By Mark Muro, Alan Berube, and Jacob Whiton

The astonishing highs and lows of tech this year have thrown new attention onto just who is working in the sector and with what set of biases, assumptions, and priorities.

In a previous post, we used data from our new report “Digitalization and the American Workforce” to assess the state of women’s access to work, opportunity, and pay in the industry. The picture that emerged then was mixed—reflective of both significant acquisition of tech skills on the part of women and the industry sexism highlighted by former Google engineer James Damore’s explosive memo last summer.

But what about the state of racial and ethnic representation in tech? This, too, has become a huge topic of concern. And so we again want to build on our earlier work on the digital workforce to examine a mixed state of affairs, albeit one with several bright spots.

What do the data say? To dig beneath the broad story of steadily increasing digital skills among black and Hispanic workers found in our report, we used the 4.6-million-worker computer and mathematical (C&M) occupational group as a useful stand-in for “tech employment.” And here the picture for blacks and Hispanics nationwide is especially mixed.

Nationwide progress for blacks and Hispanics—but significant underrepresentation overall

Broadly speaking, blacks and Hispanics have made genuine progress in penetrating the nation’s tech sector. Blacks, for example, have increased their presence in several important tech occupations, such as computer programming and operations research. Likewise, Hispanics have increased their representation in the overall C&M occupational group, moving from 5.5 percent of workers in the sector in 2002 to 6.8 percent of workers in 2016.

And yet, with that said, the presence of blacks and Hispanics in computer and math jobs remains starkly inadequate at the national level. Blacks make up 11.9 percent of all workers but only 7.9 percent of C&M workers. The gap is even larger for Hispanics, who make up 16.7 percent of all workers but only 6.8 percent of C&M workers.

Moreover, in some instances the diversity picture is growing bleaker. Black representation in C&M occupations is moving in the wrong direction, slipping from 8.1 percent of the field in 2002 to 7.9 percent in 2016. Only 4.4 percent and 2.9 percent of database administrators were black and Hispanic, respectively, in 2016, but even those small shares reflect declines since 2002. While tech companies across the nation have over and over said they must do better, blacks and Hispanics still remain underrepresented in tech jobs by nearly 50 percent.

Wide variation—and some bright spots—at the local level

Beneath the national picture, data from metropolitan areas reveal sizable variations of black and Hispanic representation in tech.

To be sure, blacks and Hispanics each remain almost universally underrepresented in C&M occupations across U.S. metros. Blacks’ share of C&M employment, for example, falls short of their workforce share in 89 of the 97 largest metros for which we have data, with some of the worst representation gaps in metros like Jackson, Miss. (-17.4 percentage points); Memphis, Tenn. (-15.9 percentage points); and Richmond, Va. (-14.0 percentage points):

And Hispanic underrepresentation extends to 88 of 90 large metros, with some of the worst underrepresentation in C&M occupations occurring in Bakersfield, Calif. (-27.9 percentage points), Los Angeles (-27.6 percentage points), and Oxnard-Thousand Oaks-Ventura, Calif. (27.3 percentage points):

Relatedly, 49 and 43 of the largest metros have seen their representations of blacks and Hispanics, respectively, actually decline since 2010. Representation has slipped the most for blacks in Columbus, Ga. (by 22.6 percentage points); Huntsville, Ala. (-12.1 percentage points); and Durham-Chapel Hill, N.C. (-11.3 percentage points). Parallel problems have affected Hispanics in Sunbelt metro areas: McAllen, Texas (-12.4 percentage points); Tucson, Ariz. (-5.9 percentage points); and Stockton, Calif. (-3.7 percentage points). Overall, black underrepresentation appears worse in the historically “black-white” metros of the Midwest and South, whereas Hispanic underrepresentation appears worst in the Southwest, where inroads into tech remain modest and Hispanic populations remain heavily clustered.

And yet for all of these daunting statistics, the story varies in some places, and suggests both variation across many places and genuine progress in some spots.

In regional terms, blacks and Hispanics are doing relatively better in the West and East, respectively, where each groups’ underrepresentation appears less pronounced.

In terms of absolute representation, black underrepresentation is currently less than 5 percent in 39 out of 97 larger metros across the country, with Deltona, Fla.; Las Vegas; and Greenville, S.C. appearing to be some of the most inclusive markets:

In parallel fashion, Hispanic representation achieves the same standard in a third of the 90 larger metros for which we have data, with St. Louis, Pittsburgh, and Rochester appearing the best places for inclusion, followed by Buffalo, Cincinnati, Columbus, and Baltimore:

In short, while blacks’ and Hispanics’ inclusion in tech is unacceptably low almost everywhere, a few increments of progress are being made in a variety of cities—most often markets characterized by longstanding black or Hispanic presence.

Community and company actions can help combat exclusion in tech

What can be done to improve things? Reversing black and Hispanic populations’ broad exclusion from tech occupations that promise significant upward mobility is for sure critical to long-run efforts to address racial inequality. Along these lines, our recent “digitalization” agenda stresses efforts on all fronts to expand the diversity of the IT talent pipeline, at the high end of the skills continuum, as well as to radically expand exposure to entry-level tech skills. Many of these efforts can take root at the local level.

At the high end, industry, the education and training ecosystem, and communities need to bear down much more on special models and outreach designed to engage underrepresented populations in incumbent-worker upskilling, work-based training, accelerated learning solutions, strong IT career on-ramps, and inclusive IT/computer science (CS) education from K-12 to university. Such models include work-oriented skills development such as Year Up; ethnically oriented “bootcamps” and coding programs like Hack Reactor’s Telegraph Track, Black Girls Code, or #YesWeCode; as well as Code.org’s state-by-state campaign to draw more students and teachers of color into its efforts to expand access to CS in schools.

Other models emphasize exposure to entry-level tech opportunities; organizations like Byte Back and Per Scholas are providing frequently overlooked computer and IT certification classes, while tech giants Microsoft and Google are rolling out focused online curricula for certifications in basic office productivity programs or IT support. A robust set of innovative tools now exist to support tech inclusion.

Increasing opportunities for people of color in one of the fastest-growing and highest-paying sectors of the economy is going to require more insistent efforts to bring about real change.

And yet, our data—which find blacks’ and Hispanics’ shares in high-tech occupations like programming and database administration moving in the wrong direction—suggest that current skills-based solutions won’t by themselves solve the nation’s digital inclusion problem. It is telling that black and Hispanic presence in high-tech occupations remains patchy despite the investment of as much as $1.2 billion over the last five years in tech-pipeline diversity programs including at non-profits and universities, which is why increasing opportunities for people of color in one of the fastest-growing and highest-paying sectors of the economy is going to require more insistent efforts to bring about real change.

Specifically, true inclusion will only occur when now-underrepresented groups accrue genuine power in tech, as noted by OpenMIC.org (and as the scholar Mar Hicks argued in our recent discussion of gender imbalances in tech). One path to that future of greater black and Hispanic power in tech is through entrepreneurship. For example, Atlanta-based entrepreneur Rodney Sampson, through his Opportunity Hub initiative, is working to promote inclusive high-tech innovation, entrepreneurship, and investment ecosystems as well as potentially breakthrough experiences for students from minority-serving higher-ed institutions. When new people accumulate true wealth, it changes things.

However, broad-scale progress is also going to require the tech sector to change its hiring and promotion practices to bring about a more equitable distributions of power within companies—power that can change hiring outcomes. Along these lines, tech companies need, as Intel//Dalbergnotes, to publish more data on how they are doing on diversity and commit to time-bound goals, with built-in accountability mechanisms. With goals like Intel/Dalberg’s “five in five” in place—that is, commitments to a five-point increase in racial, ethnic, and female representation in five years—tech company boards could then make them stick by linking compensation and incentives to the achievement of the goals.

In sum, with aggressive actions such as those mentioned above, companies could benefit themselves as well as advance the economy inclusively, all while benefitting employees of all ethnicities and genders. In the context of a growing affluence gap in the U.S., increasing the opportunities of people of color in one of America’s highest-paying, fastest-growing sectors would be a transformative step towards widening the circle of inclusive growth.


See here for data and rankings on blacks’ and Hispanics’ shares of Computer & Mathematical occupations across all metro areas.