Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a university professor in the University of Toronto’s School of Cities and Rotman School of Management, and a distinguished fellow at New York University’s Schack Institute of Real Estate and visiting fellow at Florida International University.
A new study maps digital-skilled jobs across industries, metro areas, and demographic groups, revealing deep divides.
We’ve long heard about America’s digital divide, but the nation is facing a parallel and deepening digitalization divide, too. According to a new Brookings Institution study, this digitalization divide is reflected in the increasingly uneven spread of high-paying digital jobs across the economy and workforce, by gender, race, and ethnicity, and across cities and metro areas.
Digitalization has transformed just about everything we do—from the way we work to our entertainment choices, and how we communicate with one another. It’s not just knowledge workers, smart phones, and laptops we’re talking about. Our cars, televisions, and appliances are loaded with software. Smart thermostats like Nest and voice activated “assistants” such as Alexa permeate our homes. Formerly manual jobs, like those of auto mechanics, are now more about the use of software than mechanical skills. Already, there is one smart phone for every person on Earth. And today, more than a tenth of Americans use a voice-activated personal assistant, like the Amazon Echo or Google Home, around the house.
But what is the impact of such sweeping digitalization on the geography of our workforce and economy?
The report’s authors—Mark Muro, Sifan Liu, and Siddharth Kulkarni of Brookings’s Metropolitan Policy program—provide perhaps the most comprehensive analysis to date of the impact of digitalization on America’s workforce. The report documents the changes in the digital content for nearly 600 occupations, using data from the Occupation Information Network, or O*NET, a database from the Employment and Training Administration (ETA). O*NET’s database offers incredibly fine-grained information on job content, tasks, skills, creativity, education, training, and more, based on surveys of thousands of individual workers.
The study, which covers the period 2002 to 2016, uses two key O*NET questions to analyze digitalization: the importance of knowledge of computers and electronics, and the importance of computers to doing a job. Based on that, it distinguishes three tiers of digitalization of work. Highly digital jobs, like software developers or financial analysts, scored 60 or above on a 100-point scale; medium-digital jobs, like nurses or sales managers, scored between 33 and 60; and low-digital jobs, like construction workers or personal care aids, scored below 33. (The study notes some limitations to these O*NET data, which are reported as aggregates, and not available on the micro level. This means that digitalization scores are assigned nationally, without regard to location, which could introduce potential inaccuracies at the local level.)
Today, more than 32 million workers are employed in highly digital jobs, while nearly 66 million work in medium-digital positions, and 41 million work in low-digital ones. Indeed, there has been a massive increase in the share of highly digital jobs between 2002 and 2016, when these jobs jumped from less than five percent to nearly a quarter of all U.S. jobs. Over the same period, the share of medium-digital jobs increased from roughly 40 to 48 percent, while the share of low-digital roles declined significantly, from 56 percent to 30 percent. Roughly two-thirds of new jobs created since 2010 required either high- or medium-level digital skills, and some 4 million, or 30 percent, required highly digital skills.
That said, all industries have become more digital. The mean digital scores for high-, medium-, and low-digital industries have all increased significantly since 2002. According to co-author Muro, “the share of them that require no or minimal digital skills is dwindling rapidly.” More and more workers—across all skill and wage groups—are seeing their jobs impacted by digitalization.
Furthermore, while computerization is typically seen as something that is likely to erode wages and eliminate jobs, the study finds that the more digital a job has become, the higher it is likely to pay. The mean annual wage in 2016 for workers in highly digital occupations was $72,896, compared to $48,274 for middle-level digital jobs and $30,393 for low-digital positions. The study found that on the whole, controlling for other factors, workers with superior digital skills earn higher wages than similarly educated people who have lower digital skill levels.
Wages are not only higher for more highly digitalized occupations, they have grown faster, increasing at 0.8 percent a year for high- digital jobs compared to 0.3 percent for medium-digital jobs and 0.2 percent for low-digital jobs. One reason for this is that 60 percent of low-digital jobs are susceptible to automation, while the same is true for only about 30 percent of the tasks in high-digital occupations.
Digitalization also reflects and reinforces America’s divides in gender, race, and ethnicity, according to the study. While women have slightly higher overall digital scores than men (as the chart below shows), men are overrepresented in the highest-scoring, and best-paying, digital jobs. Whites and Asians are over-represented in highly digital jobs, while blacks and Latinos are under-represented.
The Uneven Geography of Digitalization
Perhaps the most striking findings of the study concern the incredibly uneven spread of digitalization across America’s cities and metros. Take a look at the map below, which charts digitalization across America’s metros. It provides yet another lens onto America’s winner-take-all urbanism and deepening spatial inequality. At the top of the list are the tech hubs of San Jose and Boston, followed by San Francisco, Austin, Hartford, Seattle, Madison, Raleigh, and Salt Lake City—all epicenters of the knowledge economy.
The majority of Rust Belt and Sunbelt metros have relatively low digitalization scores, although Baltimore, Philadelphia, and Albany and Rochester, New York, score quite high on digitalization. Near the bottom of the list comes Las Vegas, as well as Stockton, Fresno, Bakersfield, and Riverside—the last three of which suggest that digitalization’s divide is fractal, and occurs even within states like California, which contains metros with both the highest and lowest digitalization scores.
Still, digitalization has increased in virtually all metros over the past decade and a half. And while this has tended to exacerbate the divide between the digital haves and have-nots—San Jose, Washington, D.C., and Austin all saw dramatic growth in digitalization, while Stockton, California, Youngstown, Ohio, and McAllen, Texas, saw much slower digital growth—Rust Belt metros like Rochester and Akron, and Sunbelt ones such as Phoenix, Little Rock, and Charleston all saw substantial growth in digitalization over this period.
When it comes to the geographic impacts of digitalization, Muro notes, two things are happening. On the one hand, the least digital places are starting to “catch up” by implementing the most basic technology and processes. But on the other hand, the most digital places continue to forge ahead, so work in these places is becoming even more digital. As a consequence, the gap between the two has continued to grow over time.
There is a close correlation between a metro’s level of digitalization and its wages, as shown by the chart below. Digital metros, like more highly educated and skilled metros, have stronger economies with higher wages, reflecting and reinforcing the broader pattern of spatial inequality.
What Can Be Done
The report outlines three key strategies for regional groups of business, labor, universities, and local and/or state governments trying to cope with, and lessen, the digitalization divide.
The first is to expand the digital talent pipeline by upgrading work and upskilling workers. There are various strategies that industry, labor, and non-profit groups are using to do this, as the report outlines, such as tuition reimbursement programs and online learning models for a range of education levels. The authors also encourage governments to incentivize company-based digital training by expanding existing tax benefits for education assistance programs.
The second is to expand digital literacy across the board, and especially for under-represented groups. As more and more jobs, including low-skilled ones, become digitalized, they can be on-ramps for employment and careers. The study encourages businesses and cities to develop initiatives and strategies which introduce everyday software, such as spreadsheets and word processors, to all Americans at a young age. This could start as early as elementary school, or in boarder awareness campaigns that tap into social media to familiarize everybody with these tools.
The third is to cultivate not just “hard skills,” like programming, but also essential “soft skills”—the quintessentially human skills of adaptability, emotional intelligence, and curiosity that a computer cannot replace. As Muro puts is, workers, companies, and places should focus on “getting better at doing what the machines can’t do.”
Digitalization both reflects and reinforces America’s deepening socioeconomic and geographic divides. Overcoming it will require broad strategies that address the root problems, and help America’s industries, jobs, and metro regions move toward more inclusive prosperity.
CORRECTION: In a previous version of this article, we reported that the O*NET database comes from the Bureau of Labor Statistics. In fact, the O*NET database comes from the U.S. Employment and Training Administration, another arm of the Department of Labor.