School of Public Policy Study: No Agricultural ‘Threshold Effect’

Posted April 20, 2020

By Michael Pearson

A new study led by a Georgia Institute of Technology School of Public Policy researcher offers confirmation that global warming does not appear likely to cause a sudden collapse in U.S. agricultural productivity.

In the study, Emanuele Massetti, an assistant professor at Georgia Tech, and co-author Robert Mendelsohn of Yale University sought to assess the validity of models that have tried, but failed, to identify a temperature threshold beyond which crops would suffer a sudden and precipitous decline as a result of climate change.

Their study used 30 years of daily temperature data in a model that allowed Massetti and Mendelsohn to evaluate the impact of numerous temperature ranges on productivity in a more flexible way than models used by some other researchers.

They found that while warming certainly will result in significant losses in productivity, as measured by land values, the decline will be gradual.

“In the global warming literature, there is significant debate on whether climate change will have non-linear impacts on the economy,” Massetti said. “This study shows that for agriculture, at least, it will not.”

Looking for a Threshold Effect in Agriculture

Massetti and Mendelsohn’s paper has to do with the “threshold effect,” a sudden and dramatic change resulting from conditions edging even slightly beyond a specific quantitative limit, such as temperatures shifting by one degree. Water ice, for instance, exhibits a strong threshold effect: at temperatures below the freezing point of water, it remains sold. Yet at even one degree above it, ice immediately starts to melt.

Climate change researchers have feared such an effect on agriculture, but the results of econometric studies have been anything but clear.

Several studies of the impact of warming on individual crops – such as corn, soybeans or cotton – have found a threshold effect tied to temperatures ranging from 84 degrees to 93 degrees Fahrenheit, above which the studies predict crop yields will quickly drop. Studies of the effect of temperature on total agricultural productivity have instead found a hill-shaped relationship where productivity peaks near the mean growing season temperature, then falls gradually as the climate warms.

Massetti and Mendelsohn wanted to know if those latter models fail to predict a threshold effect because of the restrictive way in which they are built, relying on quadratic specifications of mean seasonal temperatures. The model Massetti and Mendelsohn used was less restrictive. It grouped temperatures into “bins,” each one 3 degrees Celsius wide, and evaluated how the number of days in different temperature bins affected land values — the study’s proxy for productivity.

They found no evidence that even very hot conditions have a suddenly harmful effect on farmland values.

Assessing the Damage from Climate Change and the Role of Adaptation

Massetti and Mendelsohn do not think their result contradicts the studies that find threshold effects for individual crops. By considering aggregate agricultural productivity their model reflects the ability of farmers to adjust their practices to avoid the temperature thresholds. Farmers in hotter areas typically use crops that are better suited for higher temperatures. This is why we see cotton mostly grown in the south and wheat in the north of the United States. Farmers in very warm areas grow fruits or raise animals, much more heat tolerant than field crops.

Hundreds of different adaptations allow farmers to avoid dangerous thresholds. The same adaptations will help to reduce the harmful impact of future climate change. However, even accounting for the mitigating effect of adaptation, the damage to all farms of warming is roughly 8% per degree Celsius, according to the study. All-crop farms fare the worst, losing approximately 15% of their value per degree of warming Celsius, while mixed farms — those that include livestock — decline in value by about 4% per degree.

“If these results are confirmed and apply to other sectors, policy makers may not need to worry about sudden collapses of economic productivity, but this does not mean that much more aggressive mitigation of climate change is not needed,” Massetti said. ”The economic cost of climate change may be large, even after accounting for adaptation.”

The paper, “Temperature Thresholds and the Effect of Warming on American Farmland Value,” was published on March 27, 2020, in the journal Climatic Change.

The School of Public Policy is a unit of the Ivan Allen College of Liberal Arts.