The Geographic Impact of Automation

Projected job growth by 2030. (Darker colors indicate faster job growth). Source: McKinsey Global Institute. Click for larger image.

by James A. Bacon

A handful of megacities have captured a majority of U.S. job growth since the Great Recession and could win 60% of job growth through 2030, according to a July McKinsey Global Institute report. A middle tier of “stable” metropolitan areas and thriving niche cities will continue to see job growth, though at a more modest rate than the megacities, while the bottom tier of lagging metros and rural areas will see only marginal growth, if any.

The differential rates of job growth will be driven in part of the next wave of automation, which will displace many office-support, food-service, manufacturing, and customer-service jobs, while a dynamic economy creates more jobs in healthcare, STEM fields, business services, and work requiring personal interaction, says the report, “The Future of Work in America: People and Places, Today and Tomorrow.” “While there could be positive net job growth at the national level, new jobs may not appear in the same places. The challenge will be in addressing local mismatches and help workers gain new skills.”

If McKinsey’s “midpoint” job projections are close to the mark, the Washington metropolitan area will continue to dominate job growth in Virginia, while “stable” metros like Richmond and Hampton Roads will contribute to a lesser degree. The Shenandoah Valley and Roanoke-Lynchburg area will see marginal growth, and the rest of the state negative job growth.

These conclusions put a filigree on what everyone already knows about the challenges facing rural Virginia. What, then, is to be done? McKinsey offers some general strategies for adapting to our brave new world that sound remarkably similar to what I had to say in yesterday’s blog post about rural development. And I quote:

For rural counties, the road is [tough]. Many of these places lack the economic base or the inflows of investment or people to create new jobs. No amount of workforce retraining can solve the bigger challenge of lack of economic activity. Individual companies can help to ease this strain by considering whether there is a business case for establishing operations in more affordable parts of the country that need the investment.

Turning around places that have lost their economic dynamism is a multiyear journey, but it is possible. Each community will have to take inventory of his assets, such as available industrial space, natural attractions, local universities, and specialized workforce skills. That data can form the basis of an economic development plan built around a growth engine industry that can create jobs and spillover effects.

McKinsey is somewhat more optimistic than I am about the ability of rural areas to attract outside capital. I have repeatedly argued that, absent a reversal in the trend to globalization, the cheap land/cheap labor strategy of Virginia mill towns provides diminishing returns because, by global standards, land and labor in rural Virginia are not cheap. But McKinsey says, “The growing acceptance of remote working models could be a positive trend for creating jobs in rural counties, whether full-time work-at-home employee roles of contract work. But it will take a push to continue building out fast, affordable broadband in the regions that still need service.”

Bacon’s bottom line: Economic megatrends in an increasingly knowledge-intensive global economy favor large cities and put small towns and rural areas at a severe competitive disadvantage. Doubling down on the 70-year-old economic development strategy of recruiting light manufacturing jobs will yield limited results. Even if Virginia counties snag industrial investment, continued automation ensures that the number of jobs will be smaller than in the past. Rural Virginia communities must seek to find a more sustainable economic base — in many cases supporting fewer jobs — and learn how to provide essential local government services to a shrinking population.

Share this article


(comments below)


(comments below)


8 responses to “The Geographic Impact of Automation”

  1. Reed Fawell 3rd Avatar
    Reed Fawell 3rd

    Jim, what amazes me is that you apparently believe all this stuff put out by group thinkers pretending to predict the future, just like all the self important fools who went before them, predicting ice ages, imminent human extinction, and all the other nonsense on Steve Haner’s wondrous list found at:

    Let me let you Jim in on a little secret. The McKinsey Global Institute has not the foggiest notion what the future of the Shenandoah Valley of Virginia is.

  2. Reed, I completely agree with your premise; but somebody has to try to predict the future or it will simply happen, unforeseen and unappreciated except in hindsight. And you and I won’t be here to dazzle the worshippers with our hindsight much longer.

    As for McKinsey, I’ve known a few of their minions over the years and they have a remarkable skill for packaging common sense along with a superficial study as astute insight, then selling it to companies that are willing to pay for the fancy packaging and the McKinsey label. One reason they get away with it is that their idea of a superficial study nevertheless likely exceeds any other locally-focused economic study readily available and likely even looks profound to their target audience.

    Therefore I agree, also completely, with Jim, “Doubling down on the 70-year-old economic development strategy of recruiting light manufacturing jobs will yield limited results.” When McKinsey says, “That data can form the basis of an economic development plan built around a growth engine industry that can create jobs and spillover effects” — in fact they are simply selling their services to lead the wild goose chase after the data that will prove whatever “economic development” scheme they suspect already has local political backing. Economic Development Cautionary Tale No. 1: Warren County.

    1. Reed Fawell 3rd Avatar
      Reed Fawell 3rd

      Likely Acbar, we are both right here to a degree. The problem lies both with the experts and the receivers of expert opinion. In my opinion, experts as to the future are dead wrong the great and overwhelming majority of the time, but occasionally, very occasionally, they are right. And often times there a grains of truth in what they say. I learned this first in real estate many times over, never to follow the herd and conventional wisdom. I could give you a long list there. I learned it again and again too studying military history, how the experts as to the next war were most always wrong in depth and detail, even far into a war, though too, a few proved their view of the future right, at least in their salient part, here I speak of a few US Marines, and Earnest King and a few other Naval officers among the tens of thousand of officers who were dead wrong on most everything as to WW2 and I before that. I don’t differ with conventional wisdom because its different from mind, I now instinctively don’t believe it because history shows it is most always wrong, because the future is most always a big surprise to all but a few.

      I recently posted here a study of how the great majority of people go along with the herd, changing their opinions daily as needed to be in conformity. That is undisputed. It’s on steroids now, and in full display now, given instant communications, and gratification.

  3. djrippert Avatar

    Actually, the ability of the people derisively labeled in the comments as “experts” to predict technology trends is pretty good. In 1984 an executive from SUN Microsystems, John Gage, said, “The network is the computer”. On March 13, 2006 Amazon Web Services made its first sale of a cloud based product (S3 – Simple Storage Service) to a third party. The coming of cloud computing was predicted 24 years before its birth.

    Edgar Codd published “A relational model of data for large shared data banks” in 1970. However, the first commercially available relational database would not be released until 1979. Codd predicted the rise of relational databases almost a decade before they were born.

    On Feb 21, 2002 (4 years before Amazon launched cloud based IaaS and long before people talked about Platform As A Service) my team and I filed a patent entitled, “Distributed development environment for building internet applications at remote locations”. This effectively predicted that cloud (term was not used at the time we filed) would evolve from being an infrastructure for rent model to a full scale remote development model. This indeed is happening / has happened. If you are very tired and struggling to sleep it is US Patent #7870535

    The accuracy of predicting technological trends has always been pretty good but has improved significantly over the last 10 – 15 years. The problem is the general public willingness to believe. I was repeatedly told that companies would never run their application systems on relational databases in the early 1980s because of performance concerns. Wrong. I was repeatedly told that the cloud was no more than modern timesharing and would never be used by corporations due to security concerns. Wrong. I was told PaaS was a pipe dream. Wrong.

    In technology, it’s not the experts. It’s the half educated pseudo experts who can misinterpret anything.

    Trust me – AI and various forms of robotics (including purely software based robots like Robotic Process Automation or RPA) is coming at Virginia like a linebacker on PCP. The collision is not going to be pretty.

    Remember that Google’s AI program that played the complex Asian game Go (and beat the grandmaster) was never taught Go strategy. It simply went through thousands (millions) of Go games and taught itself strategy.

    Our General Assembly is as unprepared for the technological future as a Cro Magnum man would have been unprepared to fly an F-35.

    1. Don, I defer to your expertise in technology. If you say AI is bearing down on us like a linebacker on PCP, I believe you. And I agree that the General Assembly and state is like a deer in the headlights when it comes to the impact of technology. Let me ask you, which state legislatures impress you for their understanding of technology and anticipation of its social and economic impacts?

      1. djrippert Avatar

        Hard question because it comes in two parts. First, which legislatures understand technology? Second, which legislatures take prudent action based on the understanding of technology? The first question is pretty cut and dried. The second question requires an understanding of the partisan slant of the state legislature.

        Generally the more liberal states have the deepest technology industries and their legislatures understand the technology the best.

        For example, California has for years banned the use of non-compete agreements between companies and employees in most circumstances. Why? Because they understand that talent draws industry, industry does not draw talent. Talent friendly laws like banning non-competes hasn’t made California a pariah in the eyes of technology companies. Virginia’s elite would be aghast at such a thought. Why shouldn’t a corporation and an individual be able to voluntarily enter into a non-compete agreement? The answer is twofold – corporations are coercive by nature and such non-competes are not in the public’s best interests.

        More specifically on technology, SanFrancisco banned government use of facial recognition software. They understand how accurate that software is getting and they worry about government overreach and intrusion. So do I. Do you really want a record of your every public move, cataloged by your name stored forever in some government database? I don’t.

        One popular use of AI in the hiring process is through AI “interview bots,” which evaluate personal characteristics such as an applicant’s facial expression, body language, word choice, and tone of voice. The software then provides employers with feedback that can be used to evaluate whether to hire a candidate. Basically you get filmed in an interview with another person and then analyzed by an AI application. Especially with “over the internet” video interviews people don’t know their interview will be post processed by an AI app. Illinois has mandated disclosure, in advance, to the interviewee. Sounds reasonable to me.

        The European Union has passed the General Data Protection Regulation (GDPR) which holds companies collecting private data accountable for the use of that date and the disclosure of the collection to the person whose data is being harvested. California has done something similar with the California Consumer Privacy Act which is set to become law on Jan 1, 2020.

        In February, President Trump launched the American AI initiative, the US government’s first official view on AI, with a focus on driving tech breakthroughs. I’m not sure that this examines the consequences of AI but at least the Feds have started thing about AI.

        The UK launched the Centre for Data Ethics and Innovation, and the Office for AI – advisory groups within government to help lawmakers understand the implications of technology.

        Singapore’s government proposed an AI governance framework in January that takes what it calls a “human-centric” approach. Its plan highlights explainability, transparency and fairness as guiding principles, emphasizing the need to build public trust in AI. In parallel, the country has set up a S$150m ($109m) program, AI SG, to foster research and innovation.

        Meanwhile, little Ralphie Northam has downgraded the Secretary of Technology in Virginia to a non-cabinet level position – he basically eliminated the position.

        When you live in a state run by the plantation society elite you can’t hope for much. Northam’s longing for antebellum days was all too clear in that yearbook photo. I question whether he could successfully operate an electric can opener let alone understand AI.

        Virginia’s only hope is that the Feds will protect us. And if that doesn’t scare you, nothing will.

        1. I agree that it was a mistake to eliminate the Secretary of Technology position. Under Gilmore, Warner and Kaine, the Sec. Tech. did more than oversee the state IT infrastructure. Don Upson, Eugene Huang, and Aneesh Chopra acted as liaisons with the technology community and brought attention and credibility to their issues.

          I’ve got to ask though, where is the Northern Virginia Technology Council? As big as the tech industry is in NoVa, why doesn’t it exercise more clout in the General Assembly? Why aren’t the issues you talk about getting more traction?

          1. djrippert Avatar

            I used to go to NVTC meetings and they were almost totally dedicated to federal contracting. I didn’t do government work so I was pretty uninterested. I stopped going. I have no idea of what kind of outreach NVTC does at the state level. I am a member of a number of smaller technology associations in the DC Metropolitan area. Small, informal. Mostly dinners and discussion. We talk a lot about technology, trends, business environment, DC area startups, etc. We do discuss regulation, business opportunity and societal impacts of technology. However, we have never sought out politicians or government officials for a discussion. Who would we even contact? My delegate, Kathleen Murphy? My senator, Barbara Favola? Really, seriously? Maybe branch out and call Dick Saslaw or Tommy Norment? Lol. I know Tommy can relate to the risks of hacking … especially Ashley Madison. Maybe we’ll call him.

            No, the person we would have contacted would be the Virginia Secretary of Technology or somebody in his or her organization. But Governor Blackface decided that position was a waste of time. I mean when y’all are settin’ on your veranda over on the Eastern Shore family plantation … err … uh …. I mean farm … sippin’ a bourbon and branch water how can you trouble your mind with technology?

Leave a Reply