Tech empowers, tech polarizes

Tech is a hugely empowering force in the workplace, but it also polarizes. At a moment of disquiet about the nation’s West Coast tech giants, that’s the ambiguous synopsis of the big new report we at Brookings Metro just released on the progress of “digitalization”—the diffusion of digital technology into nearly every business and workplace.

Our report is, in this sense, a true good news / bad news story.

On the positive side, the analysis strongly confirms and quantifies the sizable economic benefits digital technologies are conferring on many of their users, whether they be individual workers, industries, or places. Workers all across the skills continuum are being rewarded for the depth and breadth of their digital skills through increased wages. In 2016, for example, workers in occupations with mid- or high-level digital skills requirements were paid significantly more than those in less-digital occupations.

metro_20171121_Figure 5 Mean annual wage by dig level

Digital technologies are, in this fashion, increasing the potential of individuals and society

In like fashion, digitalization appears to provide industries a route toward greater productivity, and metropolitan areas—which are rapidly increasing their mean digital scores—a route to improved prosperity. Digital technologies are, in this fashion, increasing the potential of individuals and society.

And yet, there is a problem. For all of its benefits, it is impossible (looking at the occupational data) to avoid the conclusion that the digital wave is contributing to the nation’s massive inequality problem. Leave aside the fact that our work suggests digital knowledge may be a partial stay against automation-driven job disruption. (I’ll say more about that in a forthcoming post). Even so, the new analysis adds to the accumulating evidence that digitalization is likely contributing to worker pay disparities, the “hollowing out” of job creation, and the divergence of metropolitan economic outcomes.

The digital wave is contributing to the nation’s massive inequality problem

How can this be? Part of the unevenness effect has to do with the inherent nature of digital technologies, and part of it has to do with the uneven distribution of digital skills. On the technological side, computers have long been recognized to have inherent “superstar” impacts that flow from the built-in power of high-speed computing and IT gear to massively augment the ability of skilled workers, firms, and places to collect and manage information, speed calculations, accelerate sharing, and multiply the reach of enterprises with low- to- no marginal cost. Tech may increase all workers’ capacity, but it especially augments the productivity of highly skilled, creative, or digitally adept workers, firms, or places. So digital technology is biased to some, not everyone. As to the matter of distribution, both digital skills and access to highly digital, good-paying roles are unevenly apportioned. Troubling digital skills gaps exacerbate the inherent skill biases of these technologies.

In keeping with these realities, the “digitalization of everything” is associated with stark and increasing labor market and geographical divides. To begin with, the wage premium for digital skills (highlighted above) has nearly doubled since 2002. That means that the mean pay of workers in higher-level digital occupations (about $73,000) now more than doubles the $30,000 wage of low-digital workers. This is labor market polarization at work.

Likewise, our data highlight extraordinary patterns of over- and under-representation of various gender and ethnic groups across occupational groups, in part but not exclusively driven by those groups’ digital skills. Take a look below how women—despite having slightly higher mean digital scores—remain heavily under-represented in highly digital computer and engineering occupations (but over-represented in office admin, education, and health occupations). Likewise, blacks are overrepresented in medium-digital occupations like office support and health care support, while Hispanics are greatly underrepresented in highly digital tech positions and overrepresented in low-digital domains such as farming, construction, and building and grounds maintenance.

metro_20171121_Table 8_over and under representation of gender racial groups

At the same time, our data add to the growing array of studies that suggest that the digital revolution has contributed to the “hollowing out” of the employment and pay continuum. This is the famous discovery of MIT economist David Autor and his colleagues, who argue that as digital technologies pervade business processes and redefine roles, they alter what types of workers are hired and what they are paid. Why is this? It’s because they “substitute” for rote work while “complementing” the more creative work of workers who perform non-routine or complex problem-solving and inter-personal tasks. And so in parallel fashion, our analysis finds that job creation and wage growth have each been relatively robust since 2010 for both highly digital computer-mathematical and business-finance occupational groups and low-digital occupational fields such as personal care and food preparation. By contrast, mid-digital occupational groups like office administration and educational categories have seen much slower job and wage growth.

metro_20171121_Figure 9_annual growth rate

And then there is the fact that while almost all metropolitan areas are seeing their mean digital scores rise rapidly, metros with the highest shares of highly digital workers—San Jose, Washington, Salt Lake, Boston, Austin, for instance– have been seeing their shares grow even faster than have other metros. In this regard, the digital rich among metro areas appear to be getting even richer, and now—as a consequence—are pulling away from the rest of metros on basic measures of prosperity. Here’s how that looks:

metro_20171121_Figure 14_wage and digital scores

This too is polarization at work. Such polarization is one more way in which the trend of digitalization looks very much like a major contributor to the widening pay and prosperity gaps that we have written about here and here and that the economist Enrico Moretti calls “the Great Divergence.”

Individuals, industries, and regions all need to sign onto the project of improving workers’ skills

As to what should be done about tech’s vast two-track impact on the economy and the world of work, some things are clear, but not everything. What’s most clear is that individuals, industries, and regions all need to sign onto the project of improving workers’ skills—including both higher-end digital skills, including coding, as well as more entry-level basic knowledge of workplace productivity technology. Such upskilling has now become urgent, especially since digital skills have become a prerequisite for maximizing the good of tech, including for basic economic inclusion. About this our report has a good deal to say and I will have more to say in coming blogs.

Less clear, however, is how to push back against the inherently polarizing nature of digital technologies, with their great power to both empower and divide. Certainly a starting point is to better understand the economic and labor-market influences of these technologies, at a time when the world is scrambling to do that. Likewise, policymakers—as they acquire greater recognition of tech’s built-in tendencies—are likely going to need to realize that those tendencies may well require more concerted efforts to mitigate negative dynamics. Dynamics like labor-market and regional polarization are not going to be simple to grasp, let alone limit, but as I wrote recently in a piece with my colleague Amy Liu, it’s time to get started on that work. At present, America possesses no serious policy approach for addressing tech-based polarization of the economy—and it needs one.