David R. Henderson  

Nordhaus's Global Warming Graph

Bob Murphy on the Heartland In... Leveraging and De-leveraging...

In a recent article in the New York Review of Books, Yale University economist William Nordhaus presents a graph showing changes in global mean temperatures from 1880 to 2011. Take a look. The data appear to be annual and he is graphing changes in temperature, not absolute level. So at first I thought he must mean that each year, the temperature changed by the amount on the graph compared to the last year. So, for example, that would mean that in 1945 (I'm eyeballing it here), where there is a huge spike, the temperature changed by almost 0.5 degrees C in one year.

But a little reflection shows that that can't be the case. (I'm not assuming that Nordhaus thinks it is the case. I'm simply clarifying for people who initially have my problem with understanding the graph.) Here's why. Look at the changes from about 1950 to 2011. They average at least 0.4 degrees C. So that would mean that the world has warmed by 60 times 0.4 degrees C or 24 degrees C! Clearly that hasn't happened. QED.

So what does all this mean? It must mean that Nordhaus is not graphing changes in temperature from the year before, but, rather, changes from some baseline temperature. It's not clear what baseline he has in mind. One might think 1880, but it's not clear. Read this way, what his graph says is that in 2011, the global mean temperature was higher by 0.8 degree C than it was in 1880 or some year earlier than 1880. Notice that that interpretation is consistent with two other well-agreed-on claims: (1) global temperatures are about 0.8 degrees C higher than they were 80 years ago, and (2) global temperatures are slightly lower than they were 10 years ago.

HT to Tyler Cowen.

Comments and Sharing

COMMENTS (19 to date)
Kevin Dick writes:

I think the term of art you're looking for is "temperature anomaly". This is the difference between the "average" temperature in a given year and the "average" temperature in a base period.

The average temperature in a give year is the product of a fairly complicated statistical procedure. The average temperature for the base period is the mean over many individual years, often 30.

IIRC, the base period for the instrumental record is typically 1961-1990 and for the satellite record is 1991-2010.

The calculation of the average temperature over a geographical area has a lot of steps to account for such issues as changes in thermometer technology, movement of measurement sites, additions and subtractions of measurement sites, and changes in surrounding land use.

It's actually pretty unclear what the reliable resolution of the temperature anomaly measurements truly is. IIRC, an economist in Canada, Ross McKitrick showed a strong correlation between positive local temperature anomalies and local economic growth.

Matt C writes:

It's nice to see Nordhaus present something that I could actually read through without quitting in disgust.

Maybe the graph is a little confusing, but the tone is civil and he proceeds with something like a reasoned argument. It is far better than most of what I bump into from global warming spokespeople. We might actually get an interesting conversation out of this one.

David R. Henderson writes:

@Kevin Dick,
@Matt C.,
Absolutely. That is how I have always perceived Bill Nordhaus. It was unfortunate that the 16 went overboard by getting into Lysenkoism. The actual example they gave that preceded that charge was pretty upsetting and I notice that Nordhaus didn’t respond to it. Rule for arguers: If you exaggerate, you can be guaranteed that your critic will pounce on the exaggeration.

mick writes:

Splendid graph of the young earth creationist mentality of the global warming movement.

Tom Dougherty writes:

The interesting question is not if the earth has warmed since the end of the little ice age (1850) but whether the warming is natural or manmade. If we break up the graph into 30 year intervals, from 1880 to 1910 we see little warming; from 1910 to about 1940 there is warming cycle; from 1940 to 1970 again there is little warming; and from 1970 to 2000 another warming cycle. The overall trend is a warming trend. But compare the recent warming period of 1970 to 2000 to that of 1910 to 1940. They are remarkably similar which says the warming is natural and not manmade. There are 7 times as much greenhouse gas emissions from 1970 to 2000 as there were from 1910 to 1940, however, the rate of warming during those two periods are remarkably similar.

Eelco Hoogendoorn writes:

Yes, that graph is temperature anomality relative to some mostly irrelevant baseline. Peg the 2012 value to the current temperature graph, and youve got your absolute temperatures. Add in the error bars, and youve got nothing :).

@Matt: im not familiar with Nordhaus, so I dont know how bad it usually is, but this article does very much curl my toes.

Quick rebuttal:

1: strawman. Nobody seriously disputes the graph presented, minor quibbles aside. Nonetheless, the actually claimed lack of warming over the past decade is indeed visible; which is more prominent in more detailed analyses. The relevance thereof can be disputed (though it would be nice if warmists remembered to hold themselves to the same standards come the next warm summer), but there is no doubt temperature developments have consistently 'underperformed' predictions.

2: "In reviewing the results, the IPCC report concluded: “No climate model using natural forcings [i.e., natural warming factors] alone has reproduced the observed global warming trend in the second half of the twentieth century.”".

In other words; I dont know how the universe was created, so the people who claim god did it must be right.

Yes, temperature is rising. At a completely unremarkable pace. No, I dont know why, every bit as much as I dont know why it happened many times before, long before human greenhouse gasses existed.

3: fair but minor rethorical point.

4: interpreting an analogy literally: another strawman I guess, or is there a more specific term for that fallacy? Intention obtuseness? No, nobody has been sent to the gulag. But the climategate emails show crystal clear evidence of systematic and blatant corruption of the scientific publication process in action. Anyone who is even surprised by this clearly never worked in an academic job.

5: " The big money in climate change involves firms, industries, and individuals who worry that their economic interests will be harmed by policies to slow climate change.". Huh? The ratio of industry support for researchers versus government money given out under the implicit condition that there is a problem to worry about to begin with that warrants this expenditure, is at least 1:100.

More importantly: indeed science is usually not driven my money, but even worse, by ego. Believing to be saving the world is much more tempting than believing you are not making any difference, especially if you cant justify your activities by the salary you get paid for it.

6: "This example shows why we should, in designing the most effective policies, look at benefits minus costs, not benefits divided by costs."

This is just grasping at straws. Benefits minus costs are negative, and benefits divided by costs are smaller than one*. Both contain the same information content.

*(im sure he showed otherwise, but its hard to consider his finding to be anything other than the outlier it is, considering the track record of intellectual honesty he manages to put on display in a single article)

[Minor edit--Econlib Ed.]

DK writes:

Even that graph is not convincing. It clearly shows flat 30 years in the past (all the while GHG output was steadily rising) and offers no reason to think that we are currently not in another flat period that can last 50 years.

happyjuggler0 writes:

From the Nordhaus link:

We do not need any complicated statistical analysis to see that temperatures are rising, and furthermore that they are higher in the last decade than they were in earlier decades.3

The first part of that sentence incorrectly uses past data to say that something is happening right now in the present tense, which is an unsupportable tense transposition. This may be a minor gaffe on his part, but it is worth noting that the proper way to read the graph is to say that "temperatures were rising. See what he did there?

The second part they are higher in the last decade than they were in earlier decades is something lots of CAGW advocates like to state as if it is important. If I put a pot of water on the stove and turn the heat all the way up, and then the water eventually comes to a boil, and then I turn off the heat, then yes, sure the last ten seconds of water temperature were hotter than earlier ten second periods, but this says nothing about whether or not to expect future higher temperatures! It is merely a function of forcings that occurred in the past. To assume that this means that GGH's are the cause of the past warming is to assume one's conclusion. In other words, it says nothing useful about the future or what caused the past.

To see how unsupportive the line of they are higher in the last decade than they were in earlier decades is, just change "global temperature" to "stock prices", and use the 1990 to 2000 time frame for the Nasdaq. If someone used that ten year reference to try to persuade you to buy stocks, would you be persuaded? (In hindsite the answer is obvious, but they are trying the same trick with CAGW. How about stocks in Greece, would you buy them today if you saw a chart like Nordhaus', or would you use forward looking indicators, such as pure fiscal insanity ahead for the the foreseeable decades?

In 1998, an extreme El Niño year, the predictions of the CAGW types were for yet more extreme temperatures, but they didn't happen. Instead there is a heat budget gap that is unexplained by the CAGW hypothesis, which they routinely try to finesse away by ignoring it in articles like like this.

David R. Henderson writes:

@Eelco Hogendoorn, DK, and happyjuggler0,
Good points, especially happyjuggler0’s point about verb tense.

happyjuggler0 writes:



I just want to correct a minor goof on my part. GGH should read GHG, a popular acronym for GreenHouse Gas[ses].

John Cunningham writes:

The AGW hoaxers constantly pretend to temperature numbers with unsupported details. consider the temp graph in Nordhaus's piece. I agree that the earth is in a long-term warming trend since 1850, but there has been massive chicanery in developing all the long-term temp records: those of Hansen at GISS, the East anglia hoaxers at HADCRUT, the GHCN series. There is an eye-opening analysis of temp records by an aerospace engineer, AJ Strata, at http://strata-sphere.com/blog/index.php/archives/18198 Essentially, all the temp series have downward adjustments for most or all temp records before 1940. none of the hoaxers ever explain why they only lower past temp records.

David Friedman writes:

He doesn't seem to deal with what I consider the weakest link in the argument--the claim that warming on the scale projected will have large net negative consequences. There is very little a priori reason to expect that, since earth's climate was not designed for our benefit and humans live successfully across a wide range of climates.

The usual argument comes down to adding up positive and negative externalities and claiming that the sum is large and negative. The problem with that approach is that almost all the numbers are uncertain, making it easy to get whatever result you want by a suitable selection of judgement calls.

Nordhaus in particular, in a webbed version of his work that I recently discussed on my blog, produces a net negative result by asking various people how likely it is that there will be some low probability/high negative consequence result, and trying to estimate an expected value from their answers. He did not ask them the symmetric question—how likely a low probability high positive consequence was—although there is no obvious reason why that would be less likely to occur. So he is including negative externalities, leaving out positive ones, and, not surprisingly, getting a negative sum.

For more details, see:


MikeP writes:

He did not ask them the symmetric question...

He furthermore does not ask the other symmetric question -- how likely are negative externalities resulting from pursuing measures to prevent global warming.

As I put it five years ago in a comment elsewhere...

I do not actually advocate carbon taxes or tariffs. I do however accept that they may be the least bad result we can expect from the politically charged atmosphere that the global warming debate has become.

Nonetheless, I find that the meager economic benefit due to the optimal carbon tax, along with the meager decrease in warming resulting from that tax, make giving governments -- not only the US government, which is sadly one of the best governments out there, but every government -- the power to collect and spend this tax a far more costly and risky proposition than simply dealing with global warming.

In the twentieth century governments killed outright upwards of a hundred million people. The damages caused by global warming are peanuts compared to that. Addressing global warming in ways that hinder economic growth and global interdependence while constructing new strategic and trade blocks merely to try to deal with carbon emissions would be a serious disaster.

MikeP writes:

It is very important to recognize that Nordhaus's work finds that the second best policy to a rising globally harmonized optimal carbon tax is to do nothing at all. And when you graph the curves next to each other, the optimal policy looks mostly like a refinement to the do-nothing policy.

Yet when governments propose or execute policies, they inevitably look like Kyoto or Stern or Waxman-Markey. These come out phenomenally worse in the cost-benefit analysis than doing nothing does.

So which is the best course on global warming going forward? Imagining governments are actually going to execute the optimal policy correctly over a century without geopolitical disaster? Or keeping the reasons and the powers out of the governments' hands and living with the modest predicted damages?

Mark Bahner writes:
It's actually pretty unclear what the reliable resolution of the temperature anomaly measurements truly is. IIRC, an economist in Canada, Ross McKitrick showed a strong correlation between positive local temperature anomalies and local economic growth.

Since the warming trend in the lower troposphere from satellite measurements is very similar to the warming trend from surface temperature measurements, it's logical to think that the measured surface temperature warming trend is reasonably accurate.

Mark Bahner writes:
2) global temperatures are slightly lower than they were 10 years ago.

The best way to evaluate temperature measurements is not to look at a particular year or short period, but to look at 10-, 20-, or 30-year running averages (or some similar long-term rolling average).

If one does that, it's clear that temperatures are rising (for example, the 20-year average for 1992 to 2012 is higher than the 20-year average from 1982 to 2002, and so on).

Mark Bahner writes:
He doesn't seem to deal with what I consider the weakest link in the argument--the claim that warming on the scale projected will have large net negative consequences.

There's another important consideration left out of the Nordhaus analysis: even if global warming resulted in, say a 5% decline in world GDP in the year 2100 versus what it would be if there were no warming, it is *still* highly morally questionable whether the people of today should make any sacrifice at all to spare the people of the year 2100.

This is because the world in 2100 will have a much higher per-capita GDP than today, even including any forseeable negative effects of climate change.

So to make any sacrifices for the people of 2100 means that the poor must sacrifice for the rich.

Here's a quote from an analysis by Indur Goklany:

Under the IPCC’s highest growth scenario, by 2100 GDP per capita in poor countries will be double the U.S.’s 2006 level, even taking into account any negative impact of climate change. (By 2200, it will be triple.) Yet that very same scenario is also the one that leads to the greatest rise in temperature—and is the one that has been used to justify all sorts of scare stories about the impact of climate change on the poor.

Misled on Climate Change

John Cunningham writes:

this is a detailed post showing systematic reduction in past temps across all stations in higher latitudes. by lowering past temps, one can create a spurious warming trend--

Bill writes:

I agree with Nordhaus' point that net present value is the preferred investment criterion (as opposed to the benefit-to cost-ratio.)However, when comparing investments of different sizes, i.e., requiring different initial investment amounts, it's important to "normalize" the comparison by asking what would be the productivity of the dollar amount NOT invested in the smaller project. For example, Norhaus compares project A, requiring an initial cost of $1 million and returning $10 million in benefits -- net benefits of $9 million vs. project B requiring $10 million initial investment and returning $50 million in benefits -- $40 million net benefits. To avoid his procedure that favors the larger project, one would want to know the productivity of the $9 million NOT spent on project A. E.J. Mishan has dealt with this issue in his many volumes on welfare economics / benefit cost analysis.

Comments for this entry have been closed
Return to top