Automation: An Inconvenient Truth

On the 2016 campaign trail, there was a lot of anti-trade rhetoric. Donald Trump called the Trans-Pacific Partnership (TPP), a trade deal that the Obama administration had negotiated for five years, a “disaster” and “rape” of the American people. Bernie Sanders blasted “unfettered free trade deals that outsource millions of jobs and shut down tens of thousands of factories.” Even Hillary Clinton, who as Secretary of State called the TPP the “gold standard in trade,” turned against the agreement.

There are many valid critiques of free trade agreements (e.g. investor-state dispute settlement clauses, insufficient environmental and labor standards, a lack of transparency). I share some of these concerns and so was ambivalent about the TPP. However, Sanders’ and Trump’s narrative that free trade deals decimated the American manufacturing sector is simply not true. The United States did lose about 5.6 million manufacturing jobs between 2000 and 2010. However, according to a study by the Center for Business and Economic Research at Ball State University, “85 percent of those job losses are actually attributable to technological change (largely automation) rather than international trade.” Sanders and Trump both missed the larger picture. Automation, not free trade, has had (and will continue to have) a more disruptive effect on the American labor market.

Oxford Dictionaries Online defines automation as “the use of largely automatic equipment in a system of manufacturing or other production process.” Due to automation, companies can carry out tasks more efficiently and produce larger profits. However, this also comes with drawbacks, such as the displacement of workers.

In November 2015, the management consulting firm McKinsey wrote a report on the near-term effects of automation in the United States. The findings are startling. As many as 45 percent of the activities individuals are paid to perform can “be automated by adapting currently demonstrated technologies.” If the technologies that process and understand natural language “were to reach the median level of human performance”, an additional 13 percent of work activities could be automated. To be clear, fewer than 5 percent of occupations can be entirely automated using current technology. However, the vast majority of occupations will change, necessitating “a transformation of business practices.”

Later, in July 2016, McKinsey issued a second report further detailing automation’s effect on jobs. According to the report, 51 percent of tasks are “highly susceptible” to automation (e.g. data collection, data processing, predictable physical work). 28 percent of tasks are “less susceptible” (e.g. stakeholder interactions, unpredictable physical work) and 27 percent is “least susceptible” (e.g. managing others, applying expertise). This demonstrates how some Americans, especially less-educated workers, will be disproportionately affected by automation.

A harder question to answer is whether there will be a net decrease in jobs. After all, just as automation destroys jobs, it creates new ones (e.g. robot monitoring professionals, data scientists, automation specialists). Forrester, a market research company, is pessimistic. The company estimated that 16 percent of US jobs will be replaced by robots by 2025. Meanwhile, the equivalent of 9 percent will be created. Meanwhile, other projections are less pessimistic. Another question is whether automation is having a negative effect on wages. In 2014, the University of Chicago asked a panel of leading economists this question. About 33 percent said “technology was a central reason that median wages had been stagnant over the past decade”, 20 percent disagreed, and 29 percent were unsure.

One of the biggest challenges in the next decade will be about responding to unemployment and underemployment resulting from automation. In December 2016, the Obama administration proposed a number of policy solutions. First, the United States should invest in and develop artificial intelligence (e.g. cyber defense). Obama believes that this will result in “productivity growth” and help the United States “stay on the cutting edge of innovation.” Second, Americans need to be trained for the jobs of the future (e.g. expanding the availability of job-driven training). Third, the federal government should “aid workers in the transition and empower workers to ensure broadly shared growth” (e.g. modernizing the social safety net).

Some Silicon Valley executives think that the Obama plan isn’t radical enough. Perhaps if automation was less disruptive to the labor market, then these proposals would be sufficient to help the unemployed and underemployed; but sadly, the rate of change is too fast, they argue. Instead, the federal government should implement a universal basic income (UBI). A UBI, put simply, is a “fixed amount, at a level sufficient for subsistence, given by the state to all its citizens, regardless of income or work status.” Proponents of UBI are split on how much to give per citizen, but it would probably be about $10,000. This would be very expensive and cost about $3.2 trillion. Side note: Hillary Clinton almost ran for president in favor of a universal basic income, which would be funded by carbon and financial transaction taxes. She abandoned the plan because it cost a lot of money and would be too hard to explain to Americans.

Just like the Obama administration’s plan, UBI has major critics. For example, former vice president Joe Biden is opposed, arguing that the United States needs to “build a future that puts work first.”


“While I appreciate concerns from Silicon Valley executives about what their innovations may do to American incomes, I believe they’re selling American workers short. The future will not change the enduring American values that got us here. Our children and grandchildren deserve the promise we’ve had: the skills to get ahead, the chance to earn a paycheck, and a steady job that rewards hard work. […] Too often I hear people talking about how the jobs of the future require skills that our workers cannot attain. That’s simply not the case. Today, workers in manufacturing plants already rely on complex computer systems. As I learned a long time ago, it is never a good idea to bet against the American people.”


A third policy, proposed by Rep. Ro Khanna (D-CA), would expand the earned income tax credit (EITC). This proposal would cost $1.4 trillion over a decade. However, there are advantages of the EITC over a UBI. First, the earned income tax credit enjoys broad support, and so an expansion could conceivably happen under the current political climate. Liberal economists like the EITC because it “[reduces] poverty through redistribution” while libertarian economists like it because it “spurs some people – particularly single parents – to enter the workforce.” In fact, back in 2016, Barack Obama and House Speaker Paul Ryan (R-Wis.) both advocated for a modest expansion of the program, a rare case of agreement between the two political rivals. Second, there is a consensus among economists that the EITC works. One study found that even a $1000 increase in the EITC “boosts employment by 7 percent and reduces the proportion of families below the poverty line… by 9.5 percent.” Another study by the Center for American Progress estimated that even a small EITC increase could have “knock-on” societal benefits of $1.7 to $3 billion.

None of these proposals are perfect and yet they merit more discussion. A good place to start would be the 2018 congressional midterms and the 2020 presidential election. Politicians of both parties need to articulate a clear plan on how to deal with automation. If we don’t act, the consequences could be devastating. In such a scenario, Americans will struggle to adapt to a changing economy. For the sake of American families, innovation, and the health of the overall economy, bold leadership is needed right now.