Richard Florida
by Richard Florida
Tue Aug 3rd 2010 at 12:00pm UTC

The Brainpower Map

The most powerful factor in economic growth is human capital – the level and concentration of skilled, energetic, and productive people. Charting the human capital levels of nations and regions is all the rage today, but Adam Smith long ago argued that human capital constitutes a critical fourth factor of production alongside land, labor, and capital.  Human capital is the key factor in both national and regional growth, according to careful empirical studies.

Most regional economic analyses measure human capital across metropolitan regions. But metro regions – which are made up of central cities and their surrounding suburbs – come in all kinds of shapes and sizes. Some have densely concentrated central cities, like Manhattan in the greater New York region; others are more continuously sprawling like Los Angeles. A while back, a reporter called to ask me about the role of human capital in Milwaukee. He wanted to know whether it made any difference that human capital levels were low in the central city compared to the region as a whole. I responded that I thought it did. Following Jane Jacobs, my hunch was that regions with more concentrated human capital would have some advantage over others where the distribution was flatter or more similar across the metro. When his article came out, a number of prominent economists more or less said the same thing. But we all added that we couldn’t be sure; most studies focus on human capital at the metro level and few, if any, have looked at the distribution of human capital within metros – that is, between the urban center and its suburbs.

So what about the distribution of human capital within metropolitan regions: Does it matter? With the help of my statistically savvy colleague Charlotta Mellander and our team at the Martin Prosperity Institute (MPI), I decided to take a look. We developed a simple way to begin to get at this question of the distribution of human capital within metropolitan areas. Basically, we compared data on the level of human capital (the percentage of adults with a bachelor’s degree and above) in the metro as a whole to the human-capital level in its primary urban center or central city. Fortunately, a nice data series on both metro and city-level human capital is readily available from the Brookings Institution’s State of Metropolitan America Indicators series. We used these data to generate a series of maps, build a simple measure of the ratio of central-city to metro-human capital, and ultimately to generate a scattergraph which enables us to see the pattern of central-city and metro human capital across the country.

The first map (above), prepared by the MPI’s Zara Matheson, shows human capital by city. Washington, D.C. is first with a whopping 58 percent of adults holding a bachelor’s degree or above. Five additional cities have human capital levels that exceed 50 percent – Raleigh-Cary (52.4 percent), Seattle (52.1 percent), San Francisco (51.3 percent), Madison (51 percent), and Boston (50.5 percent). Rounding out the top 10 are Atlanta (46.1 percent), Charleston (45.3 percent), Minneapolis-St. Paul (43.7 percent), and Austin (43.6 percent).

On the other hand, the city with the lowest level of human capital is Detroit, where fewer than 12 percent of adults hold a college degree, almost five times less than Washington, D.C., the country’s leading city. Most of the cities with low levels of human capital are de-industrialized cities of the Frostbelt that have faced high levels of out-migration: Youngstown (12.4 percent), Cleveland (14.3 percent), Allentown (15.3 percent), Dayton (16.8 percent), and Toledo (17.4 percent), along with two California cities – Stockton (15.8 percent) and Modesto (17.6 percent) – and two New England communities – Hartford, Connecticut (16.7 percent) and Springfield, Massachusetts (17.5 percent).

The second map (above) shows the human capital levels by metro. Again, Greater Washington, D.C. tops the list (48.6 percent), followed by Bridgeport-Stamford (47 percent), San Francisco (45.8 percent), San Jose (Silicon Valley, 41.4 percent), Boston (50.5 percent), Raleigh (52.4 percent), Madison (51 percent), Minneapolis-St. Paul (39.7 percent), Austin (39.3 percent), and Denver (38.9 percent). Seven cities rank in the top 10 on both the central city and metropolitan lists: Washington, D.C., San Francisco, Boston, Raleigh, Madison, Minneapolis, and Austin all appear twice.

When we look at metros as opposed to cities, Frostbelt communities drop out, save for Youngstown (22.1 percent), and the lowest-ranked spots are metros in California – Bakersfield (14.6 percent), Stockton (15.8 percent), Modesto (15.6 percent), Riverside (19.2 percent), and Fresno (19.5 percent); Texas – McAllen (16.3 percent) and El Paso (21.3 percent); Lakeland-Winter Haven, Florida (19.2 percent); and Las Vegas (22.4 percent).

But how do central-city and metro human-capital levels compare? To get at this, we developed a simple ratio of central-city to metro-level human capital. The results of this analysis are frankly a bit perplexing (see the chart above). The metro with the highest ratio is McAllen, Texas (1.84) – a region which ranks near the bottom of the list on metro human capital. High ratios are also found in several other communities – like, for example, Lakeland-Winter Haven, Florida (1.41) and Bakersfield, California (1.39) – that also rank low in terms of metro human capital. Other regions with high ratios include Charleston (1.52), Seattle (1.37), and Atlanta (1.27), as well as Little Rock (1.47), Boise (1.32), Salt Lake City (1.33), and Baton Rouge (1.23).

On the other side of the ledger, Detroit has the lowest ratio (.41), and it is again joined by several other Rustbelt cities in the bottom 10 – Cleveland (.48), Allentown (.53), Youngstown (.56), and Dayton (.62), along with Hartford, Connecticut (.44) and Springfield, Massachusetts (.56). Philadelphia (.66) and Ogden, Utah (.62) also have low ratios, along with Bridgeport-Stamford (.65), which has the second-highest level of metro human capital (47 percent) in the country.

The reason for these mixed results is an artifact of the data. Metros with high levels of metro human capital but moderate levels of central-city human capital will have higher ratios when compared with regions with relatively low levels of metro human capital and low-core human capital.

To get a better handle on the relationship between metro human and central-city human capital, we simply plot the two against each other in the scattergraph (above). There is a positive and reasonably close association (with a correlation between of .67). What’s more interesting though is the outliers on both sides of the fitted line.

The upper right-hand quadrant shows places which have a high level of metro human capital and a high level of central-city human capital. Washington, D.C. is the top performer here, occupying the furthest upper right-hand corner. Raleigh, Seattle, Atlanta, Madison, Boston, San Francisco, Minneapolis-St.Paul, Austin, San Jose, San Diego, and, perhaps surprisingly, Albany all occupy this “win-win” quadrant.

Greater New York is just slightly above the line with a slightly greater level of central-city to metro human capital. Denver and Bridgeport-Stamford are slightly below it.

In the bottom left quadrant of the graph are places where both metro and central-city levels of human capital are low. This space includes Detroit, Youngstown, Cleveland, Allentown, Dayton, Stockton, and Ogden, Utah, among others.

The distribution of human capital across metros areas and central cities is quite uneven. Economists Christopher Berry and Ed Glaeser have documented the growing divergence of human capital across U.S. metro regions – a subject I have written about previously for The Atlantic. But this divergence is, if anything, even greater across cities and urban centers. And since human capital is the key driver of economic prosperity, this means the economic fortunes of American metros and cities are diverging and quite likely to diverge even more in the future. Economic and social inequality is increasingly overlaid with a deepening economic geography of skill and of class. That’s a very serious problem – and one that’s getting worse.

9 Responses to “The Brainpower Map”

  1. Harold Jarche Says:

    Just because it can be measured (the percentage of adults with a bachelor’s degree and above) means that is what you should measure.

    For example, this HBR article raises some interesting points:

    “education is a misleading-to-malignant proxy for economic productivity or performance. Knowledge may be power, but “knowledge from college” is neither predictor nor guarantor of success.”

  2. Brian Kelsey Says:

    I know that city-level data is scarce compared to metros, but have you looked into plotting the ratios with other indicators, such as job growth or unemployment?

  3. Howard Says:

    Harold; My opinion is that this data does have some validity for analyzing trends in employment, segregation, etc. . ., but it is correlational and I would not interpret any relationships as one thing causing another. (I do like Schrage’s rant, but I think that it’s relevance is more in calling for the reform of higher ed and mindless HR hiring practices.)
    Brian; your point is also good. One of the difficulties with large populations is the endless variable combinations that could prove important, a constant bane of economists I believe.
    What I would like to see is a dovetailing of Richard’s theories with Hagel, Brown and Davison’s conception of “pull”. I think that their ideas exist at a micro level that would allow for quasi-experimental or qualitative analysis, that is, at a level where you could dig deeper, find some better measures, and theorize some causes.

  4. Deep Says:


    Having a college degree may not be the perfect measure of talent, but it is pretty accurate. According to Jason Abel of the Fed of Western NY, every time a city raises the proportion of college grads to the rest of the city population, it raises productivity by 2.3%. Even someone who has graduated with a poly sci degree from a small state college has developed skills that are valuable to employers. College graduates can learn new skills and information. They can read data critically. They can communicate. On the job training goes much more smoothly with college graduates than non graduates.

    When I was an undergrad in the early 2000’s, I took Introduction to Computer Science. The professor asked the class why should they spend 4 years slaving away in college, taking not only comp sci, but also calculus and physics? Not to mention all those English lit classes. When they could go to a technical school, get netorking or programming certification and get a job making $50,000, in a span of 7 months. He said, the college route is better because once the things you were certified in become obselete, the company has no need for you. You will have to go back and get more training, and you will not be paid for the training. With a college degree, a company will not hesitate to spend the cash for your training. Also you have more options in the professional feild with a college degree.

  5. Michael Wells Says:

    “To get at this, we developed a simple ratio of central-city to metro-level human capital. The results of this analysis are frankly a bit perplexing (see the chart above). The metro with the highest ratio is McAllen, Texas (1.84) – a region which ranks near the bottom of the list on metro human capital. High ratios are also found in several other communities – like, for example, Lakeland-Winter Haven, Florida (1.41) and Bakersfield, California (1.39) – that also rank low in terms of metro human capital.”

    Here’s a quick guess. These cities may have a few colleges downtown with faculty living nearby, and maybe some professionals living downtown — but no high tech industry that would involve large numbers of college grad employees in the larger region. So despite the low numbers, the few educated folks are concentrated. A very different phenomenon than someplace like Seattle that has lots of college grads, but a thriving business center concentration.

  6. L. Ross Says:

    Would be interested on more of your opinions on Pittsburgh….a city you spent a considerable amount of time in yet have not commented on as much. Perhaps you have a reason. Do you believe in the relatively positive press touting the region’s achievements in technology based economic development or have…a different opinion based on your own experience? Absent 2-3 local universities, namely CMU, PSU and Pitt…what happens to the brain power index when you exclude them, and in particular the foreign students that attend? Nearly 80% of CMU grads leave the region, and the school itself often attempts to market itself without invoking the fact that its based in Pittsburgh.

    Do you believe it is truly a viable city to grow high tech companies (beyond the healthcare/medical area…which has also had limited success)?

    Most importantly, what did you notice from first hand experience are the critical weaknesses of the city and can it really develop into a place that will attract people to come, live and work in the city (not just returnees…but those in the creative class looking for opportunity)?..

    Could you note any significant “cultural barriers” to modernizing and attracting, nurturing and retaining the creative class of people you speak of?

  7. A. Butcher Says:

    I think it would be interesting to layer the amount of entrepreneurial ventures started, early stage investment, and new small businesses (or social enterprises) started to evaluate what type of correlation there is, (if any) with human capital density.

    Given Echoing Green’s ( theory that an ecosystem of support needs to be present to cultivate successful social entrepreneurs is correct, than one would presume that the lower the density of human capital the lower the amount of early stage investment, small business creation, and entrepreneurial success.

    Even more so, it would be interesting to evaluate, how entrepreneurs are successful in places with low human capital density / ecosystem supports.

  8. L Donovan Says:

    While “Greater New York” traditionally included the 5 Boroughs of NYC and the adjoining Northern/NJ/CT suburbs, there is a sprawl into the counties north, west and east of that area. Without data for those areas, we are not tracking the movement of intellectual capital from the central city to its marginal areas.

  9. David Says:

    College graduation rates are a very good proxy for economic viability. Anytime this particular issue is brought up, however, people always cite exceptions as the rule. In GENERAL, the higher the level of educational attainment for a region, the more economically viable it is. It must be looked at from a macro level. Many of us know HS graduates who are very successful and college grads who are not. That does not contradict the general rule. Places like Seattle, Minneapolis and Washington DC all have much higher than average college graduation rates and have good economies. Some of the former rust belt areas have very low college graduation rates and are really struggling. A coincidence? Not likely. I should add that it is not just the ‘education’ part of being a college graduate that makes a difference: it’s the social network, the drive and ambition, productive living – all combined to create a human capital mosaic, if you will.