No question about it: how much money it takes to make ends meet varies dramatically across the country. It takes a whole lot more money to get by in New York, San Francisco, Boston, or D.C. than, say, Pittsburgh or even Portland.
These differences boil down to gaps in what economists call the “cost of living,” a figure that takes into account how much you need to do everything from pay rent to buy a gallon of milk, and which is a function of both national and local economic trends.
But what actually drives these big regional differences in living costs?
The Bureau of Economic Analysis (BEA) collects data on the various components of living costs for every metro in the country based on measured called Regional Price Parities (RPP). RPP tracks the different price levels of categories like food, transportation, housing, and education, as compared to the national level, which is set at 100. My colleague José Lobo of Arizona State University calculated average RPPs for 2008-2012 for four different categories: overall cost of living, cost of living for rent, cost of living for goods, and other. RPP is a weighted statistic, but Lobo was able to isolate the different effects of housing costs and the costs of goods on overall cost of living, or RPP.
The first map below charts the overall cost of living across U.S. metros.
Honolulu, Hawaii has the highest cost of living, with an RPP of 122.9. This is not surprising, as it is on an isolated island where shipping costs effectively drive up the price of everything from goods to housing.
New York (122.2) is second, followed by San Jose, the hub of Silicon Valley (122.0), and Bridgeport, Connecticut (121.5)—which includes many high-end commuter suburbs of New York, as well as the finance center around Stamford. Next are Santa Cruz, California (121.4), San Francisco (121.3), and Washington, D.C. (120.4). The map appears blue and dark blue, indicating high cost of living, along the Northeast corridor and in Miami, parts of Texas, Chicago and Minneapolis in the Midwest, and California and Seattle along the West Coast. In contrast, many Midwestern Rustbelt metros and older, smaller places in the Deep South had RPPs below 90.
But take a look at the second map, below, which isolates just the cost of living for the goods and services part of the RPP calculation. Now the differences among metros shrink considerably, and the map looks far, far less varied.
Sure, it’s still more expensive in the Northeast corridor and the West Coast, while parts of the old South and upstate New York are now relatively more expensive than they were in the overall cost of living metric. These are places where high prices of goods contribute more to cost of living disparities. But, overall, the range is far smaller.
The places with the highest RPPs for goods are just over 110, a mark-up that is substantially less than for the overall cost of living, which reached well over 120 for places like Honolulu, San Francisco, and New York. When looking at overall cost of living, less than a third of all metros fall within five percent of the national average of 100, with scores of 95 to 105. In contrast, more than 300—over 80 percent of all metros—fall within five percent of the national average for goods.
The third map shows the cost of living difference based just on housing or rents. The quick takeaway is that differences in living costs across metros seem to be driven almost entirely by the huge differences in housing costs.
The range is enormous. The prices in places with the highest overall living costs are roughly 20 percent above the national average, and the prices in places with the highest costs for just goods and services are roughly 10 percent above the national average. But when it comes to housing costs, the priciest metros are a whopping 50 to 70 percent more expensive than the national average.
The places where housing costs are super high are generally where you would expect: the East Coast corridor, from Maine through the Boston-Washington corridor, and in Southern Florida. New Orleans, Dallas, Austin, Denver, Chicago, Minneapolis-St. Paul, Salt Lake City, Phoenix, and many of the West Coast metros also have housing costs that are considerably above the national average.
The San Jose metro area, the heart of Silicon Valley, has the highest RPP for housing of any metro—170.4. Its overall RPP, however, is only 119.8, and its RPP for goods only is an even more modest 109.7. In nearby San Francisco, the overall cost of living RPP is 119.7, and its RPP for goods is just 109.9; but its RPP for rents is 167.5. The center part of the Honolulu metro—the one with the highest overall RPP in the nation—comes in second in terms of housing costs, with an RPP of 167.5.
The same pattern holds for the nation’s two largest metros. Los Angeles has an RPP of 154.6 for housing and just 103.4 for goods, giving it a total cost-of-living score of 115.2. New York has an RPP for housing of 153.9, compared to a cost of living for goods only of 107.7, for an an overall cost of living score of 121.3.
In several cases, high housing costs offset even below average costs of living for goods. In Oxnard-Thousand Oaks, California; Trenton, New Jersey; and Boston-Cambridge, Massachusetts, RPPs for goods only are slightly below 100. But housing prices between 30 and 60 percent higher than the national average contributed to an overall cost-of-living-score of 11 or 12 percent higher than average.
This means that housing costs are even more of a contributor to cost-of-living differentials than you might expect. The variation in RPP when we isolate just housing costs is a range of nearly 100—the highest is 170, while the lowest is just 57. There are about 25 metros whose RPPs for housing are over 120, while 10 have housing RPPs over 150.
• • • • •
I wanted to look a bit more systematically at what might be behind the huge gap in the housing component of cost of living across metros. My MPI colleague Charlotta Mellander ran a basic correlation analysis of the key economic and demographic factors that might be associated with this. I give my usual warning that correlation does not imply causation but merely points to associations between variables.
The housing components of cost of living track the size and density of metros, being closely associated with population size (.46) and even more so with high population density (.64).
It also tracks closely with key dimensions of the high-tech knowledge economy. The housing cost of living is positively associated with the share of creative class workers (.46), the share of college grads (.58), and concentration of high tech industry (.54). Conversely, the housing component of cost of living is negatively associated with the share of workers in blue collar jobs (-.50). Of course, this says nothing about the direction of causation. These are the key factors that contribute to higher regional productivity, and thus to higher wages and incomes and, ultimately, housing prices.
The housing component of cost of living is also associated with key markers of artistic amenities and openness to diversity, a point I have made before. The housing component of cost of living is correlated with the Bohemian Index, which tracks artists (.42) and even more so with the concentration of gay couples (.69).
Most of all, our statistical analysis suggests that, when we talk about differences in costs of living we are mainly talking about differences in housing costs. The RPP for housing is much more closely correlated to this overall RPP metric (.94) than the correlation between overall living costs for goods and services is (.78). And the correlation between the cost of living for housing and for goods is weaker still (.58).
In essence, what it all comes down to is that housing—which makes up one of the largest single expenditures for most American families—is the big driver of variation in costs of living.