How Technology Affects Employment

Introduction

Stunning technological advances in robotics and artificial intelligence (AI) are being reported virtually on a daily basis, from the versatile mobile robots of Boston Dynamics to autonomous vehicles, vessels, and drones, to 3D-printed buildings and new breakthroughs in machine learning from firms in Silicon Valley and beyond.

Technological progress is a key driver of productivity, quality of life, and economic development. The current revolution, however, might displace more jobs than it creates – a fear that has fed some fantastic headlines about robots coming for our jobs.

Even if these fears aren’t new, things really might be different this time. In a pair of installments, we’ll identify key issues and analyze several different economic forecasts to illustrate the breadth of opinions on this topic. We have our own views of course, but here we’re just interested in presenting information.

(Note that all of the charts embedded in this article are supported by datasets that we encourage you to work with.)

The big picture

A number of international agencies have recently flagged issues relating to the future of employment, and the consequences of automation and deindustrialization (ADB; ILO; IMF; UNCTAD; UNDP; UNIDO; World Bank; and OECD). The International Labor Organization (ILO) recently launched a Global Commission on the Future of Work.1  Many have concluded that trade, offshoring, and immigration don’t appear to be the biggest threat to jobs in developed countries. Automation, it turns out, is much more important.2

These aren’t new fears, of course. Over the last two centuries, there have been periodic warnings that automation and new technology would wipe out large numbers of middle-class jobs. For example, take the Luddite movement of the early 19th century, in which a group of English textile artisans protested production automation by trying to destroy some of the machines.3

In the past, however, technological innovation — including automation — has always in the end created many more new jobs than it has destroyed, raising productivity, spurring sustained increases in living standards, and shifting the balance of work and leisure.4 But people concerned about automation and employment today are quick to point out that past interactions between automation and employment don’t necessarily hold true for a future where the emergence of advanced computing power, artificial intelligence, and robotics might replace labor on a scale not previously observed.3

We’ll focus on two questions here: Why didn’t automation increase unemployment in the past, and will it be different today? In the next installment, we’ll look at the data and evaluate some strategies to combat potential job loss.

“In the British census of 1841, 22 percent of citizens registered as agricultural workers; today that number is below one percent.”

Why automation hasn’t wiped out all jobs before

The wide adoption of labor-saving technologies in the past raises a question: Why are there still so many jobs?3 Economists have proposed a few explanations.

First, economists often cite agricultural and manufacturing revolutions as vivid examples of automation’s impact on employment in the past. And indeed, if we look at these on an isolated industry-by-industry basis, the shifts seem radical. In the British census of 1841, 22 percent of citizens registered as agricultural workers. Today, that number is below one percent.1 In the United States over roughly the same period, the agriculture share of employment declined from 58 percent of total employment to 2.5 percent.4

Developed Economies

Employment by sector, labor productivity, and per capita income

We can then look forward to the second half of 20th century, when the number of workers employed in industry sectors in developed countries began to drop.

At the same time, however, total employment, labor productivity, and per capita income grew. Displaced agriculture workers got absorbed by the manufacturing sector, and later the advent of the services sector absorbed still more displaced labor.

Clearly, then, labor-saving advances didn’t decrease total employment in the past.

Developing Economies

Employment by sector, labor productivity, and per capita income

Experts break the industrial revolution into five stages: the first (steam) in the late 1700s; the second (iron) in the mid-1800s; the third (steel and electricity) in the 1890s and 1900s; the fourth (electromagnetic and chemical) in the 1950s and 1960s; and now the current revolution, information and communications technology.5

This raises a natural question: Why are there still so many jobs? Economists have proposed a few explanations.

1. Production consists of many stages, and automating one or several of those stages increases the value of human work in other stages. (You need less labor input to produce an equal or even greater amount of output.) Automation increases worker productivity, so real wages rise. Workers who make more money increase demand for goods and services, which also creates new jobs.3 This is why productivity growth should align with growth of aggregate demand.

2. Growth in productivity reduces production costs. These savings are passed on to consumers in the form of lower prices, which in turn stimulates new demand. A population will therefore consume more products and services, not less or the same amount.3

3. Technological progress creates new industries that absorb labor displaced by new technologies.

4. Humans can adapt to these changes, acquiring new skills through education and training.6

One more theory comes from Positive Theory of Capital (1889), by Austrian economist Eugen von Böhm-Bawerk. According to Böhm-Bawerk, every new technology and labor division adds complexity to production processes. This complexity requires more production stages and more time to achieve the same outcome. In the end, then, new technologies would actually require more overall labor.

This explains why historical evidence shows that employment remains strong even in periods of dramatic change. Historical trends also show that journalists and even expert analysts tend to overstate the extent of machine substitution for human labor, and at the same time ignore the synergy between automation and labor that increases productivity, earnings, and demand for labor.3

However, the transition isn’t always smooth. For example, real wages stagnated for nearly 50 years during the Industrial Revolution in 19th-century England, and only picked up again at a time of substantial social policy reforms.4

“The digital revolution might displace labor at rates not observed before, and possibly not even at rates we anticipate today.”

Will this time be different?

Many experts insist it will. Powerful digital technologies (such as artificial intelligence, advanced robotics, mobile platforms, machine learning, neural networks, cloud computing and the internet of things) are rapidly translating to all sectors of the economy.7 Unlike the previous four revolutions, the digital revolution will substitute not just labor, but for human cognition as well.8 Finally, largely because of the pace of advancement, it will become difficult for workers (including high-skilled workers) to adopt by means of education and training.

Another concern is the rate of innovation and automation adoption. In the mid-2000s many experts suggested driving could not be automated, but in 2009 Google announced the launch of its self-driving cars program. This demonstrates that the rate of innovation and adoption in the coming decades might be exponentially faster compared to the technological waves in the past. The digital revolution, then, might displace labor at rates not observed before, and possibly not even at rates we anticipate today.

In our next blog post, we’ll analyze estimates on how many jobs might be vulnerable and which are most at risk. We’ll also evaluate a few strategies to combat potential job loss.

Looking to work directly with the data? Explore our curated dashboards on the topic in Knoema:

References

  1. Lukas Schlogl and Andy Sumner, 2018 “The Rise of the Robot Reserve Army: Automation and the Future of Economic Development, Work, and Wages in Developing Countries.” CGD Working Paper 487. Washington, DC: Center for Global Development. Retrieved from Link
  2. The New York Times (2016). “The long-term job killer is not China. It’s Automation” Retrieved from Link
  3. Autor, D.H., 2015 “Why Are There Still So Many Jobs? The History and Future of Workplace Automation” Journal of Economic Perspectives. Volume 29, Number 3. Summer 2015 p. 3–30
  1. MGI, 2017 Jobs lost, jobs gained: Workforce transitions in a time of automation. Link
  1. Robert D. Atkinson and John Wu (May, 2017). “False Alarmism: Technological Disruption and the U.S. Labor Market, 1850–2015” Information Technology and Innovation Foundation. Working Series. Link
  1. Carl Benedikt Frey, Michael A. Osborne, September 2013. “The Future of Employment”. Published by the Oxford Martin Programme on Technology and Employment. Working Paper. Link
  2. Oxford Economics and CISCO. (December 2017). “The A.I. Paradox. How Robots Will Make Work More Human”.
  1. Arntz, M., Gregory, T., & Zierahn, U. (2016). The risk of automation for jobs in OECD countries: A comparative analysis. OECD Social, Employment and Migration Working Papers, 2(189), 47–54.

Accuracy In Macroeconomic Forecasting

Introduction Stunning technological advances in robotics and artificial intelligence (AI) are being reported virtually on a daily basis, from the versatile mobile robots of Boston Dynamics to autonomous vehicles, vessels, and drones, to 3D-printed buildings and new breakthroughs in machine learning from firms in Silicon Valley and beyond. Technological progress is a key driver of …

Measuring Technological Impact On Employment

Introduction Stunning technological advances in robotics and artificial intelligence (AI) are being reported virtually on a daily basis, from the versatile mobile robots of Boston Dynamics to autonomous vehicles, vessels, and drones, to 3D-printed buildings and new breakthroughs in machine learning from firms in Silicon Valley and beyond. Technological progress is a key driver of …