What lies behind the inequality of American cities? The conventional explanation blames the rise of the globalized, knowledge economy which has eliminated family-supporting factory jobs and cleaved the workforce into high-paying, high-skill and low-paying, low-skill jobs. But, as I wrote in my previous post, wage inequality only explains a very small part of income inequality.

How to explain this apparent discrepancy? What other factors lie behind rising inequality across America's cities?

To answer that question, I reviewed several powerful theories that try to explain persistent economic and social disadvantage across cities.

The first focuses not just on trends in skills and wages, but on shifts in populations. Christopher Berry and Edward Glaeser noted the divergence of human capital levels across cities in 2005. In his book, *The** Big Sort*, Bill Bishop shows how America is becoming increasingly sorted and divided by skill, economic position and political differences. Writing in the magazine, I dubbed this the "means migration."

A detailed study published by the National Bureau of Economic Research found inequality to be higher in larger cities and metros (which attract more highly skilled people). Size (measured as population) alone accounted for roughly 25 to 35 percent of the total increase in economic inequality across metros over the past three decades, after other key factors were taken into account.

A second calls attention to declining rates of unionization. In *The Great U-Turn**,* economists Bennett Harrison and Barry Bluestone blame the attack on and the decline in unions for undermining wages not just for unionized workers but for reducing the so-called wage floor for workers in the broader economy.

A third focuses on the intersection of race, poverty and economic disadvantage In his classic, *The Truly Disadvantaged,* sociologist, William Julius Wilson identified race and poverty as persistent factors. Robert Sampson's *The Great American City* shows the enduring legacy of location and neighborhood effects in perpetuating economic disadvantage.

I decided to look at how each of these competing theories might help to explain the inequality puzzle across America's cities.

With the help of my Martin Prosperity Institute (MPI) colleague Charlotta Mellander, I examined the factors that might be associated with both wage and income inequality, probing the role of poverty, race, unionization, skills, human capital, and other variables identified by previous studies. It’s important to remember, of course, that correlation and causation are not the same thing; other factors that we didn’t consider at all might play an even larger role. Our broader research on the subject also included more detailed multivariate analysis.

The results give us much to ponder. They suggest that full story of inequality across American cities goes beyond technology, globalization, skills and wages, and includes unions, race and poverty.

W**age Inequality**

According to our analysis, wage inequality is closely correlated (.61) with human capital (measured as the percent of adults that are college grads) and even more so (.68) with the share of the work force employed in knowledge, professional, and creative occupations. It is similarly associated with the two key skill groups that define high-skill jobs – analytical skills (.44) and social skills (.55). Wage inequality is also closely related to the technology-intensity of metros, with a whopping .74 correlation to high-tech industry concentration.

Wage inequality is also associated with the wealth and affluence of cities more generally. It is closely related to economic output per capita (.48), average incomes (.46), and average wages (.56). It is also related to both the density (.38) and even more with the population size (.48) of metros.

Interestingly, wage inequality across cities is only modestly related to housing costs as a share of income (.19).

**Income Inequality**

The picture changes substantially when we turn to income inequality. Wage inequality explains *just 15 percent* of the variation in income inequality What accounts for the rest?

First off, most of the factors that are associated with wage inequality have little relation to income inequality. There's little correlation between income inequality and the wealth of cities. There is no correlation between income inequality and average incomes or wages, and only a weak relationship to economic output per capita (the correlation is less than .2).

Income inequality across cities is far less associated with factors like skills and human capital. The correlation between human capital and income inequality is less than .3, less than half of that between it and wage inequality. There is even less of a correlation to the creative class (the correlation is less .2). Income inequality is also not related to city size (with a correlation of .2), and has no correlation whatsoever to density. And income inequality has no relation at all to the share of income devoted to housing costs.

So what is related to income inequality?

Unionization for one. The correlation is -.3, about double that for wage inequality. Unions appear to play some role in mitigating income inequality across U.S. metros.

Income inequality is even more closely linked to race and poverty. It is significantly associated with the share of the population that is black, with a correlation of .3 (in contrast, there is almost no correlation between race and wage inequality).

The number one factor in income inequality, by far, is poverty. Income inequality is quite considerably associated with the overall level of poverty, with a correlation of .5, one of the highest of our analysis (there is no correlation between poverty and wage inequality by the way).

The graph above shows the association between income inequality and the share of population below the poverty line. While the regression line is strong, there is also considerable variation. Some metros have high income inequality alongside high rates of poverty (Gainesville, Florida; McAllen, Texas; Athens, Georgia), while others have high inequality but much lower rates of poverty (Bridgeport, Connecticut; Naples, Florida; New York).

To further confirm this, we also ran a multivariate analysis of the factors that explain income inequality, including poverty, race, unionization alongside human capital and skills in our analysis. Our findings show that unionization, race and especially poverty are closely associated with income inequality, even when human capital and skills are taken into account (the full paper is here).

This connection between income inequality and poverty and race also helps to explain why such inequality is so much more pronounced in the U.S. than in other knowledge-based economies like Canada or the Northern European nations.

The inequality of American cities is bound up with not just the transformation of the United States from an industrial to a global knowledge economy, but of persistent tragedy at the bottom of America’s socio-economic order. Policy-makers can no longer put off dealing with this issue. We need to create more good jobs, and to develop strategies to turn the tens of millions of current low-wage, low-skill jobs into higher-wage jobs by more fully engaging worker sand leveraging their skills. But we must also build a more robust social safety net to address the persistent legacies of poverty and race which plague the truly disadvantaged.