Early February brought freezing weather, and equity markets began to shiver. The connection is not entirely fatuous: Research now suggests that the weather really does have some bearing on market returns and that it could even form the basis of an index-trading strategy.
The existing literature on the relationship between the weather and stocks has been largely inconclusive, but this new research – by Ming Dong of York University’s Schulich School of Business in Toronto and Andreanne Tremblay of Quebec City’s Laval University – sought to remedy this problem on two fronts.
First, Dong and Tremblay took account of how the same weather conditions can have different psychological effects in different regions. Investors in colder climates, for example, are more likely to appreciate comparatively warmer weather than those in a desert environment. Second, they considered a combination of five weather variables – sunshine, wind, rain, snow and temperature – where others have previously focused on temperature or cloud cover.
Dong and Tremblay looked at those conditions in relation to the daily performance of 49 countries’ main equity indices between 1973 and 2012. The countries were categorized as having cold, mild or hot climates based on the average annual weather in the city of their primary national stock exchange. The US was therefore placed in the cold basket, given New York’s tendency toward severe winters.
The academics had two hypotheses: that ‘comfortable weather,’ relative to the region’s climate, would lead to better investor mood and higher stock returns, and that the effects of the weather on returns would be stronger when people spend more time outdoors or when the marginal utility of outdoor time was higher.
Broadly, these hypotheses were validated. Sunshine was found to have a positive effect on daily stock returns in each temperature category, with the strength of this effect correlated with the amount of time spent outdoors.
In cold climates, wind and rain had a negative effect on returns in both the summer and the spring, which Dong and Tremblay attributed to a detrimental impact on mood and outdoor activity in those months. On the other hand, wind and rain had a positive effect during summertime in hot countries.
Meanwhile, temperature did not interact with equity performance in any straightforward way. In cold regions, returns were negatively influenced by high summertime temperatures, while in hot countries returns were higher on cool days in the summer and on warm days in the winter.
Both tendencies can be put down to the sluggishness caused by oppressive heat, but a rise in stock returns when winter weather is harshest in cold and mild climates was more surprising. ‘This is incompatible with the comfortable weather hypothesis, but it is consistent with evidence from experimental psychology that at very low temperatures subjects tend to exhibit increased aggression or risk-seeking behaviors,’ Dong and Tremblay wrote.
Time to shine
The next logical step was to enquire whether these observations could be implemented as an investment strategy, rotating between country indices on the basis of their daily weather. Dong and Tremblay looked at the weather at the beginning of each day of their review period and went long the country index with the conditions most associated with the highest local returns, according to their data.
Taking the whole universe first, this approach yielded an average daily return of 6.1 basis points, far in excess of the 1.6 basis points delivered per day by global equities over this period. While those figures are gross of transaction costs, Dong and Tremblay noted that the alpha was sufficient to cover even conservative estimates of the fees for trading index products.
Of course, the greater practical challenge would be overcoming time differences. It is impossible to wake up in New York, buy an Australia ETF because Sydney had enjoyed a perfect morning and reap the rewards of that day’s feelgood Australian rally. Dong and Tremblay therefore also restricted the tradable markets by time zone. When they did so just for North America, the daily average return was 3.7 basis points, still ahead of the global benchmark.
Such patterns should not occur in a theoretically efficient market, reinforcing the importance of understanding investor behavior. ‘Our results do indicate that there exist substantial weather effects on stock returns, that the patterns of the effects differ across climates and seasons and that the strength of the weather effects tends to vary with the time spent outdoors,’ Dong and Tremblay concluded.
‘These findings suggest that the weather influences asset prices in ways consistent with projection bias, whereby temporary emotional states influence individuals’ judgments about long-term prospects.’