Yes, it’s ridiculously hot outside. Unless you’re reading this from the northern reaches of Alaska (low of 37 degrees tonight!), chances are you’re sweating or panting (that is, if you’re a dog, in which case someone should probably alert the presses that a dog can read). Here in New York City, the sidewalk seems to be melting, and as The New York Times put
it, “the afternoon sun feels as though something hot and fierce has jumped on your shoulders and bitten your neck.” Aside from making us hotter, can weather patterns change the choices we make and the things we buy? To be fair, there are some things that are going to vary seasonally; it’s not particularly noteworthy that gelato sales are higher in July than, say, November. But can heat waves (and cold fronts) affect bigger purchase decisions as well?
Economists Meghan Busse, Devin Pope, Jaren Pope, and Jorge Silva-Risso addressed this question by looking at two of the largest purchases that people make: cars and homes. When we buy things that last for a long period of time (like a car) we necessarily need to make some predictions about whether we’ll enjoy using that thing in the future. Theory from economics would suggest that we make these predictions in a straightforward, rational manner: I should be able to pretty accurately predict just how much usefulness I’ll get out of a product both now and in the future. The reality, however, is that this isn’t the case. Psychologists and economists have found that people succumb to what’s known as the “projection bias”: we mispredict how much our future tastes and preferences will match our current tastes and preferences. We shouldn’t go to the grocery store, in other words, on an empty stomach. That jumbo pack of Reese’s may seem ideal when you’re starving, but once you’ve had your dinner, you might not want 16 peanut butter cups lying around your house (bad example, that actually sounds pretty great). The point is that we often fail to consider the wants of our future selves and consequently make decisions in the present that might not be best over time.
Although this concept makes intuitive sense in contexts such as grocery shopping, what hasn’t been examined is whether people fall prey to the projection bias when making monumental decisions like buying cars or houses. It’s not hard to think about what such evidence would look like. If consumers were perfectly rational when making large decisions, and weren’t affected by their visceral states, then we would expect them to make similar decisions for cars and houses no matter what it was like outside. If, on the other hand, projection biases were at play, we might expect that warm weather would lead people to buy a disproportionate number of convertibles and homes with swimming pools. And even though it’s hard to even imagine what cold weather is like this time of year, we might also expect more 4-wheel drives to be sold in the winter months. Busse and colleagues examined these questions by looking at 40 million car sales from 2001 to 2008 and 4 million housing purchases from 1998 to 2008.
Their data show strong evidence for the projection bias. The paper is chock-full of insightful analyses, but let me highlight just a few of the interesting ones here. Sales of convertibles – the classic summer car – peak in April for 7 of the 8 years of data. Right when it starts to get warm, in other words, we see a spike in sales of a car that seems ideal for warmer weather. And, sales of 4-wheel drives reach their peak at the beginning of winter.
But maybe this isn’t really projection bias – as unlikely as it might be, perhaps people are buying cars seasonally, like sweaters and shorts. To combat this point, the authors look at car sales based on what a city’s weather is generally like. Namely, they examined whether abnormally high temperatures, no matter what time of year, also increase sales of convertibles. Indeed, there is a strong correlation between hotter-than-average weeks in Chicago (a city that shows a lot of weather fluctuations over a year) and sales of convertibles. On average, a mean temperature that is 20 degrees warmer than usual will lead to a .22 percentage point jump in in the percentage of total cars that are convertibles. Given that convertibles account for 2.6% of all cars that are sold, this represents a significant increase. Importantly, this was only the case in the winter, spring, and fall, when a change in temperature would really be noticeable and would increase the immediate value that consumers might place on a convertible. In Miami, where it is usually warm, these correlations don’t exist.
Along similar lines, home sales also seem to be affected by weather patterns. Houses with swimming pools that go under contract in the summer months sell for an average of .2 percentage points more than expected. Homes with swimming pools that go under contract in the winter months sell for an average of .16 percentage points less. These changes represent about a $1400 swing in prices. Importantly, these analyses remain significant when controlling for other relevant factors, like number of bedrooms, the size of the lot, etc.
Yes, we might make minor decision errors when we’re tired or extremely hungry but it turns out that we may be at the whims of our visceral states when making consequential choices as well. Indeed, in Busse and colleagues’ paper, they find that cars that were purchased on abnormally nice days are more likely to be quickly traded in for a different car. The authors suggest the need for more regulation in the car and housing markets, such as extended “cool-off” periods (no pun intended). Of course, one upside to all of this is that it’s way, way too hot outside to even thinking about leaving my apartment to go buy a car or a house right now.