Brrr. It's cold outside - still. Winter has been rough for much of the United States, with temperatures plummeting far below normal. Here in San Francisco, where Trulia is headquartered, we're setting record highs, not lows, but we sympathize with the rest of you. In fact, our economics team hails entirely from the lake-effect snow belt of upstate New York, so we know what it's like to suffer through the cold and snow.
And while our Trulia team members have up and moved to California, leaving the brutal winters of our childhoods behind, it's apparent that many of you are also dreaming of warmer locales. We analyzed search traffic on Trulia between December 1, 2013, and January 21, 2014, to see how daily temperature fluctuations affected home-search patterns (see note below). It's clear as a bone-chilling winter morning: when the cold wind blows, home searches increase - especially for homes in warmer parts of the country.
Searching for Warmth When the Mercury Drops
Nationally, home searches increased by 2.6% overall for every 10 degrees Fahrenheit the temperature dropped. Why? In part because cold weather keeps people inside where they do more indoor activities, including searching for real estate online. But cold weather doesn't simply cause people to do more of everything to an equal degree. When temperatures plummet, searches for homes within the searcher's own metro rise 2.2%, while searches for homes outside the searcher's own metro rise 2.9%.
And the colder it gets, the better warm looks. For every 10-degree temperature drop that occurs where a house hunter resides, we see a 4.4% increase in searches for homes in warm regions, which is bigger than the increase in searches overall. While some of this searching might reflect the desire to move to a warmer place and leave winter behind permanently, the increase in searches for homes in sunny vacation spots is even higher: a 5.5% jump for every 10-degree temperature decline. In other words, searches for homes in warm vacation destinations increase more than twice as much as home searches overall.
Honey, I'm Freezing. Let's Move to Miami.
Breaking these search patterns down further, we can see which metro's homes get the biggest search boost when winter weather strikes. Miami benefits most, with a 7.3% climb in searches for every 10-degree temperature drop in wintry regions, followed by Phoenix and Jacksonville. And among the 10 metros that get the biggest rise in searches when temperatures plunge, only one - Dayton, OH - is outside the South and West.
On Second Thought, I Do Love the Summer at Home. Let's Be Snowbirds.
The table above showed that, in cold weather, searches to warm vacation areas jumped even more than searches to warm places in general. But you don't need to move to Florida to avoid winter - you just need a second home there. For our analysis of vacation areas, we identified ZIP codes where vacation homes account for at least 25% of the housing stock, according to the Census, and grouped them together within each county to define vacation areas. (This is the same method as in this post.)
Vacation areas where search traffic jumped most include several beach towns in Florida. But when winter hits hard, the number one vacation spot where searches increased was Deschutes County, OR, near Bend; searches for homes there jumped 9.1% for every 10-degree drop in temperature elsewhere.
As January comes to an end, temperatures are once again dropping in much of the country. Try to stay warm: bundle up, sip hot chocolate, and look at homes in Miami and Phoenix.
Note: this post is based on web searches on Trulia between December 1, 2013, and January 21, 2014. Data on daily high temperatures come from the National Climatic Data Center. We excluded searches by people in regions with warm winter climates, which include much of the South, Southwest, and California. This is based on regression analysis of daily searches by people in a metro to homes in regions, metros, or counties, and the explanatory variable is daily high temperature in the searcher's metro; we included fixed effects for the searcher's metro and for the day of week, as well as a quadratic time trend. All of the reported percentages are statistically significant at the 5% level.