Rise of the Robots: Technology and the Threat of a Jobless Future. Martin Ford. 2015.
The just right “Goldilocks” period of the US economy after World War II ended in the 1970s when growth slowed and capital began to take a larger share of gains than labor. Consequently, inequality soared from 1973 to 2013, when productivity gains of 107% were almost all retained by business owners and investors while production workers’ wages fell by 13%. This era marked a fundamental shift in the relationship between workers and machines made possible by enormous advances in digital technology. This shift is just beginning. Due to accelerating advances in computers and artificial intelligence, machines are now poised to move from merely enhancing the productivities of workers to actually becoming workers themselves. Resultant massive unemployment could cause the collapse of the feedback loop between productivity, rising wages, and increasing consumer spending that supports our economy. Thus the author asks if accelerating technology could disrupt our entire system to the point where a fundamental restructuring may be required if prosperity is to continue?
In the Goldilocks years after World War II, technological progress markedly increased productivity but did not adversely affect labor because it occurred in areas like mechanical, chemical, and aerospace engineering that allowed workers to become more valuable and command higher wages. This pattern was transformed in the 1980s when the increasing roles of information technology and automation displaced or deskilled workers and led to decoupling of the historical correlation between rising productivity and rising incomes. (See fig. 2.1) This resulted in the following seven deadly trends:
- Stagnant wages—from 1973 to 2013 weekly wages fell from $767 to $664 in 2013 dollars, and the increase of median household annual income from $50,000 to only $61,000 (compared to $25,000 to $50,000 from 1949 to 1973) was due to women entering the workforce. Since the 1960s, the inflation-adjusted minimum wage has actually fallen by 12%. 2. Income share decreasing for labor and increasing for corporations (See fig. 2.3). 3. Declining labor force participation—consistent decline for men from 86% to 70% from 1950 to 2013 but increased combined rate for men and newly arriving women, which peaked at 67% in 2000, then declined. 4. Diminishing job creation, lengthening jobless recoveries, and soaring long-term unemployment (See fig 2.6). 1998 to 2013 saw increases of US output by 42% and population by 40 million with no increase in hours of employment. 5. Soaring inequality—95% of income gains were by the top 1% from 2009 to 2012, with the obvious risk of political capture by financial elites. 6. Declining incomes and underemployment for recent college graduates—for those with only bachelor’s degrees, income fell by 13% from 2000 to 2010, and only half found jobs utilizing their education. 7. Job market polarization and part-time jobs—solid middle income jobs destroyed are mostly replaced by lower-wage service jobs and a smaller number of high skill jobs.
Practitioners of economics and finance point to earlier overstated warnings about technology and tend to dismiss anyone who argues that this time might be different. They point out that previously new jobs were created as old jobs were destroyed, as with the transition from horses and buggies to motor vehicles, and they offer globalization, financialization, and politics as alternate explanations for the deteriorating status of workers. The author counters that there is a fundamental difference between the earlier innovations that displaced workers from older industries to newer ones with similar jobs and the more recent rapidly accelerating digital innovations that are just beginning to eliminate enormous numbers of workers at all skill levels without generating replacement jobs. Occupations amounting to nearly half of US total employment may be vulnerable to automation within the next one to two decades, according to a 2013 Oxford University study.
With respect to globalization, the fraction of US workers in manufacturing has fallen steadily due to automation since the 1950s, long before NAFTA in 1990 and the rise of China in the 2000s (See fig. 2.8). Also, in 2011, 82% of goods and services purchased in the US were produced in the US, and only 3% were imported from China. With respect to financialization, its share of GDP has increased from 2.8% to 8.7% since 1950; its share of corporate profits has increased from 13% in 1978-1997 to 30% in 1998-2007, and its employees have 70% higher salaries—with much of this activity geared toward rent-seeking. However, this may be viewed as a ramification of the growth of information technology. Powerful computers now account for nearly two-thirds of stock trading and are essential for innovations like exotic derivatives and CDOs. With respect to politics, conservative business interests have successfully conducted a sustained and organized assault on forces that counter inequality, including regulations, high marginal tax rates, and private sector unions. If a nation fails to implement policies designed to mitigate the changes brought on by advancing technology, should we label that as a problem caused by technology or politics?
Information technology has evolved to a general purpose technology that is simultaneously transforming many sectors of the economy. Automation has replaced workers with devices like ATMs and self-service check-out systems and has the potential to replace 50% of fast food workers. On-line retailers like Amazon and Netflix, have moved jobs to the internet and warehouses where they are much more easily automated. Retail employment is threatened by fully automated vending systems like Redbox movie-rental kiosks that now serve Chicago with seven employees rather than the seven for each of dozens of previous Blockbuster stores. Expansion of business on the internet also increases inequality because the income from on-line activity nearly always follows a winner-take-all distribution.
Commercial software already produces automated articles in sports, business, and politics and has the potential to produce 90% of news articles within fifteen years. Commercial software is available to almost completely manage most aspects of the execution of business projects. IBM’s Deep Blue computer used AI, machine learning, and enormous capacity to defeat world chess champion Gary Kasparov in the 1990s, and its successor Watson accomplished the more difficult task of defeating the Jeopardy champions in 2011. The migration of these capabilities into the cloud is almost certain to be a powerful driver of white-collar automation. Only the health care and education industries are relatively resistant to these changes, with the result that their relative costs are accelerating. (An excellent brief discussion is provided for why health care is simply not comparable to other markets.)
In the near future, rapidly increasing capacity in fields like big data, artificial intelligence (AI), machine learning, artificial neural networks, and 3D printing is poised to replace human employment as never before. Advancing AI and machine learning will eliminate many white collar positions, such as in middle management and the professions, even in the fields of computer science and engineering. As a result, even increasing education, which is advocated for students as protection against automation, is experiencing diminishing returns. With up to 50% of jobs at risk to automation and up to 25% at risk to offshoring (presumably with some overlap), the potential impact on employment is staggering.
These examples point to the theoretical catastrophic endpoint of relentless progression toward automation—at least in the absence of policies designed to adapt to the situation. Members of the non-elite general public are already well into this trajectory due to rising inequality, stagnant wages, a decreasing education premium, and loss of worker power from globalization and automation. Their ability to face the coming crisis is further jeopardized by the rising political power of those elites who have promoted policies to limit business responsibility and to lower taxes by weakening the safety net. As a result of the transition from defined benefit pensions to often under-funded 401K plans, 50% of American households aged 65-69 have retirement balances of $5,000 or less. Their plight could be compounded in coming decades by another severe financial crisis or by dislocations from global warming, since the same elites are determined to prevent or reverse policies designed to lessen those possibilities.
Unfortunately for elites, future decades will likely be precarious for them as well, even for the portion who are pushing for oligarchy (see the main post in darkcashnet.net for their identities). The demand required by the economic system to generate their riches comes mostly from consumer spending—more than two-thirds in the US and more than 60% in other developed countries. If jobs are automated away, consumers will lack the purchasing power necessary to drive this demand, and a fundamental restructuring of our economic rules will be required. If this occurs, some form of direct redistribution of purchasing power will become essential if economic growth is to continue.
The author, a Silicon Valley CEO, believes the most likely solution to restore purchasing power is likely to be some form of basic income guarantee, such as by returns from a centrally managed sovereign wealth fund. For example, a $10,000 annual basic income would require new revenue of about $1 trillion after adjustment for reducing or eliminating numerous antipoverty programs. Since ever more taxable income is rising to the very top, taxation should be restructured to mirror the income distribution. In addition to concern for the health of an economy with extreme inequality, there is the moral question of whether a tiny elite should be able to, in effect, capture ownership of society’s accumulated technological capital. Today’s innovations do not really compare to the groundbreaking work of pioneers like Alan Turing and John von Neumann. The computer technology that makes automation possible exists in some measure because millions of middle-class taxpayers supported federal funding for basic research in the decades following World War II. Hence, the general public has a legitimate claim on the returns of decades of enormous incremental development that it supported, particularly when the dysfunctional nature of the alternative is considered.