Weather that Moves Markets

At a time when the climate crisis imposes its agenda on both governments and businesses, finance is also beginning to deal with weather unpredictability. Extreme weather events — once exceptions, now increasingly the norm — not only impact daily life or the balance sheets of insurance companies, but are also beginning to affect financial markets. In particular, the US natural gas futures market, the NYMEX Henry Hub, is showing increasing sensitivity to weather forecasts, especially when they herald extreme cold or heat waves. 

This is the topic at the heart of the study Can extreme weather forecasts lead to a risk premium? Evidence of a non-linear response in US natural gas futures, written by Stefano Caselli, Professor of Finance at Bocconi University and Dean of the SDA Bocconi School of Management, along with Manou Monteux, Maria Cristina Arcuri and Gino Gandolfi, all faculty at the University of Parma and affiliated with SDA Bocconi.

The study analyzes 30 years of data — from 1990 to 2019 — combining observed temperatures and short- and medium-term forecasts (up to two weeks), with the daily returns of natural gas futures contracts. The goal is ambitious: to understand if and how extreme weather forecasts can generate a risk premium, i.e. an additional return linked to the risk that operators take when betting on future price trends.

A premium for those who take risks with the weather

"Our research clearly shows that it is not observed temperatures that move gas prices, but extreme forecasts compared to seasonal averages," says Professor Caselli. "It's the discrepancy between what is expected and what is predicted, especially when you get out of the seasonal patterns, that triggers reactions in the markets."

The analysis is based on a complex architecture that distinguishes between "actual" and "predicted" temperatures, normalizing the data with respect to historical climatological averages. In operational terms, the authors tested a spread trading strategy, buying the contract with the closest expiry (NG1) and selling the contract with the next expiry (NG2), on days when forecasts indicate unusually cold weather (below the 10th percentile compared to normal). In winter, this configuration is associated with higher demand for gas for heating, which has a direct impact on prices in the short term.

The results are surprising: such a strategy would have produced, on average, a compound annual return (CAGR) of 12% over a 30-year period, far exceeding the return of the S&P 500 over the same period (7.6%). But even more significant is the risk/return ratio: the Sharpe ratio of the strategy linked to extreme weather forecasts is 1.3 — about three times that of the US stock index.

The extreme metric

The paper highlights an interesting dynamic: as the time horizon of the forecast increases (from one to two weeks) and as its "extremity" increases (i.e. how much it deviates from seasonal averages), the return that can be obtained from strategies based on this information also increases. It is a non-linear relationship, reflecting the inherent uncertainty of weather forecasts and the growing relevance of weather conditions for energy markets.

"Two-week forecasts are inherently more uncertain, but precisely for this reason they incorporate a higher premium," explains Caselli. "Those who bet on an incoming cold wave expose themselves to a double risk: the real weather and the one related to the accuracy of the forecast. The market recognizes and remunerates this risk."

And if it seems that all this depends on an imperfect efficiency of the market, the paper offers an answer: the observed temperatures — in themselves — do not generate significant movements in prices. It is only when forecasts are wrong compared to reality that prices adjust. This is a "Bayesian" behavior, as the authors define it, which confirms the efficiency of the market in processing the information available in advance.

Finance and climate change

The study is part of an emerging strand of climate finance, which studies how markets react to signals related to climate change. If extreme forecasts are increasingly frequent, also due to the alteration of atmospheric patterns, it is reasonable to expect that these effects will be amplified.

"Our analysis shows that climate change is not just an environmental issue, but is redefining pricing mechanisms in financial markets," concludes Caselli. "Understanding these mechanisms is crucial for proper risk management, especially for anyone working in the energy sector."