Fact: In the past ten years, there has been a 119% increase in EBIT but only a 31% increase in the number of employees. Hourly earnings in US manufacturing are up only 22%. The digital age has as a result decreased labor’s share of GDP.
The historical observation that shares in national income accrue to capital and labor in a consistent manner has broken down. Increasing demand for low-skilled workers and a surge in productivity helped boost employment and wage growth over the course of the 19th century. The switch to electricity and the rise of multinational corporations shifted the demand to higher-skilled workers into mass production.
Innovation in the last few decades benefits the few rather than the many. The rise in inequality is indisputable and can be to some degree explained by skilled based technological change. Employment in middle-skill manufacturing and clerical occupations experienced a secular decline in the 1970s due to the impact of computers. The Computer Revolution in the 1980s favored workers with a college education. Work involving complex but manual tasks like cleaning or driving trucks became more plentiful. In general, low skill and low income service jobs experienced significant job growth. In the 1990s this situation changed. Computers and industrial robots together have substituted for a much wider class of routine work, reallocating workers to manual service occupations. As a result, much of the job growth in the US in recent decades has been in high-skill occupations where automation cannot be accomplished and to low-skill occupations where automation is too expense. The result is a polarized labor market. High-skill workers are taking on jobs traditionally performed by low-skill workers, pushing low-skill wages down even more, and to some degree out of the work force.
There is a long-term decline in routine occupations that occurs as these jobs are lost during recessions. As labor costs go up, companies will respond by increasingly turning to automation in order to keep costs down and profits up, leading to a global gap between profit growth and job creation. Future recoveries will likely be jobless as technological advancements allow distressed companies to shed middle-income jobs in favor of automation—a phenomenon occurring across industries like manufacturing, wholesale and retail trade, and financial services.
The coming digital age will be more disruptive than previous revolutions as it is happening faster and is changing the way we live and work. Robotic technology, in the wide sense, makes it possible to substitute labor for capital, so productivity improves while wages do not. The most recent and comprehensive survey of AI experts concludes we will probably reach overall human ability in machine intelligence by 2040-50. Skilled workers have benefited from the digital age as producers. Technological levels that enable machines to replicate even highly skilled work imply an acceleration of the trend to greater income inequality.
The top 1% income share has more than doubled over the past 30 years, but this is not all attributable to technology and globalization. The 1% income share has gone up much more in the Anglo-Saxon countries compared to France, Germany, and Japan. That said, digital technology takes the form of capital that substitutes labor and as such income inequality is likely to continue to surge in all industrialized countries.
The booming stock market is also a part of the inequality story. Productivity is a key driver of long-term economic growth. Growth is the key driver of long-term corporate earnings. Earnings growth is the key driver of long term stock market returns. Equity investors have been the clear beneficiaries of the economic gains from automation. Machines don’t take sick leave and machine don’t require holidays.
As robots and AI substitute for labor, they also reduce the demand for capital. Innovators and entrepreneurs, not workers or investors will be the main beneficiaries. For example, the operating income per employee at Google is 6x that of IBM and about 12x that of General Motors. Slowing global population growth and the rate of capital absorbing innovation will cause net savings at full employment to grow and net investment to fall.
AI, robotization, and related technologies require significant capital investment up-front, accompanied by fewer workers needed to build out the new capital, relative to technologies of the past. McKinsey estimates that the number of industrial robots installed globally by 2025 will rise to 25 million, up 15 million from the current level, implying 25% to 30% average annual growth in robot sales, which is considerably higher than the average growth rate over the last two decades. This would require considering investments totaling from about $900 billion to $1.2 trillion. Sounds steep, but this fixed investment will automate virtually all of global manufacturing capacity.
Workers have benefited from automation as consumers, but there is a dark side to this benefit. But the surge in income inequality has been accompanied by an increase in borrowing, which has helped to maintain consumption levels. This is unsustainable. Easy credit means borrowing from the future to boost unemployment despite stagnating incomes. The result is that high-income individuals save more, more borrowing among low-income workers, and low consumption inequality compared to income inequality. The capital share of income will only increase further, benefitting those with a lower propensity to consume. This explains the savings glut and slower growth characterized as “secular stagnation”.
Savings and investment decisions drive the gap between potential and actual GDP. Macro policy depends on whether actual GDP keeps pace with potential GDP. As economies become more digitized, potential GDP running above actual means that slack and overabundance is abundant causing downward pressure on prices, including wages. Investment opportunities will continue to taper off, a rising share income to profit, and a persistent savings glut. If output gaps are widening and price pressures lean to deflation, policy makers will address it. Fiscal and tax policy will have to address the collapse in income for a swath of workers replaced by robots. QE to boost effective demand will be the standard.
The Luddite riots between 1811 and 1816 were meant to stem mechanization, but the Crown and the guilds did not have the political influence to halt creative destruction. This time it will be no different.