Richard Florida
by Richard Florida
Sat May 29th 2010 at 12:31pm UTC

The Density of Smart People

Clusters of smart people of the highly educated sort that economists refer to as “human capital” are the key engine of economic growth and development. The standard way economists measure this is to take the percentage of people in a country, state, or metropolitan area with a bachelor’s degree or higher. Jane Jacobs argued that the clustering of talented and energetic in cities is the fundamental driving force of economic development. In a classic essay, “On the Mechanics of Economic Development,” the Nobel prize-winning, University of Chicago economist Robert Lucas formalized Jacobs’ insights and argued that human capital, or what can be called Jane Jacobs externalities, are indeed the key factor in economic growth and development. Still most scholars measure human capital in terms of population, not in terms of its geographic concentration.

So I was intrigued by this fascinating analysis by Rob Pitingolo (h/t: Don Peck) which takes this question head on. To get at the issue of human capital clustering, Pitingolo compiled a neat measure of what he calls “educational attainment density.” Instead of measuring human capital or college degree holders as a function of population, he measures it as a function of land area – that is, as college degree holders per square mile.  As he explains:

I compiled the data at two geographic levels: first at the city level and second at the “urban county” level. I realize that comparing these geographies is not always entirely fair. That’s why I’m giving away the spreadsheet with all of my work to anyone who wants to build upon this analysis (download it here). I picked these cities by looking at the 50 largest metro areas by population and pulling what I deemed to be the “primary city” from each. In two metro areas, the Twin Cities and Bay Area, I pulled two “primary cities.”

He goes through a variety of analyses in his post, which I highly recommend. But let me just show the results of  his analysis of college degree density for the 50 largest cities.

San Francisco and New York are far and away the leaders in human capital density with 7,031 and 6,357 college degree holders per square mile, respectively. Boston (3,871), Washington, D.C. (3,395) , Seattle (2,853), and Chicago (2.543) all have human capital densities in the range of 2,500 to 3,500 degree holders per quarter mile. Silicon Valley has a human capital density of 1,259 degree holders per square mile. Also in this range and above the 1,000 threshold are Minneapolis (1,997), Providence (1,711), Philadelphia (1,664), Miami (1,633), L.A. (1,596), Oakland (1,596), Baltimore (1,336), St. Paul (1,293), Pittsburgh (1,289), San Jose (1,259) Portland (1,194), San Diego (1,071), Atlanta, (1,035) and Denver (1,023). Interestingly, noted, smart, high-tech clusters Austin and Raleigh are slightly below this level with 857 and 799 college degree holders per square mile, respectively. The lowest human capital densities are in Oklahoma City (159) and Jacksonville (167). Human capital densities of less than 500 degree holders per square mile are found in Birmingham (210), Louisville (250), Nashville (268), New Orleans (285), Kansas City (288), Memphis (313), Virginia Beach (370), Indianapolis (408), Detroit (425), Salt Lake City (445), Cleveland (453), San Antonio (469), and Phoenix (470). The median density in his data set is 792.

Pitingolo goes on to provide interesting analyses of human capital density at the county level, and also to conduct a residual analysis which enables him to identify places that are better or worse than expected on “predicted degree density.  He raises an important question about the distribution of human capital within a metro region, calling attention to the issue of ”human capital sprawl.” As he defines, this occurs when human capital density is lower in the central city than its surrounding county. He finds preliminary evidence of this type of  human capital sprawl in five places - Louisville, Jacksonville, Oklahoma City, Nashville, and Indianapolis – and notes that: “This preliminary result is particularly worrisome if you believe that metro areas need strong central cities and strong central cities need a lot of smart people.”

33 Responses to “The Density of Smart People”

  1. Wil Says:

    San Francisco is crawling with financial retirees who have degrees, yet produce nothing. Most innovation in the Bay Area occurs in the suburbs, such as silicon valley, Berkeley, and Emeryville.

  2. Charles Says:

    The companies might be located in Emeryville or the Silicon Valley, but many of the employees live in San Francisco. That’s why, for example, Google runs a bus service from SF to Mountain View. I’ve worked in Berkeley, Marin, Menlo Park, and Mountain View but I live in SF, because it’s a reasonable commute to any of the tech centers in the bay area.

  3. Joe Zuccaro Says:

    Did you intend the title about the “Density” of “Smart People” to be a double entendre?

    A better title, however might have been “The geographic distribution density of people who have had formal education past high school,” only because “smart” does not equal formally educated. (how many of those “smart” people work for CEOs who dropped out of college? probably more than you realize.

    But this graph means little without the original dimension ignored – the proportion of “smart” people to the relative population in the city, which can create a culture of innovation and free enterprise instead of a cloistered sub-culture.

    That makes a whole new pecking order – divide the density of college gradates by the density of the population at large for Raleigh, and you will discover it it jumps from #26 to #4. The Triangle region has a been a steady economic growth engine, perhaps not in absolute numbers but nevertheless has maintained a culture that still makes it an attractive investment area. Types of degrees may make a difference – the economic development of an area with more life sciences degrees may be characteristically different from one with more IT degrees.

    Lastly, the “smart” people in New York and Washington DC that have brought this country to the brink of disaster maybe need some common sense from the not-so-smart people in other areas. DC probably has more law degrees; NYC more finance degrees. Go figure.


  4. nap Says:

    if only college degree status was a true indicator of intelligence… :p

  5. marc Says:

    I live in NYC and am an editor at a magazine. I’m 28 and never went to college. In fact, I almost didn’t graduate high school because I cut class so much.

  6. cognominal Says:

    I am a fan of Jane Jacobs but also of Marshall McLuhan and Lewis Mumford. According to the naive prediction of MCLuhan, an Internet consequence, the materialization of his electronic world, is to create a global village. Indeed physical distances are abolished but many logical topologies have materialized; some just for the sake of greedy business or totalitarian countries, and so at the general expense.

    I am just reading again Jane Jacobs and trying to translate her concepts to the virtual world of Internet.
    I am not talking second life sillyness, meaning aping the real world but of the properties of the virtual world we dwell in. Instead of a global village, we have logical cities with various sizes and properties. Topologies and boundaries translate to protocols and frameworks that define how information circulates and people interact.
    Net denizen belong to many communities. The big city blocks decried by Jane are the Disney or AOL gated worlds and so on. Really the works of Jane Jacobs or Lewis Mumford should be read again in the light of the modern Internet and its increasing social dimensions.

  7. addicted Says:

    Why do people get so touchy about the use of college degrees as a proxy for intelligence?

    That doesn’t mean that everyone who has a degree is smart, or that you have to have a degree to be smart. Its just the easiest measurable that is reasonably correlated to intelligence.

  8. David Says:


    You are so right. In the aggregate, it an excellent proxy for economic viability of an area. The S.F Bay area and Seattle are good positive examples – many Rust Belt cities are negative examples.

    Of course, there are examples of non-college degree holders succeeding (Bill Gates), but in general, the better educated a populace, the better off it will be economically.

  9. Michael Wells Says:

    Interesting. One of the things Jane Jacobs talked about in density was the number of people you were likely to come into contact with in daily life. Dense apartment-oriented cities like NY, SF, Boston have many more people coming into contact with each other. But Jacobs and others also talk about the creative energy that comes from different kinds of people living & coming in contact with each other. While college degrees are a pretty broad measure, they still are a limited universe in terms of diversity.

    We were just in Bellevue, Washington home of Microsoft. One of the things we noticed downtown was how homogenous it was. Lots of 30-somethings, dressed alike and geeky. So the density doesn’t give a lot of diversity that I could see.

  10. Ryan Says:

    If this where Civilization IV we would be losing the cultural victory.

  11. brillega Says:

    “degree holder” does not imply “smart person” although it may generally imply “hireable office/knowledge worker”

    Also, using this chart is biased to begin with, because it’s not normalized to take into account the higher density of PEOPLE IN GENERAL, smart or otherwise in NY and SF.

    Without comparing this chart to the chart for density of all humans, it could be giving the exact same information as a chart showing “Density of Circus Clowns Per Square Mile” or “Density of Kanye West Fans”

    Thought experiment: what would this look like next to a chart showing “Density of Homosexuals per square mile”? I bet THAT chart would actually give some useful info.

  12. Fred Says:

    King County (Seattle) is about 40% national forest and protected watershed. Thus, the correct land area to use in calculating density is considerably less than the figure used. I suspect the east side of Lake Washington from Mercer Island north has a greater density of degrees than Seattle itself. I agree with Michael Wells that there is a high concentration of geeky 30 something’s in Bellevue – a lot of whom are immigrants to the US. Consequently, although there is not a lot of educational and income diversity it is a cosmopolitan group.
    Even though it is possible to stay connected with friends around the world and jointly work on projects with colleagues from different geographic areas I still think the casual unexpected chances to meet new people face to face in a dense area are important in formulating new approaches to problem solving.

  13. Anen Says:

    This is more a measure of the fact that the most expensive areas to live in require a college degree to obtain a job which pays well enough to live there.

    Outside of that, just packing degree-holding persons together may not result in some massive measure of economic creativity. We don’t think of San Francisco when we think of the top job producing productive centers of the world.

  14. David Says:

    From the St. Louis Federal Reserve Bank:


    One of the fundamental sources of good job growth is an educated labor force. Within the last three decades, the demand for highly educated workers has grown dramatically in the United States. In 1980, the average proportion of workers across all 200 industries with some education at the college level was 32 percent. By 2000, it had risen to 51 percent. In fact, no industry saw its proportion of college-educated workers decrease over this period.

    At the same time, it is also true that high-paying jobs tend to have a particularly strong demand for college-educated workers. Among the top 25 percent of jobs in the sample, the average proportion of workers with a bachelor’s degree rose from 18 percent in 1980 to 36 percent in 2000. The average proportion of workers with a bachelor’s degree in the bottom 25 percent of jobs also increased over this period, although by a much smaller amount: 10.8 percent to 12.9 percent. These results suggest that the growth of good jobs can be expected to occur in cities with highly educated populations.

    The evidence strongly supports this conclusion. A 1 percentage point increase in the share of a city’s adult population (i.e., at least 25 years of age) with a bachelor’s degree is associated with a 1.2 percentage point increase in the rate at which good jobs are created over the next 10 years. Other measures of education yield similar results. Cities with larger numbers of colleges and universities and employment accounted for by institutions of higher education (a measure of the extent of the university community) tend to exhibit a significantly faster growth rate for good jobs.”

  15. SleepyJean Says:

    Yeah, he’s right when he says that this comparison isn’t entirely fair. LA is way too spread out to compete against cities like SF and NYC where people are living on top of each other.

    Also, LA doesn’t really have a “primary city” so I’m not sure what area was considered. And SF gets TWO primary cities? What … Berkeley, where there is an enormous University and many professors? Hogwash.

  16. Kiko Says:

    As other people have mentioned, a college degree is not the greatest determinant of intelligence. Having a college degree shows that you are capable of following through with instruction and know how to work the system to get what you want. Teachers and curriculum nowadays are too lenient. Many graduates can achieve a 4.0 gpa but all this signifies is that they can follow the rules and turn things in on time. The system is set up for students to complete assignments but nothing is said about learning. Rather than intelligence, assertiveness should be seen as the driving factor. The study should have also looked at population density. Ofcourse there is going to be more people with college degrees in New York and San Francisco because they have a high population density. Next, type of jobs that are available in these areas should be factored in. Applicable experience is prefered over degree noteriety in some areas. You need to way in the cost effectiveness of jobs, available resources/geographical location, the location of school versus cost to attend, jobs that are present in the area………..there are so many factors that he did not include that this study should only be considered introductory and maybe just a request for futher research/funding to conduct this study properly.

  17. Scott W Says:

    Wouldn’t a per-capita measure be a better indicator of education density? This just shows how closely people live together and the list would probably be in the same order for adults who didn’t graduate from high school.

  18. David Says:

    Those of you who are countering that a degree is necessary for success are attacking the wrong argument. You are looking at individual cases, or at substrata of people who have degrees and perhaps shouldn’t, or at those who succeed without them. It is a macro metric, not a micro one. Think of it as an economic indicator. Areas with higher proportions of the populace with at least a bachelor’s degree fare much better economically. Why is that case? A whole host of reasons, both elementary and complex.

    For further reading on this topic, I strongly recommend the work of the Harvard economist, Ed Glaeser.

  19. David Says:

    Meant to say “degree is not necessary for success” in the comment above.

  20. Contributor Says:

    New Haven, CT would be tied with San Jose CA on this.

  21. Daniel Carins Says:

    “Retirees who produce nothing”.

    Are you suggesting the virile, competent college graduates put them in a gas chamber and euthenase them, Wil? You could call it “The Final Solution”.

    Just because someone is retired doesn’t mean they produce “nothing”. How about unpaid childcare, consumption of goods and services that support jobs, how about payment of taxation on taxable income and consumption, how about simply living and contributing to civil society?

  22. Steve Mouzon Says:

    These numbers are a bit misleading, as the cities with the highest density of smart people likely have the highest density of idiots, too… because they’re generally the cities with the highest density. Put another way, walking down a street in Manhattan doesn’t necessarily make a chance encounter with a total stranger any more likely to occur with a really smart person than with the cerebrally challenged. For that, you need intelligence per capita. And for that, places like Cambridge, Mass., North Carolina’s Research Triangle, and my almost-completely unwalkable hometown of Huntsville, Alabama (Marshall Space Flight Center, etc.) are all near the top of the list.

    One caveat… once a network is established, it’s clearly better to have the nodes of that network closer rather than more distant… but not that much better… yet. Until driving becomes a major disincentive, places like Huntsville will thrive at innovation because all the smart people can still drive to their meetings with other smart people. And because they’re smart people, many of them are already figuring out ways of meeting seamlessly (not your garden-variety videoconference, but rather, really cool stuff) even after the Automobile Age is over.

  23. Deep Says:

    For every CEO that didn’t finish college, there are tons of college dropouts that are making less than they would be had they stayed in school. Opportunities for people to work their way to the top is diminishing in most industries. Whether a college degree makes you smart or not is debatable, but a college degree makes you significantly more prepared in the new economy. Even if you have a philosophy degree, you’re more marketable than someone who hasn’t gone to college. The data shows for every 1% increase in the proportion of college educated residents a city has, it sees a 2.3% increase in productivity. This is why cities like Boston, Austin, and San Francisco are so productive, over 30% of their residents have a college degree. On the flip side, Detroit and Buffalo have around 10% of its population with a college degree.

  24. Deep Says:

    @Michael-Bellvue is a company town. It is really no different than towns in Central Pennsylvania that were controled by the coal companies. There are very little reason for people in other industries to go there. Why go there when everything is in Seattle or Vancouver? Microsoft is big enough to not have to be in the middle of everything.

  25. George M Says:

    The places with the most degrees are also the places where for every one job opening there are 50+ applicants! (aka- the rat race!)

  26. justin kruger Says:

    This should be done based on commute time.

    Density of intelligence based on a 45min commute. That should normalize across all geographies and map well to how employer and organizations recruit.

  27. Wil Says:

    One thing I have learned from being an employer is that college degrees are overrated. Take a look at this list of famous successful dropouts, it might make you wonder if dropping out is the key to success.

  28. David Says:


    Lists like that do not prove a lot – of course, there are people that are going to be big successes regardless of whether they have a phd or dropped out of school in the 8th grade. Bill Gates wasn’t a blue collar worker who one day become an I.T. pioneer. He clearly benefited from his education, and did attend Harvard. Seattle has one of the highest rates of college graduation in the U.S. I’d bet that a lot of Microsoft employees are college graduates.

    Once again, the macro point is being missed and individual cases are cited. In general, those with college degrees tend to be hard-working, ambitious, persistent, adaptable, goal-oriented, etc. The unemployment rate for college grads is 5%, HS grads 10% and HS dropouts 15%. Clearly the “market” values a college education.

    Suggested reading:

  29. Cliff Lippard Says:

    Mr. Pitingolo is correct about there being problems with the comparisons. Nashville, Indianapolis, and Louisville are all metropolitan counties, meaning that their city borders include large areas outside of their true urbanized area. This would distort the density measure. I think to really compare these cities with others one would need to look at a much finer level, perhaps using census tracts.

  30. Cliff Lippard Says:

    A quick and dirty calculation for Nashville using Census’ urbanized area boundaries and education attainment data returned a degree holder density of 424 per sq. mile.

  31. Mike Says:

    First I must agree with many before me that formal education does not equal “smart” and neither has an impact on the ability to strive or excel in business. For those that have forgotten, Henry Ford did not finish high school, Michael Dell dropped out of college, and Richard Branson became an entrepreneur at 16. I don’t think anyone would argue these men are all captains of industry.

    Next, we need to properly define Human Capital. It is more about the investment in the individual than it is about degrees and population. Human Capital represents the value of a particular or team of individuals to the organization they work for. Retaining human capital is a function of management and there is little relevance on the employee or their education.

    I must completely disagree with the authors definition of economic stimulation and recovery. A true sign of recovery is when company’s must do more to retain their “human capital” for fear they will leave for other opportunities, and has nothing to do with the population holding college degrees. An electrician with 30 qualified mechanics, none of which hold a degree, has more human capital at stake than any financial institution employing hundreds of analysts who regurgitate publications and obvious trends to sound intelligent.

  32. SwBratcher Says:

    I’m with Joe Zuccaro on this one. His comment above sheds light from a perspective the author seems to lack.

  33. halten_Sie Says:

    The data need be normalized by the population density for each respective city.

    Dimensional analysis:

    (degrees/area) * (area/population) = degrees/population

    This would be a dimensionless ratio. Your data are given in units of degrees and inverse area which by itself is meaningless.