VoxEU Column Energy Environment Financial Markets

The market is betting on climate change

Understanding beliefs about climate change is important, but most of the measures used in the literature are unreliable. Instead, this column uses prices of financial products whose payouts are tied to future weather outcomes in the US. These market expectations correlate well with climate model outputs between 2002 and 2018 and observed weather data across eight US cities, and show significant warming trends. When money is at stake, agents are accurately anticipating warming trends in line with the scientific consensus of climate models.

Scientists overwhelmingly agree that climate is changing because of human activity. But public opinion in the US remains mixed. As of 2016, less than half of Americans believe that the Earth is getting warmer due to human activity, a number which hasn’t budged much since the Pew Research Center started asking the question back in 2006. These views vary greatly across geography, political affiliation, educational status, and economic sector (Leiserowitz et al. 2016). Likewise, politicians in the US have questioned the evidence on climate change, with some famously calling it an ‘elaborate hoax’. However, public statements and poll responses on highly politicised issues like climate change can be unreliable.

In our recent paper (Schlenker and Taylor 2019),  we look at weather derivatives in financial markets to assess beliefs about climate. We find that traders have been pricing in a warming trend that is closely aligned with the projections of scientific climate models, as well as the observed weather outcomes during that time. 

Our data come from the Chicago Mercantile Exchange, which offers monthly futures contracts for eight cities on two main weather products – cooling degree days, which measure how much cooling is necessary during hot temperatures in summer months, and heating degree days, which measure how much heating is required during cold temperatures in winter months. The contracts are indexed to 65°F, a common standard for utility companies because cooling and heating systems tend to be turned on above and below that level, respectively. For example, a mean daily temperature of 85°F degrees would count as 20 cooling degree days. These daily degree days are then summed over the course of a month or season. 

Introduced in 2002, the futures contracts allow companies to mitigate losses stemming from fluctuations in the weather (e.g., a heating oil distributor can hedge against low sales stemming from a warm winter). They also allow traders to bet on whether a particular month will be hotter or colder than average. And since they are traded prior to the month on which their price is based, the contract price provides a direct measure of the market’s view on future climate beyond the ten to fourteen-day window in which weather forecasts have predictive power. At the beginning of June, for example, traders begin placing bets on how hot July will be, an outcome unknown at the time the trades are made. We test whether this price has changed over the years.

For the eight cities from 2002 to 2018, we sum up the yearly predictions for cooling degree days between July and September, and for heating degree days between November and March. We then compare annual price trends, actual temperature trends, and the projections made in 2006 by an ensemble of climate models. All show significant warming, i.e., an increase in cooling degree days in summer and a decrease in heating degree days in winter. In other words, the market’s predictions are aligned with observed warming trends. The relationship is weaker during winters than summers, but still robust. It holds after controlling for changes in ocean circulation patterns such as El Nino and the North Atlantic Oscillation, which have been shown to be strong predictors of temperature (Zebiak and Cane 1987). 

Figure 1 Trends in futures prices, weather station data, and climate models

Note: The figure estimates non-parametric trends using lowess regression on the average annual residual among the eight city airports. The left panel shows cumulative degree days for summer (June-September) and the right panel for winter (November-March).

Futures prices closely follow the predictions of climate scientists, which, on average, appear to have materialised, thus validating the climate models. This close agreement between markets and models implies that traders are taking into consideration the scientific consensus on climate change when making trades. Overall, we find that the market has been accurately pricing in climate change, largely in line with global climate models, and that this began occurring at least since the early 2000s when the weather futures markets were formed.

Our findings are relevant to understanding climate adaption. Economists have studied the costs and benefits of a changing climate (Auffhammer 2018), often by estimating damages by looking at the impacts or random year-to-year weather fluctuations (Dell et al. 2014). But if people are adapting to what they see as a permanent change in climate – and shifting their behaviours and investments accordingly – then it may not be appropriate to assume these observed weather sensitivities over the long term. And before people can adapt, they first need to believe that climate is changing. Therefore, knowing true beliefs about climate change is crucial for both successful policymaking and adaptation.

Anyone doubting the observed warming trend can make a significant profit by betting against it in weather markets. However, the observed annual trend in futures prices shows that the supposedly efficient financial markets agree that climate is warming. When money is on the line, it is hard to find parties willing to bet against the scientific consensus.

References

Auffhammer, M (2018), “Quantifying economic damages from climate change”, Journal of Economic Perspectives 32(4): 33-52.

Dell, M, B F Jones and B A Olken (2014), “What do we learn from the weather? The new climate-economy literature”, Journal of Economic Literature 53(3): 740-798.

Leiserowitz, A, E Maibach, C Roser-Renouf, S Rosenthal and M Cutler (2016), “Climate change in the American mind: November 2016”, Yale Program on Climate Change communication.

Schlenker, W and C A Taylor (2019), “Market expectations about climate change”, NBER working Paper 25554. 

Zebiak, S E and M A Cane (1987), “A model El Niño-Southern oscillation”, Monthly Weather Review 115: 2262-2278.

6,824 Reads