If January construction and home sales data – which come out this week and next – show declines, the “polar vortex” will be prime suspect #1. But how much did January’s weather really affect housing? Last month, we showed that cold weather boosts online home searches, particularly for homes in warmer regions. On the flip side, construction and sales activity should decline in bad weather since many construction sites are unprotected from the outdoors, and sales depend on people braving the elements to go to open houses, signings, and the like. But does it?
To estimate how much January weather affected construction and sales, we looked back at more than a decade of weather and housing data. We calculated the historical relationship between abnormal weather (temperature and precipitation) and month-over-month changes in five types of housing activity (construction starts, construction permits, new home sales, existing home sales, and pending home sales) in each of the four regions (Northeast, Midwest, South, and West).
Did Housing Freeze Up Last Month?
Applying the historical patterns to last month’s actual temperature and precipitation shows that January weather probably contributed to a small decline in all five housing activities. The January month-over-month housing data coming out in the next two weeks should be 1-2% lower than it would have been if last month’s weather had been in line with January norms:
| Housing activity |
Predicted January weather effect |
January data release date |
| Construction starts (Census) |
-1.1% |
2/19 |
| Construction permits (Census) |
-1.4% |
2/19 |
| Existing home sales (NAR) |
-1.1% |
2/21 |
| New home sales (Census) |
-2.1% |
2/26 |
| Pending home sales (NAR) |
-1.3% |
2/28 |
Relative to the amount of volatility in month-over-month housing numbers, 1-2% is a pretty small effect. Why didn’t the polar vortex matter more for housing activity? Three reasons:
1. Last month was cold, but it wasn’t the coldest January of the Millennium. The average daily high in the Midwest, for instance, was 27 degrees in January 2014 – just 6 degrees below the historical norm. In fact, the Midwest had a slightly colder January in 2009, and the Northeast had colder Januaries in 2003, 2004, and 2009, too. Even though the polar vortex was brutally chilling, some days in January were downright balmy. For example, in New York, the daily high temperature was above the January average 12 out of 31 days this year. And – at the risk of rubbing it in from here in San Francisco – last month the West had its second warmest January in 15 years.
2. Precipitation affects housing activity, too. In January 2014, the Northeast and Midwest had near normal precipitation, the South was dry, and the West was very dry. Winter precipitation reduces home sales, particularly in the South, so the dry January was, if anything, a slight help to housing, offsetting the effect of the cold.
3. The polar vortex hit hardest in regions with less housing activity. The Midwest and Northeast suffered the severest weather in January, but the South and West together account for the majority of housing activity in the U.S. Combined, those two regions made up 76% of national construction starts and 64% of existing home sales in December. January’s harshest weather, therefore, was not where most of America’s housing activity is.
Winter Weather is a Wobble, not a Hobble
Here’s what the weather wobble means for interpreting the forthcoming January construction and sales data. Because the weather had a slight negative impact on housing activity, flat month-over-month numbers for construction or sales would mean that other market forces were strong enough to offset the negative effect of bad weather. But if housing activity fell month-over-month in January by more than the predicted weather effect, don’t pin the entire drop on the cold. That means if existing home sales fall by 5% month-over-month in January, for example, then only a bit of decline (1.1%) should be blamed on weather.
The regional patterns in housing activity will also help reveal whether weather mattered. The impact of January’s weather on starts should be most negative in the Northeast and Midwest, so if starts decline most in the South and West, then weather’s not the culprit.
Finally, housing activity tends to bounce partway back the month after bad weather (unless that next month is unusually bad, too). Rain and cold don’t last forever, and neither do their effects on housing. Therefore, bad winter weather will only delay some construction and sales activity, rather than make it disappear. Severe winter weather may cause housing activity to wobble, but cold, rain, and snow won’t hobble the housing recovery.
Note: this analysis is based on seasonally adjusted regional data on construction starts, construction permits, and new home sales (Census) and existing and pending home sales (NAR). Weather data come from the National Climatic Data Center. Monthly regional data were calculated from daily weather station data, using weighted averages by county households. The estimates are based on regression analysis of month-over-month changes in housing activity on contemporaneous and lagged deviations of temperature and precipitation from monthly norms. Only November-March data were used in the analysis since weather affects housing activity differently in warmer months.
