It’s a well-worn story that it’s much more expensive to live in big cities than smaller ones—in say, Brooklyn, New York, than in Brooklyn, Iowa. Housing prices are a huge part of this, obviously: The average property listing price in the New York City borough last week was just under $1.25 million; in the Iowa city, it was $166,000.
But while most people think just about everything else costs more in New York, a recent study by Jessie Handbury of the University of Pennsylvania and David E. Weinstein of Columbia University suggests that groceries and other staples may actually be cheaper in big cities than in small ones—when one does the proper sort of comparison.
Conventional comparisons of food prices are based on metrics like the Consumer Price Index (CPI), a Bureau of Labor Statics metric that tracks the price of a representative basket of identical goods and services sold in the same stores across metros on a month-to-month basis. Regional price indexes, another comparison method, look at similar goods purchased in different stores across many metros. But as Handbury and Weinstein point out, theses metrics suffer from two significant biases.
The first type of bias, “heterogeneity bias,” occurs when comparing what seem to be similar but are really different products. This problem arises in when one compares the prices of, for example, a half-gallon of milk in different cities, but doesn’t control for features such as brand (store brands are generally less expensive) or organic production (organic milk is generally more expensive). As Handbury and Weinstein note, even though such features often dramatically affect prices, they’re not always accounted for in standard food price indexes.
The second type, “variety bias,” arises from a failure to account for the fact that some goods are sometimes unavailable in the locations studied. There are some cities, for example, where specific fruits are available year-round, and others where they are only available on a seasonal basis. An index with such a problem built in creates a situation worse than comparing apples to oranges, the researchers point out: These sorts of indexes compare apples to nothing.
So what happens to the comparative cost of living in different places once such discrepancies are taken into account?
To get at this, the researchers employ a dataset from Nielsen collected in 2005 on the prices of nearly 350,000 individualized products scanned by barcode and sold in comparable stores in 49 U.S. cities. To ensure precise comparisons, Handbury and Weinstein only compare the prices of identical goods, such as the same brand of 2 percent milk.
First, the study finds that smaller cities offer a much more limited assortment of goods than larger ones. Handbury and Weinstein calculate that there are, for example, four times the number of products for purchase in New York (2005 population: 8.2 million) than there are in Des Moines, Iowa (2005 population: 194,000). The chart below, from the study, shows the relationship: The larger the city’s population, the more products. (“UPC” stands for Universal Product Code, or the unique barcode carried by each product and scanned during purchase.)
Overall, the study finds that a doubling in a city size leads to a 20 percent increase in the number of available products.
Second, when the researchers compare the prices of identical goods across cities, they find that the heterogeneity bias accounts for a whopping 97 percent of price discrepancies.
Once they adjust for both variety and availability, Handbury and Weisntein find conventional wisdom flipped: People in bigger cities actually pay less, in real dollars, for their groceries. Someone moving from Des Moines to the Big Apple would see a 1.5 percent increase in prices for commonly available groceries. But this higher common goods price index would be offset by the greater number of goods available in New York. As the researchers write, “The gains due to variety mean that the theoretically rigorous exact price index indicates that the price level in New York is actually 4.2 percent lower than Des Moines.”
As I discovered in my own analysis a year ago, the large gap in living costs between these two different sorts of places is mostly due to diverging housing prices. If you’re planning on moving from Iowa to New York, you’ll pay a whole lot more for housing, but you’ll save a bit on groceries and other day-to-day purchases. You’ll also have a whole lot more to choose from, which is just one of the reasons you’re paying more for housing.