Income inequality has been growing in America, driven by technology, globalization and other factors. It's caused tensions between the haves and have-nots, which often get played out at the local level, and these tensions have erupted into fights over housing affordability and public services.
Are growing income gaps limited to particular metros, or is this trend widespread? To untangle the facts about local income inequality, we compared the incomes of rich, median, and poor households in the 100 largest metros in 2012, 2006, 2000 and 1990, using Census data (see note below). A rich household is defined as being at the 90 percentile - which means being above 90% of all households in the metro; the median is at the 50 percentile, while poor is defined as at the 10 percentile. Our main inequality measure is the ratio of incomes at the 90 and 10 percentiles (the "90/10 ratio"), which shows the size of the gap between the rich and the poor. A higher value of the ratio means incomes are more unequal; among the 100 metros, the 90/10 ratio ranges from below 9 to above 18.
Taking this approach, we found that some metros are much more unequal than others, and the most unequal metros tend to have higher housing costs and slower economic growth. Despite these differences, income inequality has increased in nearly all metros over the past two decades and has accelerated in recent years.
Income Gap Widest in Fairfield County, San Francisco, and New York
The most unequal metro in America isn't a well-known big city; it isn't even bankrupt or overrun with rich tech workers. It's Fairfield County, CT, home to the tony towns of Darien and Weston but also to the city of Bridgeport, where one third of children are below the official poverty level today and which tried to go bankrupt back in 1991. There, the 90 percentile of income is 18.5 times the 10 percentile. San Francisco, New York, Boston, and Detroit - which did successfully go bankrupt last year - round out the top five. Among the top 10 most unequal metros, four are in New England.
At the other extreme, the least unequal metros in America include three in Florida that are popular with retirees: Lakeland-Winter Haven; Cape Coral-Fort Myers; and Palm Bay-Melbourne-Titusville. But equality isn't just for places with lots of older folk: Salt Lake City and Raleigh are also among the least unequal metros despite having relatively young populations.
Why Are Some Metros More Unequal Than Others?
The 90/10 ratio is twice as high in the most unequal metros when compared with the least unequal metro. What explains these huge differences in income inequality? It's hard to disentangle all of the factors that might cause or be affected by inequality, but some simple correlations can help. After examining a wide range of demographic, economic, and housing measures, two factors stood out as being most strongly correlated with local income inequality: (1) housing affordability and (2) longer-term growth.
First, let's take a look at housing affordability. Affordability and equality typically go together, and the least affordable housing markets tend to be the most unequal. The correlation between the 90/10 income ratio and the percent of for-sale homes affordable to the middle class (defined as households with median income) - is -0.4 and statistically significant. As the scatterplot shows, the four most unequal metros (Fairfield County, San Francisco, New York, and Boston) are also among the most unaffordable. Yet there are many exceptions to this pattern: Detroit, Toledo, New Orleans, and Philadelphia are also quite unequal but are relatively affordable, and on the flip side, Ventura County, Orange County, and Honolulu are more equal but far less affordable.
Why are unaffordable housing markets more unequal? Here's a potential explanation: expensive housing pushes out many low- and middle-income households, and, in turn, richer residents bid up home prices and rents. At the same time, to compensate for high housing costs, many expensive big cities have policies or programs - like rent control or inclusionary zoning - that preserve or build some housing for lower-income residents (though, ironically, such policies could reduce overall housing affordability if they discourage new housing supply).
The second factor is growth. Faster growing metros tend to be less unequal. Whether growth is measured using employment change or construction activity, the correlation between growth since 1990 and inequality today (using the 90/10 ratio) is between -0.4 and -0.5. The link between growth and income equality is especially strong for low-income households; in other words, slower job growth (as well as less construction) is associated more with the poor falling farther behind, rather than with the rich pulling farther ahead.
Higher inequality, therefore, is associated with both lower affordability and slower growth. (A more sophisticated analysis would be needed to identify strategies for reducing inequality and to determine whether improving housing affordability through more housing construction could, by itself, reduce local income inequality.) While the reasons for local income inequality aren't crystal clear, the trend toward bigger income gaps is clear - as the final section shows.
Everywhere, the Rich are Pulling Away from the Pack
Income inequality has grown in nearly all of the 100 largest metros. Between 1990 and 2012, the 90/10 ratio increased in 94 of the 100 largest metros - above all in San Francisco, Fairfield County, and San Jose. The 10 metros where inequality increased most include four in California and five in New England, as well as Honolulu.
Just six metros bucked the trend of widening inequality, all in the south-central region of the U.S:
Not only has inequality increased over the past two decades, but it has also accelerated in the last few years. Averaging over the 100 metros, the 90/10 ratio increased by the same amount in the most recent six years alone (from 11.1 in 2006 to 12.2 in 2012) as it did in the previous sixteen years (from 10.0 in 1990 to 11.1 in 2006). Furthermore, between 2006 and 2012, inequality increased in 84 of the 100 largest metros, so the growth in inequality was more widespread in those years than in 1990-2000 and 2000-2006, when fewer metros saw inequality widen.
More strikingly, the growth in inequality at the metro level since 1990 has been more about the rich pulling away from the pack than the poor falling farther behind. To demonstrate, inequality can be measured by comparing rich households' incomes to the median using the ratio of the 90 to 50 percentiles and separately comparing poor households' incomes to the median using the ratio of the 50 and 10 percentiles. Here's what these measures show: between 1990 and 2012, the 90/50 ratio increased in all 100 metros, while the 50/10 ratio increased in 68 metros and decreased in 32. That means that every large metro has seen the rich get farther ahead of the typical household, but only two thirds of metros have seen the poor fall farther behind.
The tensions over growing local inequality, therefore, are backed by the facts. Nearly every major metro has seen the income gap between the rich and the poor grow since 1990, and the gap has widened faster in the last six years. The growth in inequality is more about the rich getting richer than the poor getting poorer, though both are happening. And where housing is less affordable, income inequality is more extreme.
Note: Inequality is measured as the ratio of different percentiles of household income at the metro level. All income data come from the Census: the 1990 and 2000 decennial Censuses, and the 2006 and 2012 1-year American Community Surveys. Metros were identified based on Public Use Microdata Areas (PUMA's) mapped to consistent 2009 definitions for metro areas and divisions. Census data were obtained through IPUMS, which requests to be cited as such: Steven Ruggles, J. Trent Alexander, Katie Genadek, Ronald Goeken, Matthew B. Schroeder, and Matthew Sobek. Integrated Public Use Microdata Series: Version 5.0 [Machine-readable database]. Minneapolis: University of Minnesota, 2010.