20150111

Inequality, Automation and the Political Class

The president focused on income inequality in his State of the Union address to Congress. And it is a safe bet that income inequality will be a major Democratic campaign theme headed into the midterm elections, as Democrats struggle to increase base Democratic turnout in a non-presidential election cycle.

Economic inequality strikes a chord in contemporary America. A recent NBC News/Wall Street Journal poll found that 45 percent of American adults believe that "reducing income inequality between the rich and poor" is an "absolute priority for this year." And, more importantly, Hollywood has clearly latched on to this trend with films such the science fiction action film Elysium highlighting the issue with a futuristic movie critiquing the gulf between a cognitive, off-world elite and a poor and crowded earth below. In this cinematic dystopia the elite living on Elysium need very little from the writhing masses back on earth. Capital, replete with robots and advanced technology has decoupled from labor.

The angst is palpable. As just one example, a CBS News survey from January found that 53 percent of Americans believe the next generation will be worse off.

A Democratic focus on inequality in the midterm elections is clearly a defensive political maneuver in a difficult political year and likely good politics for them. Is it politically expedient to attempt another pivot away from voter angst over the Affordable Care Act? Yes. Democratic campaigns need to change the subject and boost turnout in order to hold on to control of the Senate.

But, from a policy perspective, focusing on the end state of income inequality misses the wider context of exponential technological progress, of robots and algorithms automating away many of the jobs Americans hold today. It is as though our political discourse is frozen in the past while our technology races forward.

This is not surprising. Both political parties are essentially backward looking as they defend constituencies that have grown up over time within their respective political ecosystems.

But, robotics and software companies are not incentivized to preserve mid-20th century political constituencies. And this is the problem with nearly all of the political rhetoric we hear today. Leaders in both parties were socialized in America's industrial era and are comfortable with a 20th century paradigm for America. The 20th century is familiar and safe intellectual terrain for them. America's political class would prefer to keep the discourse safely symbolic, put the vote buying on the federal credit card and retire into a comfortable lobbying or association post. In sterile business-speak, our time horizons are misaligned.

But reality intrudes.

American manufacturing is racing ahead, but American manufacturing jobs are not. In 1953, 32 percent of American workers were employed in manufacturing. Only 9 percent work in manufacturing today.

Within the context of public opinion, it helps to go back to the 1980s. In January 1981 the Roper Organization asked 2,000 Americans if they had "read or heard anything about factories using robots instead of people on assembly lines." Forty-three percent of Americans had already heard about automation. By November 1989 Gallup asked 1234 American adults if "most assembly line workers will be replaced by robots." Fifty-two percent of Americans expected robots to replace most assembly line workers by the year 2000. Although US manufacturing jobs appear to have peaked in the late 1970s, automation has not yet eliminated "most assembly line" jobs. But, significant investments in robotics continue on a global scale. In 2011, for example, the Chinese tech manufacturing titan, Foxconn, announced plans to purchase and embed one million robots across its operations. And, as Liu Kun, a Foxconn spokesperson told China Daily "We have canceled hiring entry-level workers, a decision that is partly associated with our efforts in production automation."

Closer to home we may see a renaissance in American manufacturing due to advantageous energy costs, automation and lower transportation costs. The good news is that American manufacturing may come back. The bad news is that the jobs won't. As Nicholas Negroponte, founder of the MIT Media Lab and One Laptop per Child told an annual conference of futurists last summer, "whatever the future of jobs may be, the future is unequivocally not in manufacturing." This is all summed up in a wonderful textile country joke you're likely to hear much more often. The joke states that a 21st century textile mill requires only a man and a dog. The man is there to feed the dog. The dog is there to keep the man away from the robots.

At some point both Republican and Democratic leaders will need to address the changing nature of the economy in an honest fashion. And, this is not limited to manufacturing. Many rote, white-collar jobs are threatened by software automation. Unfortunately for America's political class, the facts are not friendly.

Three books define the automation challenge clearly:

Race Against the Machine
Second Machine Age
Average Is Over

Each explains how automation is driving 21st century "technological unemployment," where blue collar labor is replaced by robotics and any white collar labor that relies on repeated, linear tasks is replaced by algorithms.

Analyzing how software eats 20th century jobs is instructive. For example, a startup in Evanston, Ill., called Narrative Science has found a way for machine intelligence to write basic sports reporting. This is not a fantastical science fiction plot. The New York Times reported on Narrative Science's work in 2011. ElaCarte, a Silicon Valley company, has created a tablet computer for restaurants, called the Presto, which allows patrons to order food from a device, taking the waiter out of the equation. Applebee's will deploy 100,000 of these devices this year. Most large retail establishments will use RFID or touch screens to eliminate cashiers over the next decade. All of this has been well-documented and should surprise no one.

Truck drivers and taxi drivers are clearly headed for the history books when self-driving vehicles begin to dominate short and long hauls. In a telling example, one of my 2010 classmates at the University of Houston's Strategic Foresight program was a trial lawyer specialized in representing the victims of truck-car accidents. He was ahead of his time. Nevada didn't make autonomous driving legal until 2012. But, he realized that automated trucking would complicate future lawsuits. Who would be held liable with no driver at the wheel? That may not be clear, but the fate of truckers, Teamsters and taxi drivers appears much clearer.

While the automation of blue-collar, manufacturing work has been well documented, the automation of white collar and service sector work has not been as well covered. But, at least some opinion data exists. A 2008 NORC, University of Chicago survey found that 16 percent of American workers had "heard of persons in your firm having their jobs replaced by computers, computerized equipment or other forms of automation in the past three years." More troubling is a recent Oxford University study paper titled "The Future of Employment: How Susceptible Are Jobs to Computerisation?" This report, published in September 2013, will not be heavily referenced by America's current political class. After analyzing 702 occupations, the authors estimate that "about 47 percent of total US employment is at risk." The report even lists the probability of "computerization" for each occupation. The occupations most at risk are: telemarketers, title examiners, hand sewers, mathematical technicians, insurance underwriters, watch repairers, cargo and freight agents, and tax preparers.

But, just as telling is the list of least at-risk occupations. These are: recreational therapist, mechanic supervisors, emergency management directors, mental health and substance abuse social workers, audiologists, occupational therapists, orthotists and prosthetists, healthcare social workers, oral and maxillofacial surgeons, fire fighter supervisors, dietitians, lodging managers, choreographers, sales engineers, physicians and surgeons, instructional coordinators, psychologists, police supervisors, dentists and elementary school teachers. Ten of the 20 least automatable careers are in health care, largely serving retiring baby boomers. All are jobs requiring human interaction and strong EQ (emotional quotient).

Of course, this is a list of relatively safe jobs that exist today. The more interesting line of inquiry is around the jobs that will exist tomorrow and the economy that will support them. In this regard Robin Hanson and Tyler Cowen, both in George Mason University's Economics Department, are well out in front. Generally speaking, technology eliminates some jobs and creates the growth platform for many others. But the transition, when abrupt, can be difficult. And, at present, best estimates of future jobs fall in several categories that we could define as either jobs that work with machines or jobs that machines simply cannot do. Unfortunately, the former appear to require very high IQ and the latter appear to require very high EQ.

Simply put, in order to prosper in the 21st century workers will need to either become very adept at working with machines and computer code or find niches in which robots and algorithms cannot replace "the human touch." Unfortunately for many workers, this requires a significant hard skill upgrade or an enhanced EQ.

And, the blunt truth is that the economy of the future, much like the movie Elysium, may not need many low-skill Americans. The best writing on this is from James H. Irvine and Sandra Schwarzbach at the Naval Air Warfare Center in China Lake. Their article, titled "New Technologies and the World Ahead: The Top 20 Plus 5" takes a serious look at emerging technology and social stratification. As they note:

"This introduction of robotic labor will replace one-third to one-half of 'pick-and-place operation' human labor in some categories of the industrial and service sectors. Human language interface automation may cut service sector labor requirements by between one-fourth and one-third in some categories. This precipitous fall in low-end labor will probably occur in about a five to seven year period."
This is not the happy talk you'll hear on the campaign trail. Instead, it is rapid, disruptive economic change. A technological and organizational jump is made. Those that can make the jump enjoy massive, first mover advantage. Those that cannot, are left playing catch up. In evolutionary terms we might think of this as "punctuated equilibrium." In the defense community the term is "revolution in military affairs" and describes how those able to master the jump to a new level (like chariots, iron, gunpowder, rifling, blitzkrieg, air power, atomic weapons, precision munitions, cyber, etc.) enjoy strategic advantage over their adversaries. This is all well and good for the winners, but in a society where many will struggle with economic disruption, the question quickly becomes how we help our country through what could be a very rough transition.

Most business leaders know this abrupt transition is coming. According to PwC's annual CEO survey, 56 percent of American CEOs and 47 percent of CEOs globally are somewhat or very concerned about "the speed of technological change" as a threat to their growth prospects.

A more entertaining analysis of the economic dislocation brought on by rapid technological progress is Cory Doctorow's book Makers. The story, set in the near future, follows the lives of a small group of highly creative, everyman entrepreneurs. But the backdrop is a world in which manufacturing jobs have vanished, many white collar jobs have been algorithmed away and America is littered with abandoned shopping plazas that could not adjust to e-commerce and 3D printing. The American people may instinctively sense this. When asked in December 2013 if they were optimistic or pessimistic about "the opportunity for most people to achieve the American dream," a majority (54 percent) told a National Opinion Research Center poll that they were pessimistic.

Obviously, this kind of wrenching change presents a significant challenge to America and every other society.

But, unfortunately, the challenge is much greater than upskilling the young and the working, because most western countries (and many in the Pacific Rim) are aging rapidly and will need to spend scarce government resources on elder care. This is made all the more difficult by their generally high national debt levels. Taken together it means that at the exact time that nation states should be heavily investing in training and upskilling citizens, they can't because elder care and debt service have constrained their options.

So, what do we do? Looking forward in time, nation states like America have three basic response options to the challenge of automation and exponential technology.

In option one, resource constrained and with an aging population, nation states play at the margins and hope that the market, society and individuals quickly adjust, upskill and (to paraphrase Erik Brynjolsson and Andrew McAfee) "race with the machine." This is a classic "muddle through" scenario.

Free market enthusiasts will rightly point to Sal Khan's Khan Academy, initiatives like Code Academy and technologies like MOOCs (massive open online courses) as bottom up, entrepreneurial responses to the challenge. And they may be right that resources will be better deployed and innovation more quickly dispersed if the market is left unfettered. A free market response would likely build an alternative learning and credentialing system with skills certifications built alongside the formal education system. A parallel, commercial education sector focusing on applied skills is already under progress. Mozilla Open Badges are just one example of this.

But, free market enthusiasts miss the fact that American policy has always supported the country as it moved from one economy to another. This explains our history of expansionist land policies, support for the railroads, protective tariffs during our industrial revolution, the GI Bill and government R&D spending.

In option two, nation states focus on redistributing downstream wealth either through piecemeal support programs or through a guaranteed basic income. Proponents reason that so much wealth will be thrown off from advances in automation and the GRINTechs (Genetics, Robotics, Internet and Nanotech) that society will be able to afford to guarantee citizens a basic income. A group called Generation Basic Income is pushing for this in Switzerland today with a plan to guarantee every citizen a yearly income of 30,000 Swiss francs (around $33,000 USD). The Swiss system allows for direct democracy through referendum and Generation Basic Income has gained the signatures needed to put their idea on the ballot. This is a very old idea we can expect to see more of, despite the moral hazards and siphoning of resources away from business creation and investment.

We may be seeing a de facto version of this now in the United States, as poorer Americans piece together food stamps, the earned income tax credit and social security disability. For example, in December 2001 18,746,286 Americans used food stamps. In November 2103 that number was 47,033,135, a little more than the 2013 population estimate for Spain. The US is also experiencing similar surges in disability payments.

In option three, nation states meet this great challenge with a logical and disciplined plan to help their citizens leap ahead with the advances in technology. Here the United States should analyze the strategies of other nations, and study the United Kingdom's response. This is because England, under Education Secretary Michael Gove is mandating that computer programming be taught to all children beginning with their entry into elementary school and continuing through what we call high school in four stages. This bold, clear-eyed thinking is an approach that America should emulate. We can expect smaller, more agile and tech savvy nations like Singapore and the Nordic countries to follow suit. But, will America? Within the American context, we should look to Silicon Valley, where our best and brightest have clustered, for inspiration. We need to give students today the skills that they need to compete in the 21st Century. But, we also need to reskill and upskill today's workers. And only information technology can provide the scale for us to do this.

As a large country the United States has many power centers, but three have outsized influence -- Washington, New York City and Silicon Valley. As my colleague Jamaal Mobley is fond of saying, so much of contemporary business, innovation, policy and politics are driven by the interplay between The Hill, The Street and The Valley. Given our extreme challenge in preparing America for the next great wave of technological and economic change, The Hill and The Valley should spend far more time together.

No comments: