The loss of economic momentum in several sectors during winter is a regular phenomenon and statistical techniques are employed to adjust economic data for this loss in momentum (and also for the gain in momentum that follows during spring). However, if the loss in economic momentum is more pronounced than usual due to a stronger-than-usual winter, then the seasonally adjusted economic data should turn out weaker than expected. This is confirmed by a study on the impact of the mild winter 2001-02 on the US economy (Lessons from the unusual impacts of an abnormal winter in the USA. Changnon & Changnon. Meteorological Applications 12, 187-191 2005. The abstract can be found here). More specifically, the study found that the mild winter reduced costs of heating, led to reductions in transportation problems, lower road/highway maintenance costs, increased construction activity, reduced insurance losses, greater retails sales and increased home buying.
On the other side, if we have a cold winter and especially if there are severe snowstorms such as happened in the US and continental Europe since the start of the year, then one should expect retails sales to be weaker than usual (people will stay at home more often), higher insurance losses, less home buying, weaker construction activity, more frequent transportation problems (which also means that people will get to work later, hurting output). On a brighter note, utility output should have profited. The impact of these effects should not be underestimated. I already presented at the start of the year a small model which enters a receiver or payer position in the 10y USD swap rate depending on whether a certain month was colder or warmer than usual. The model enters a receiver position if a winter month is colder-than-usual or a summer month warmer-than-usual (and a payer position if it is the opposite). The position is entered at the start of the following month and held for one month. I used the population weighted heating degree days and cooling degree days vs. the norm for that particular month to determine the positioning (the actual HDD/CDD as well as the normal HDD/CDD can be found here on the website of the CFTC. Additionally, the website provides a description of the methodology used to calculate the degree days). Over the past 11 years this would have resulted in a profit of 1130bp (excluding transaction costs) with gains in both bearish as well as bullish market environments.
Profitable trading rule suggests that the weather has indeed a significant impact on short-term market behaviour
This profitable simple model should highlight that indeed the weather has an impact on the economy and therefore financial markets which should not be underestimated. Winter in the US and continental Europe has been colder-than-usual with more snowstorms since early December. Also January has been colder (albeit not that much) as have the first weeks in February. In turn, near-term economic data is likely to continue painting a weak growh environment. However, most of this weather related economic loss should not be a permament one. It is much more likely that once the weather improves, there will not only be the usual seasonal strength of spring but rather also some catch-up in the form of improved retails sales etc. Again, this is confirming my neutral strategic outlook for government bond markets (with a horizon of approx. 3 month). However, my bearish tactical call seems to be a bit premature.