In my last post, I discussed the Manhattan Institute’s Oren Cass’s May 16 testimony before the House Committee on Science, Space, and Technology in a hearing on climate change. Cass’s overall thesis is that many “official” estimates of potential climate change damage do not adequately incorporate human adaptability, and so they grossly exaggerate the likely harms.

In my post, I reviewed Cass’s argument about ozone and PM2.5, where the alleged dangers of unchecked warming will lead to higher atmospheric concentrations of these pollutants by the year 2100, despite the fact that they will at that point still be lower than they were in the year 2000. In effect, the studies warning of the dangers of more ozone and PM2.5 did not adequately take into account the dramatic reduction in these pollutants that has occurred just in the last two decades.

In the current post, I’ll summarize another of Cass’s excellent points, which is that researchers did not incorporate the documented decline of heat mortality rates into their projections, thus grossly exaggerating deaths from global warming.

Historical Heat Mortality Rates

In a previous IER post, I blogged about Oren Cass’s policy study when it first came out. In that post, I explained how some of the alarmist studies in this literature commit the simple fallacy of assuming that northern cities will be just as vulnerable to hot days in the year 2100 (when they are—in the computer simulations—more the norm) as they are right now, when they are still rare. Yet these projections assumed that, say, people in Pittsburgh in the year 2100 would suffer eighty times the mortality rate from heat as people in Houston in the year 2000.

As I pointed out, the problem here is that the researchers involved were not taking into account the fact that a northern city would adapt to become more like a southern city, if in fact that climate really changed as the computer simulations indicated. I used an analogy going the other way, showing that Atlanta would respond much better to snowstorms if they became the norm, because the city government would buy more plows and salt the roads, while the drivers would become better in the snow.

In the present post, let me move to a different line of research that Cass also critiques. Specifically, he analyzes the study by Olivier Deschênes and Michael Greenstone, “Climate Change, Mortality, and Adaptation: Evidence from Annual Fluctuations in Weather in the US,” Applied Economics 3, no. 4 (Oct. 2011): pp. 152–85. These authors predict a net increase of 63,000 annual deaths in the U.S. due to higher temperatures by the year 2100—this comes from 59,000 fewer cold-related deaths coupled with 123,000 more heat-related deaths.

Now the method behind the Deschênes-Greenstone calculation is better than a completely naïve approach in which a given city in the north is assumed to have the same vulnerability to an extreme heat day in the year 2100, even if such days become much more frequent by then.

However, even though the Deschênes-Greenstone study incorporates the adaptations already deployed geographically (by combining the mortality of extreme heat days across various U.S. cities), it does not adequately account for the improved adaptability we’ve observed over time (as I explain in the next section).

Things Get Better With Age

In a subsequent study a few years later (in which Deschênes and Greenstone were co-authors), the large estimates of future heat deaths were scaled back. Specifically, in Alan Barreca et al., “Adapting to Climate Change: The Remarkable Decline in the US Temperature-Mortality Relationship over the Twentieth Century,” Journal of Political Economy 124, no. 1 (Feb. 2016): pp. 105-59, the researchers found that heat mortality rates declined significantly as air conditioning and other adaptations were implemented. As Cass summarizes, the researchers found “the mortality effect [of extreme heat days] falling by half from the 1960–79 period to the 1980–2004 period.”

Once we take into account the fact that the historical mortality rates over a longer period are unnecessarily pessimistic—because they incorporate deaths during years when the U.S. had less extensive air conditioning—this obviously makes projected deaths from extreme heat in the year 2100 much lower. Again I quote from Cass’s discussion:

Applying the Deschênes-Greenstone estimate of 42.3 additional >90°F days by 2100, they estimated that climate change could cause roughly 60,000 additional deaths in 2100 at the 1960 level of air-conditioner adoption. But at the 2004 level of air-conditioner adoption, “the null hypothesis that additional 80°F–89°F and >90°F days would have no impact on mortality cannot be rejected.” Or, to put this in plain English: additional extremely hot days could mean zero additional heat deaths. (Bold added.)

Note in the above that this result is quite conservative. The actual adoption of air conditioning in the United States in the year 2100 will surely be much more extensive than it was in 2004.

Yet even though the literature contains this well-documented ability of humans to adapt to hot days, Cass shows that some of the “official” estimates of the dangers of climate change chose to ignore this element of the research when constructing the headline figures that captured the attention of the press and policymakers. Nobody wants to hear, “With more air conditioning, it’s possible that no additional deaths will occur in the year 2100 because of 43 additional extreme heat days.”


In this post I summarized yet another example from Cass’s House testimony in which the alarming projections of U.S. deaths from climate change have clearly been exaggerated. The mortality rate from extreme heat fell in half from the period 1960-1979 to 1980-2004. Even if we stipulate the increased global temperatures in the computer simulations cited by government agencies, the estimates that take into account recent evidence on air conditioning show that the impact on human mortality in the U.S. could be negligible at the end of the century.

It is surely not fanciful or unscientific to say that estimates of climate change damage should look at the impact of air conditioning on historical mortality rates. And yet, as Oren Cass’s testimony shows, the most alarming projections of damage—and these figures are used in “official” documents guiding policy—ignore such basic considerations.

Print Friendly, PDF & Email