is nonsense when applied to Global Warming, by setting up a hypothetical statistical test in which the causal relationship between temperature and GDP is considered in relation to both the Global and Local factors that determine temperature. I'm going to pretend (as Nordhaus himself did: Nordhaus and Sztorc 2013, p. 11) that the relationship between GDP and temperature is quadratic, where GDP of a given location is determined by the change in temperature squared—but this time there are two quadratics, one for changes in Global Temperature (relative to the pre-Industrial global average of +15°C), and another for changes in Local Temperature (relative to the current USA average of 12°C).

Given the assumption that the relationship between temperature and GDP is quadratic, the only thing to be calculated is the value of the coefficients for the two quadratics. Let's call the first coefficient and the second , and let's use to signify the difference between global temperature at some point in time and pre-industrial global temperature, and to signify the difference between the average temperature of a given State, and the average temperature of the USA. The basic proposition being tested is that GDP per capita () at a given time and location is a function of global and local temperature deviations at that time and location:

Doing a statistical fit of the data points in Figure 8 using this equation will give you an estimate for , but—to labour the obvious—no information on , because the global temperature is the same for all locations.

What Mendelsohn and Nordhaus effectively did is assume that and were identical: the assumption "that the observed variation of economic activity with climate over space holds over time as well" is literally assuming that .

It can't be said often enough: this is nuts!

Imagine applying the same argument to hiking and altitude: that whichever way you walk, it's all just movement, so the direction in which you walk doesn't matter, your altitude will change by the same amount. Imagine that you have no data on the slope of a mountain in East-West direction, but you know that it's quite flat in the North-South direction. Does that mean that it is therefore safe to walk East-West? Of course not! You may be lucky enough to find yourself on a perfectly symmetrical mountain, and thus survive this idiotic thought. But what if you are on, for example, El Capitan?

Figure 10: El Capitan. Nice and flat North South, rather different East-West

In other words, "distance isn't just distance": it does matter in which direction you walk. The same applies here: "temperature isn't just temperature": it matters not only how warm it is, but how you change that warmth. Nordhaus's damage function, and that of all the "Integrated Assessment Models" (IAMs) that used this data, tells us absolutely nothing about how GDP will be affected as global temperatures rise.

One reason that economists have been able to get away with this ridiculous assumption—apart from the fact that most people simply trust that they have done their work well, and therefore don't analyse their work as closely as I am doing here—is that humanity has no experience of temperatures that far above current levels. We can't know what such a world will be like by extrapolating from our current situation. We simply don't know what a world 3°C warmer than today—Nordhaus's "optimal" temperature target, given that we're already about 1°C above pre-industrial levels—would be like.

However, we do know what the world was like when it was 4°C colder than pre-industrial levels, since that was the temperature 20,000 years ago, during the last Ice Age—see Figure 11. Ironically, this lets us assess the realism of Nordhaus's predictions for the economic impact of Global Warming, because his damage function is symmetrical: as a quadratic, it predicts precisely the same level of damage to GDP from a fall of, for example, 4°C in the global average temperature, as it does for a rise of 4°C.

Figure 11: Earth's average temperature over the last 22,000 years

According to Nordhaus's "Damage Function", if the world were in fact facing Global Cooling of 4°C, rather than Global Warming of the same magnitude, a 4°C fall in temperature would reduce global GDP by 4%.

Really? What would such a world look like? Unlike Global Warming, we know, because it happened just 20,000 years ago. This a brief enough period on a geological timescale for incontrovertible evidence of what that world was like to be easily discovered by geologists today. Greenland-style ice sheets then covered all of modern-day Canada, much of the northern USA (including Chicago and New York), most of the UK, and most of Europe north of Berlin (see Figure 17 on page 18, and Figure 18 on page 25). Global Cooling of 4°C relative to pre-industrial levels would ultimately reproduce those ice sheets.

Figure 12: Northern European ice sheets during the Last Glacial Maximum, about 20,000 years ago (Svendsen, Alexanderson et al. 2004, Figure 16, p. 1257)

The final outcome of a kilometre or more of ice obliterating Canada, Ireland, Scotland, northern Germany, Poland, Denmark, Norway, Sweden, etc., would of course take many centuries, but the chaos such a process would trigger would be immediate. A 4°C fall in global temperature at the same speed as Global Warming itself is expected to happen would trigger a human exodus from high latitude countries, destroy large slabs of the planet's industrial capacity, drastically alter the fertility of current agricultural land via huge changes in temperature and precipitation (with the rate of change possibly pushing many of the plant and animal species on which humans rely into extinction themselves), and drive what was left of humanity into the currently underdeveloped tropical regions of the planet—with all the political and military conflicts that such a shift would trigger.

Figure 13: North American Ice sheets during Last Glacial Maximum (Hughes 2013, Figure 1, p. 31)

The assertion that such a transformation would reduce global GDP by a mere 4%, compared to a world in which Global Cooing did not occur—the very idea that the conditions of that world could be extrapolated from the conditions that prevail today—is absurd.

The same applies to the actual situation we face of Global Warming. You simply can't extrapolate from today's temperature and GDP data to predict what GDP would be in a 4°C warmer world. The very fact that these Neoclassical economists even considered using statistical techniques to predict the consequences of Global Warming shows that they don't appreciate what Global Warming is in the first place.

The only way that I can make sense of them even attempting to do so is that they were far more fixated on playing with their statistical methodology (known as "econometrics") than they were on the topic itself, ignoring whether those techniques were applicable to the data, let alone whether the "data" they used was relevant to the topic.

This impression was strengthened via a tweet from Richard Tol, announcing a new paper that estimated a 7.22% fall in GDP by 2100—relative to what it would be in the complete absence of Global Warming—from an annual increase in global temperature till then of 0.04°C (see Figure 14).

Figure 14: Tol's tweet promoting Khan, Mohaddes at al.'s paper giving estimates of the damage from Global Warming (Kahn, Mohaddes et al. 2019)

My first reaction was to laugh at the faux precision: predicting that a further 3.44°C increase in global temperature (0.04°C a year for 86 years from their base year of 2014) would cause a 7.22% fall in GDP in 2100, relative to a world without Global Warming? When today's estimates of today's GDP are to only one digit of accuracy, and even those are frequently subsequently revised by several percentage points? Any decent statistician knows to round estimates to the level of the reliability of the data—though even a 7% estimate would be in the same ballpark as existing "statistical" estimates, and therefore prima facie absurd. My second was ennui: having waded through Nordhaus and Mendelsohn's delusional statistical papers, did I really have to wade through yet another one, especially since it hadn't yet been published in a journal? Its status as a Dallas Federal Reserve working paper (Globalization Institute Working Paper 365) motivated me to overcome the ennui and read it.

"Long-Term Macroeconomic Effects of Climate Change: A Cross-Country Analysis" (Kahn, Mohaddes et al. 2019) had at least one redeeming feature compared to its predecessors in the "statistical" literature: it acknowledged that assuming that today's temperatureàGDP data could be used to predict the impact of climate change ignored that change over time was very different to change over space:

there are a number of grounds on which the econometric evidence of the effects of climate change on growth may be questioned. Firstly, the literature relies primarily on the cross-sectional approach (see, for instance, Sachs and Warner 1997, Gallup et al. 1999, Nordhaus 2006, and Dell et al. 2009), and as such does not take into account the time dimension of the data (i.e., assumes that the observed relationship across countries holds over time as well) (Kahn, Mohaddes et al. 2019, p. 2. Emphasis added).
 

So at least this paper was going to take time into account—on the surface, a plus compared to the existing literature. But that plus soon evaporated when I saw how they had purported to take time into account: they had taken data on GDP and temperature change between 1960 and 2014, fed those trends into an economic model in which deviations of temperature from historic norms affected labor productivity, and used that model to extrapolate the 1960-2014 relationship forward for the next 86 years:

This paper investigates the long-term macroeconomic effects of climate change across 174 countries over the period 1960 to 2014… Climate change could affect the level of output … through … lower labour productivity. We … develop a theoretical growth model that links deviations of climate variables from their historical norms to changes in labour productivity …
 
Our theoretical model postulates that labour productivity in each country is affected by … country-specific climate variables… in a historically cold region, a rise in temperature above its historical norm might result in higher labour productivity, whilst for a dry region, a fall in precipitation below its historical norms is likely to have adverse effects on labour productivity…

 
We start by documenting that the global average temperature has risen by 0.0181 degrees Celsius per year over the last half century …
 
Specifically, we show that if temperature rises above its historical norm by 0.01°C annually, income growth will be lower by 0.0543
percentage points per year…

 
We show that an increase in average global temperature of 0.04°C per year— corresponding to the Representative Concentration Pathway (RCP) 8.5 scenario, which assumes higher greenhouse gas emissions in the absence of mitigation policies— reduces world's real GDP per capita by 7.22 percent by 2100. (Kahn, Mohaddes et al. 2019, pp. 1-4)

 

My feeling of ennui gave way, once more, to incredulity. The change in temperature between 1960 and 2014 (1°C according to their estimate) was substantial, but it had not obviously triggered large-scale qualitative changes in fundamental features of the Earth's climate. But another 3.5°C increase—their estimate of the impact of rising Greenhouse Gases till 2100, given no mitigation efforts—surely would do so. Climate scientists have been considering this issue since 2008 (Lenton, Held et al. 2008), and by 2018 they had concluded that, for example, the Greenland and West Antarctic ice sheets would almost certainly collapse if temperatures rose another 2°C from current levels (Steffen, Rockström et al. 2018, Figure 3, p 8255).

Surely these economists took this likelihood of drastic qualitative change in the planet's climate into account? Unfortunately, no. There was a wealth of discussion of econometric techniques to cope with trends in data, the right type of lags to choose in setting up a model, how to cope with endogenous versus exogenous variables… but no discussion of whether the level of temperature change they were contemplating might make enormous qualitative differences to the planet's climate that would render irrelevant their quasi-linear extrapolations from current data. I searched in vain for the word "tipping", for any consideration whatsoever that such a huge increase in the retained energy in the biosphere would make any qualitative changes that could render their extrapolations useless.

Figure 15: Figure 2 from (Kahn, Mohaddes et al. 2019, p. 6), showing previous economists' estimates of climate change damages versus their own (shaded area).

This made their otherwise praiseworthy attempt to include change in global temperatures as well as comparisons of local temperatures useless. It was as if they'd done a study of ice-skating speeds, concluded that speed skaters move 1% faster for every 1°C increase in the temperature of the ice, and therefore predicted a 4% increase in skating times if the temperature of the ice increased from minus 2°C to plus 2°C.

Reading Kahn and Mohaddes (2019) alerted me to one additional study in this "statistical" tradition that prima facie was worth examining—if only because its predictions of a 23% hit to GDP from a 4°C increase in global temperature was in a different ballpark to the other economic studies: "Global non-linear effect of temperature on economic production" by Burke, Hsiang and Miguel (Burke, Hsiang et al. 2015).

The study was certainly superior to its distant companions in Figure 15 by acknowledging that the impact of global warming on economic output was highly nonlinear:

Numerous basic productive components of an economy display a highly non-linear relationship with daily or hourly temperature. For example, labour supply, labour productivity, and crop yields all decline abruptly beyond temperature thresholds located between 20°C and 30°C. (Burke, Craxton et al. 2016, p. 235)
 

But it shared the same defect of taking functions fitted to today's nonlinear relationships between temperature and GDP (between the years of 1960 and 2010), and then extrapolating those to predict the impact of Global Warming. If these current nonlinear relationships hold in a 4°C warmer world, then this would be a useful exercise: but a world where the biosphere's retained energy levels have risen by that much will undoubtedly be ruled by entirely different nonlinear relationships.

Figure 16: Burke et al's Figure 2, showing the nonlinear impact of temperature variations between 1960 and 2010 on output between 1960 and 2010 (Burke, Craxton et al. 2016, p. 236)

Data Appendix

Figure 17: The unsustainable trends in private debt and credit in the USA that alerted me to the inevitability of a financial crisis

Figure 18: The correlation of credit and unemployment, which is still ignored by mainstream economists

Temperature and GDP per capita by State data for the USA

A summary of average temperatures by State data is available at http://www.usa.com/rank/us--average-temperature--state-rank.htm. Statistics can be derived for every State on a one-State-per-query basis from https://www.ncdc.noaa.gov/cag/statewide/time-series.

The Bureau of Economic Analysis has data on per capita real GDP by State in the series SAGDP10N, which can be accessed via https://apps.bea.gov/iTable/index_regional.cfm.

Table 2: Population, average temperature & 2018 GDP per capita data for the Contiguous USA States

        

State


Population


Temp °C


GDP2018pc

 

Alabama


4817678


17.0


40279

 

Arizona


6561516


18.9


43096

 

Arkansas


2947036


15.6


38467

 

California


38066920


16.2


67698

 

Colorado


5197580


7.9


59057

 

Connecticut


3592053


9.2


67784

 

Delaware


917060


12.5


66023

 

Florida


19361792


22.1


43052

 

Georgia


9907756


16.8


49663

 

Idaho


1599464


8.0


39843

 

Illinois


12868747


10.8


59980

 

Indiana


6542411


11.0


48738

 

Iowa


3078116


9.0


54101

 

Kansas


2882946


12.6


52297

 

Kentucky


4383272


13.1


41659

 

Louisiana


4601049


19.3


49606

 

Maine


1328535


6.2


42356

 

Maryland


5887776


12.6


60886

 

Massachusetts


6657291


9.0


72635

 

Michigan


9889024


8.1


46858

 

Minnesota


5383661


6.1


59057

 

Mississippi


2984345


17.5


34029

 

Missouri


6028076


12.6


46064

 

Montana


1006370


6.8


42173

 

Nebraska


1855617


9.9


58141

 

Nevada


2761584


14.1


48189

 

New Hampshire


1321069


6.9


55744

 

New Jersey


8874374


11.1


62263

 

New Mexico


2080085


11.7


44187

 

New York


19594330


9.0


73531

 

North Carolina


9750405


14.8


47778

 

North Dakota


704925


4.9


67308

 

Ohio


11560380


10.5


51456

 

Oklahoma


3818851


15.6


48954

 

Oregon


3900343


10.7


50996

 

Pennsylvania


12758729


9.9


55602

 

Rhode Island


1053252


9.6


50827

 

South Carolina


4727273


16.5


39883

 

South Dakota


834708


7.5


51997

 

Tennessee


6451365


14.5


47695

 

Texas


26092033


18.9


58417

 

Utah


2858111


9.8


49740

 

Vermont


626358


6.5


47921

 

Virginia


8185131


13.1


55929

 

Washington


6899123


10.3


67242

 

West Virginia


1853881


11.7


39495

 

Wisconsin


5724692


7.1


51575

 

Wyoming


575251


6.4


66413

References

Arent, D. J., R. S. J. Tol, et al. (2014). Key economic sectors and services. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. C. B. Field, V. R. Barros, D. J. Dokkenet al. Cambridge, United Kingdom, Cambridge University Press: 659-708.

Berrittella, M., K. Rehdanz, et al. (2008). "The Impact of Trade Liberalization on Water Use: A Computable General Equilibrium Analysis." Journal of Economic Integration
23(3): 631-655.

Bosello, F., R. Roson, et al. (2007). "Economy-wide Estimates of the Implications of Climate Change: Sea Level Rise." Environ Resource Econ
37(3): 549-571.

Bosello, F., R. Roson, et al. (2006). "Economy-wide estimates of the implications of climate change: Human health." Ecological Economics
58(3): 579-591.

Bosello, F., R. Roson, et al. (2008). "Economy-wide estimates of the implications of climate change – a rejoinder." Ecological Economics
66(1): 14-15.

Burke, M., M. Craxton, et al. (2016). "Opportunities for advances in climate change economics." Science
352(6283): 292-293.

Burke, M., S. M. Hsiang, et al. (2015). "Global non-linear effect of temperature on economic production." Nature
527(7577): 235.

Field, C. B., V. R. Barros, et al. (2014). IPCC, 2014: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom, Cambridge University Press, .

Hughes, P. D. (2013). 8.3 Quaternary-Pleistocene Glacial and Periglacial Environments, Elsevier Inc.

Kahn, M. E., K. Mohaddes, et al. (2019) "Long-Term Macroeconomic Effects of Climate Change: A Cross-Country Analysis." DOI: https://doi.org/10.24149/gwp365.

Keen, S. (1993). "The Misinterpretation of Marx's Theory of Value." Journal of the History of Economic Thought
15(2): 282-300.

Keen, S. (1993). "Use-Value, Exchange Value, and the Demise of Marx's Labor Theory of Value." Journal of the History of Economic Thought
15(1): 107-121.

Lazear, E. P. (2000). "Economic Imperialism." Quarterly Journal of Economics
115(1): 99-146.

Lenton, T. M., H. Held, et al. (2008). "Tipping elements in the Earth's climate system." Proceedings of the National Academy of Sciences
105(6): 1786-1793.

Lomborg, B. (2019). On climate change, humanity is not 'evil'. The Globe and Mail. Toronto.

Nordhaus, W. (2018). "Nobel Lecture in Economic Sciences. Climate Change: The Ultimate Challenge for Economics." from https://www.nobelprize.org/uploads/2018/10/nordhaus-slides.pdf.

Nordhaus, W. and P. Sztorc (2013). DICE 2013R: Introduction and User's Manual.

Nordhaus, W. D. (2007). "A Review of the "Stern Review on the Economics of Climate Change"." Journal of Economic Literature
45(3): 686-702.

Ronneberger, K., M. Berrittella, et al. (2009). "[email protected]: Introducing Biophysical Aspects of Land-Use Decisions into a Computable General Equilibrium Model a Coupling Experiment." Environ Model Assess
14(2): 149-168.

Steffen, W., J. Rockström, et al. (2018). "Trajectories of the Earth System in the Anthropocene." Proceedings of the National Academy of Sciences
115(33): 8252-8259.

Svendsen, J., H. Alexanderson, et al. (2004). "Late quaternary ice sheet history of northern Eurasia." Quaternary Science Reviews
23(11-13): 1229-1271.

Tol, R. S. J. (2009). "The Economic Effects of Climate Change." The Journal of Economic Perspectives
23(2): 29–51.

Weitzman, M. L. (2011). "Fat-Tailed Uncertainty in the Economics of Catastrophic Climate Change." Review of Environmental Economics and Policy
5(2): 275-292.

.

Become a patron to

276
Unlock 276 exclusive posts
Listen anywhere
Connect via private message