Shaping structural change in an era of new technology
The founder of the Information Technology and Innovation Foundation argues threats of a fourth industrial revolution are overstated. Will change be more modest than predicted?
In recent years an idea has spread, repeated at countless conferences around the world and in op-eds, tweets, books and reports, that we are in a fourth industrial revolution, more transformative than any change in human history, that will, among other negative effects, lead to massive job losses and unemployment. Indeed, it seems you cannot attend Davos, a G20 summit, or a TED talk – one organised by the media organisation TED (Technology, Entertainment, Design) – without hearing the warnings. Robots and artificial intelligence (AI) are coming to put most of us out of work, except for a growing number of ‘gig economy’ workers who cobble together a meagre income using internet platforms dominated by large US tech companies like Uber and Airbnb.
The ‘fourth industrialists’ are wrong. This next wave of innovation will be modest, but progressive, particularly by enabling an uptick in productivity that, with the right policies, will lift incomes for workers around the world. It is therefore time to think clearly and deflate the growing techno-panic before policymakers actually implement many progress-killing ideas like taxing robots and implementing a universal basic income.
To think clearly about technological transformations in the labour market, it is important to examine two factors:
- What will be the nature and pace of the next technology wave?
- What are its likely impacts on jobs, employment relationships, income inequality, job quality, firm disruption and labour market disruption?
This essay examines these questions and concludes that policymakers should ignore techno-Cassandras and instead embrace this next wave of innovation, while at the same time ensuring that workers are well equipped to prosper from it.
We are not facing a fourth industrial revolution
Pundits use a variety of terms to refer to the supposed technological transformation that is currently under way: ‘the Second Machine Age’, ‘the Rise of the Robots’, ‘the Coming Singularity’ and others. But the term that has caught on the most is ‘the Fourth Industrial Revolution’, which was coined by K. Schwab, head of the World Economic Forum. He breathlessly writes, “We stand on the brink of a technological revolution that will fundamentally alter the way we live, work, and relate to one another. In its scale, scope, and complexity, the transformation will be unlike anything humankind has experienced before” (Schwab 2016).
If this were true, it might be cause for concern, for it suggests that history provides no guide to the present. But in fact, it is not true. First, the next innovation wave is not the fourth, it is the sixth. In Schwab’s sweeping but shallow historical telling, the first revolution of steam power was in the late 1700s and early 1800s. Then came electric power in the early 1900s. Then a few years ago digital technologies. Now the fourth is upon us.
For historians of technology such periodisation makes little sense. Those who follow on the work of Joseph Schumpeter and who study technological long waves generally agree that there have been in fact five waves to date:
- the first industrial revolution of the steam engine in the 1780s and 1790s
- the second revolution of iron in the 1840s and 1850s
- the third revolution of the 1890s and 1900s based on steel and electricity
- the fourth revolution in the 1950s and 1960s based on electromechanical and chemical technologies
- the fifth of our present era based on information technology and communications technology (Atkinson 2005).
According to this periodisation, a sixth wave will emerge, likely grounded in AI, robotics and perhaps nanotechnology and biotechnology, but not before an intervening period of relative stagnation of perhaps as long as 20 to 25 years, a period the global economy appears to be currently suffering through now.
This more accurate periodisation points to several important conclusions. First, despite all the breathless talk about us being in the midst of a fourth industrial revolution, the next technology wave is not here yet and will not be for at least a decade. This, more than any other factor, explains the slowdown in global productivity over the last decade (Atkinson 2016). The current digital technology system has reached a spot on the ‘S-curve’ where it is difficult for it to continue to drive productivity at a robust rate.
Second, there is no reason to believe that this coming technology wave will be any different in pace and magnitude than past waves. Each past wave led to improved technology in a few key areas (eg steam engines, railroads, steel, electricity, chemical processing and information technology) and these were then used by many sectors and processes. But none completely transformed all industries. Within manufacturing, for example, each wave led to important improvements, but there were still many processes that required human labour.
The next wave, grounded in AI and robotics, will be no different. While it will no doubt affect many industries and processes, many will remain largely untouched: think of fireman, pre-school teachers, massage therapists and trial lawyers. Moreover, this technology will replace some workers, as all pasts waves have done, but they will also augment others. AI, for example, will not replace doctors, but it will help them make better diagnoses and treatment decisions. This is why the Information Technology and Innovation Foundation (ITIF) estimated that only about 8% of jobs were at high risk of automation by 2024, and why the McKinsey Global Institute estimated just 5% were at such risk (Atkinson 2017; Chui, Manyika and Miremadi 2015).
In response to this argument, ‘fourth industrialists’ tell us that computer systems with powerful artificial general intelligence (AGI) are just around the corner. For them AGI will eclipse the full range of human ability – not only in routine manual or cognitive tasks, but also where more complex actions or decisionmaking are involved. But there is about as much chance of AGI emerging in the next century as the earth being destroyed by an asteroid. As MIT computer science professor R. Brooks (2015, 111) puts it:
The fears of runaway AI systems either conquering humans or making them irrelevant aren’t even remotely well grounded. Misled by suitcase words, people are making category errors in fungibility of capabilities – category errors comparable to seeing the rise of more efficient internal combustion engines and jumping to the conclusion that warp drives are just around the corner.
To be sure, there is progress in AI, including in machine learning, but these are still and will remain discrete capabilities, not general (recognising fraud in financial transactions, for example).
This relates to the second important issue: the pace of change from the technologies. If the wave increased economy-wide productivity by 75%, but it took 30 years to do so, this would mean a modest annual rate of growth of less than 3%, on a par with past periods in developed nations where labour force adjustment proceeded apace. But if this happens over 10 years, it surely would mean a much faster rate of dislocation. And here again, without evidence, the ‘fourth industrialists’ assert that the coming pace of change will be unprecedented.
But past long-wave transformations have taken at least 30 years to work their way through developed economies. There are three reasons for this relatively slow pace. First, new technology systems do not emerge fully formed. Early versions are less advanced than later ones. We say this with the electric motor in the early 1910s, where it took decades for improvements in power, price and quality to enable electric motors to be transformative. Going forward we will likely see this pattern in autonomous vehicles. The best (and quite expensive) current autonomy technology is at what is referred to as level 3, where drivers are still necessary for many functions. Level 5 cars that are affordable – where the human can go on a long, complicated trip asleep in the backseat – are decades away. Second, even though new technologies are better than old, old technologies are usually not completely scrapped, at least until their value is significantly depreciated. Trucking companies, for example, will not suddenly toss all their expensive semis in the junk yard. Third, not all organisations are first-adopters. Some adopt early, most adopt in the middle after the technology is de-risked, and the rest late.
So, yes, there will be a next wave of innovation, but it will not be an unprecedented tidal wave of transformation, but rather a moderate increase in innovation that will hopefully kick in by at least the mid-part of the next decade and will likely take at least 20 years to diffuse through economies, leading to an increase in economy-wide labour productivity to at best 3–4% growth per year.
Major issues of concern
Notwithstanding that the next wave of innovation will not be unprecedented, there still could be negative impacts that policymakers need to prepare for and seek to mitigate. However, there will also be benefits, something ‘fourth industrialists’ usually ignore. Most importantly, the next wave will raise productivity growth rates. European productivity has been growing at anaemic rates for years, and in the UK it has virtually ceased. Without productivity growth to create a ‘bigger pie’ there is no way for European living standards to increase, especially given that the working age to old person ratio will drop from 3.5 today to 2.2 by 2040. But this does not mean that there may not be some negative impacts from the next wave of innovation. However, most of these fears are unwarranted and the main one, job dislocation, can and should be addressed by smart policies.
Let us start with unemployment. The ‘fourth industrialists’ claim that the next wave will lead to massive job losses. Yet academic studies, historical data and logic all suggest that increased rates of productivity growth will not lead to higher unemployment (Atkinson and Wu 2017; Miller and Atkinson 2013). If anything, higher productivity growth in nations has been associated with lower rates of unemployment. The reason is simple and ignored by ‘fourth industrialists’: companies invest in process innovation (innovations to boost productivity) to cut costs, and because of competitive markets, they pass the vast share of those savings on to consumers in the form of price cuts, and some to workers in the form of higher wages. This added purchasing power is not buried; it is spent, and that spending creates new jobs. This dynamic is the same if productivity grows at 1% a year or 5%. Moreover, higher productivity growth creates a ‘rational exuberance’ where consumers and businesses feel more confident, and spend and invest more, leading to even more growth and job creation. So, the bottom line, absent ill-advised policies such as universal basic income to pay people for not working, and higher unemployment from the next technology wave will not happen (Atkinson 2016).
Even if unemployment rates will not rise, many ask whether the new jobs from the next wave will be good ones. But for two reasons this is not the right question to ask. First, the new jobs created will be largely related to how people spend their new added income, which will likely be on things like education, personal services, hotels and other lodging, entertainment, insurance, air travel, new cars and trucks, and major appliances. Some of this will create good jobs (eg education), others not so good jobs (eg personal services).
Second, rather than fret about what industries and occupations are growing and shrinking, policymakers should focus on raising productivity. That some jobs pay more than others is because they are more productive. A main reason janitors are paid less than software engineers is because the latter’s output per hour is much higher. Therefore, the most important question regarding the mix of jobs is whether the next innovation wave will raise productivity.
It will be even better if the next wave raises productivity more in lower-wage occupations. If it does, there will be relatively fewer workers employed in low-wage occupations and the wages of everyone, including the remaining low-wage workers, will increase. To see how, imagine that the next technology wave boosts productivity by 25% only for the bottom 25% of wage earners. In the US this would allow the tasks these workers currently do to be performed by just 23.4 million workers, instead of the current 31.2 million. That means 7.8 million workers freed up, and as the savings from lower prices are spent, they could be employed doing other work. Because the prices of goods and services produced by low-wage workers would fall, this spending would be distributed in the same shares as it is currently, with 12.9% going on goods and services produced by workers in the first wage quartile, 17.6% in the second, 27.2% in the third, and 42.3% in the fourth.[i] As a result, most of those 7.8 million workers would see a wage increase as they move to higher-wage jobs. So too would all other workers because the real prices of goods and services supplied by low-wage workers would now be lower.
Labour market status
Even if most people will be working, ‘fourth industrialists’ warn that an increasing share of workers will be contingent workers, doing work through platforms of American tech giants. To be sure, such ‘gig economy’ work has grown in the last decade, but much of this has been a fall-out of the Great Recession, when full-time, permanent work was scarce compared with today. So even with the growth of Uber, Airbnb and other work-sharing platforms, in 2015 only about 600,000 people were employed this way. Moreover, the share of the US workforce that is self-employed is at an all-time low of less than 7% (US Bureau of Labor Statistics 2016). There is no reason to believe that self-employment will grow in the future (see Arnold et al. this volume).
‘Fourth industrialists’ warn that the next wave of innovation will bring massive growth in inequality. There is no doubt that income inequality has grown in Europe and the US, although by considerably less than Thomas Piketty would have us believe (Rose 2014). But very little of this growth has been from occupational changes driven by technology. The Economic Policy Institute finds that inequality did not increase because jobs in middle-wage occupations were eliminated by productivity gains (Bivens and Mishel 2015). Rather, virtually all the increase was within occupations, with some individuals making winner-take-all incomes at the expense of other workers in the same occupation.
It is important to realise that this had nothing to do with technological productivity and everything to do with socio-political factors. To take an example from US pro basketball, income inequality in the National Basketball Association (NBA) did not grow because technology eliminated middle-skilled players, it grew because of political economy factors, such as the introduction of free agency that allowed the LeBron J. and Steph Currys of the world to make vastly more money than the NBA stars of the 1970s. As J. Rothwell showed in a study for the Brookings Institution, the one-percenters are largely professionals and financiers: 6% of the top 1% of earners are in the financial services industry, 7% in law, 7% are doctors, 7% work in hospitals, and 4% are dentists (Rothwell 2016). This growth in earnings inequality has nothing to do with productivity.
Not convinced, ‘fourth industrialists’ will say the future will be different, especially if the next wave of innovation impacts lower-wage occupations more than higher-wage ones. That indeed is likely to happen, as ITIF found that there was a modest (−0.38) correlation between the risk of a US job being automated and the levels of education needed for the occupation (Atkinson 2017). But this pattern of automation would actually reduce, not increase, inequality. One reason is that 40% of adult European employees report that they have higher skills than are required to perform their current job. (European Commission Skills Panorama 2014; McGowan and Andrews 2015). These workers are in jobs that require fewer skills than they possess, presumably for most of them because there are not enough high-skilled jobs in the EU economy to employ them. If the next technology wave has a larger impact on eliminating low-wage jobs, this would by definition mean that a greater share of jobs would be in middle and higher-wage employment. And many European workers now in low-wage jobs have more than enough skills to move into these jobs. More fundamentally, even with robust minimum-wage levels and tax-based redistribution measures, it is extremely difficult to raise significantly the after-tax income levels of people working in low-productivity, often low-skill-level, industries for the simple reason that wages cannot exceed the output of the worker. Automating low-wage jobs will lead to not only fewer low-wage jobs and more middle- and higher-wage jobs, but usually higher output per worker in the remaining workforce, so those workers’ wages can more easily be increased.
This positive outcome depends on relative price declines from automating low-income jobs so that demand for goods and services grows. But ‘fourth industrialists’ say there will be no price reductions because all the savings will go to the increasingly fewer owners. Owners of capital will somehow no longer have to compete on the basis of price and will be able to make exorbitant profits, immiserating the proletariat. But this scenario of a few ‘robot owners’ making ‘trillions’ while the rest of us are unemployed strains credibility. The reality is that if one ‘robot owner’ jacked up prices and made massive profits, another robot owner would lower prices to gain market share, just as this process of competition has worked since the beginning of market economies (see discussion of competition and innovation in Atkinson and Lind 2018).
While unemployment will not increase from the next wave of innovation that does not mean that there will not be modest or even significant rates of businesses disruption. Just as internet platforms today are disrupting a range of industries, including private transport, retail, lodging, and telephone and cable TV, one could imagine (and hope for) emerging technologies disrupting even more industries. For example, ‘fin-tech’ could disrupt the traditional banking industry. But progressives, more than people in other political camps, should understand that government’s role is not to protect businesses from risk; it is to protect consumers from business opposition to change. For-profit businesses, big or small, are more than happy to reap the profit upside of success, but are all too quick to run to government to protect them from the downside of competitive loss. Progressives should focus not on protecting companies from technology-based disruption, which fundamentally helps consumers, but on helping workers make transitions to new employment. Uber, Lyft and other car services are a case in point today. Too many governments want to protect incumbent taxi companies at the expense of consumers who benefit from better and cheaper car services.
Worker transition and dislocation
Of the concerns ‘fourth industrialists’ raise, the only valid and important one is how to help workers adjust to the inevitably, albeit modest, higher rates of labour market churn that will be coming. It is important to note, however, that – at least in the US – the rate of labour market churn (defined as jobs created in occupations plus jobs eliminated in other occupations) has been at an historic low over the last two decades (Atkinson and Wu 2017). But as the next wave of innovation boosts productivity that rate is sure to increase somewhat.
One proposal to address this is the introduction of a universal basic income (UBI). Under this widely touted scheme, the state would somehow take money from somewhere and write monthly cheques to all adults, whether they are working or not, poor or rich. This allegedly would establish a stable floor on which everyone would build their own brighter future. This, however, is one idea progressives should loudly decry. UBI would lead to the very thing its advocates warn us technology will bring: large-scale unemployment as the government incentivises workers to be idle instead of helping pave pathways for those displaced by technology to find success in new jobs.
To be sure, the alternative should not be a return to the Hobbesian world of the 1800s when if a worker lost his job he was on his own. Progressives need to ensure that there is temporary income support for workers who lose their jobs through no fault of their own. On this score Europe is both better and worse off than the US. It is better in that it provides laid-off workers with more income support. It is worse in that at least in some European nations workers are eligible for unemployment income support for far too long, which not only encourages them to stay out of the labour market and have their skills atrophy, but by reducing the purchasing power of employed workers (who bear higher taxes), it reduces the demand for workers, leading to higher unemployment.
The lesson for Europe and the US is to copy the Nordic countries’ flexicurity model, which ties benefits to proving that workers are either actively looking for work or are in a certified training programme. Progressives should also advocate for a system of lifelong-learning accounts similar to what France recently introduced while Emmanuel Macron was economy minister and Myriam El Khomri was labour minister. These accounts can be drawn down to pay for retraining throughout the working life (see Weber this volume). Progressives should also push to disrupt the traditional higher education system, which has become too expensive and too inflexible (Kennedy, Castro and Atkinson 2017).
These and other steps to ease transitions are important because if Europe is going to have any hope of its voters embracing change and innovation, governments needs to do more to reduce employment risk for workers. At the same time, if Europe is going to reap the benefits of the next innovation wave, the last thing progressives want to do is stoke people’s unwarranted fears that their jobs are on the fourth industrial wave chopping block from all powerful ‘Terminator-like’ robots or support completely misguided policy proposals like taxing and regulating robots to slow their adoption. While slowing innovation runs counter to progressives’ policy goals of ensuring a growing standard of living for workers, it also runs counter to their political goals. When centre-left parties have succeeded in Europe or the US they have done so when they stood for a vision of growth, widely shared, not limited growth with massive redistribution.
In short, the vision should be innovation, widely shared. And with the right policies the vision can very well become the reality.
Atkinson, R. D. (2005), The Past and Future of America’s Economy: Long Waves of Innovation that Power Cycles of Growth, Cheltenham: Edward Elgar.
Atkinson, R. D. (2016), ‘Think Like an Enterprise: Why Nations Need Comprehensive Productivity Strategies’, Information Technology and Innovation Foundation, May, http://www2.itif.org/2016-think-like-an-enterprise.pdf?_ga=2.112124294.382854264.1506957444-487073861.1494271158.
Atkinson, R. D. (2017), ‘Unfortunately, Technology Will Not Eliminate Many Jobs’, Innovation Files, 7 August, https://itif.org/publications/2017/08/07/unfortunately-technology-will-not-eliminate-many-jobs.
Atkinson, R. D. and M. Lind (2018), Big is Beautiful: Debunking the Mythology of Small Business. Cambridge, Mass, and London: MIT Press, forthcoming.
Atkinson, R. D. and J. Wu (2017), ‘False Alarmism: Technological Disruption and the US Labor Market, 1850–2015’, Information Technology and Innovation Foundation, https://itif.org/publications/2017/05/08/false-alarmism-technological-disruption-and-us-labor-market-1850–2015.
Bivens, J. and L. Mishel (2015), ‘Understanding the Historic Divergence Between Productivity and a Typical Worker’s Pay: Why It Matters and Why It’s Real’, Economic Policy Institute, http://www.epi.org/publication/understanding-the-historic-divergencebetween-productivity-and-a-typical-workers-pay-why-it-matters-and-why-its-real/.
Brooks, R. A. (2015), ‘Mistaking Performance for Competence’, in J. Brockman (ed.), What to Think About Machines That Think, New York: Harper Perennial.
Chui, M., J. Manyika and M. Miremadi (2015), ‘Four Fundamentals of Workplace Automation’, McKinsey & Company, November, http://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/four-fundamentals-of-workplace-automation.
European Commission Skills Panorama (2014), ‘Skills Under-Utilisation Across Countries in 2014’, http://skillspanorama.cedefop.europa.eu/en/indicators/skills-under-utilisation.
Kennedy, J., D. Castro and R. D. Atkinson (2016), ‘Why It’s Time to Disrupt Higher Education by Separating Learning from Credentialing’, Information Technology and Innovation Foundation, https://itif.org/publications/2016/08/01/why-its-time-disrupt-higher-education-separating-learning-credentialing.
McGowan, M. A. and D. Andrews (2015), ‘Skill Mismatch and Public Policy in OECD Countries’, OECD Economics Department Working Paper 1210, 28 April, http://www.oecd.org/eco/growth/Skill-mismatch-and-public-policy-in-OECD-countries.pdf.
Miller, B. and R. D. Atkinson (2013), ‘Are Robots Taking Our Jobs, or Making Them’, Information Technology and Innovation Foundation, http://www2.itif.org/2013-are-robots-taking-jobs.pdf.
Rose, S. J. (2014), ‘Was JFK wrong? Does Rising Productivity No Longer Lead to Substantial Middle-Class Income Gains?’ Information Technology and Innovation Foundation, http://www2.itif.org/2014-rising-productivity-middle-class.pdf.
Rothwell, J. (2016), ‘Why Elites Want More Competition for Everyone Except Themselves’, Evonomics, 2 April, http://evonomics.com/why-elites-want-more-competition-foreveryone-except-themselves/.
Schwab, K. (2016), ‘The Fourth Industrial Revolution: What It Means, How to Respond’, World Economic Forum, 14 January, https://www.weforum.org/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond/.
US Bureau of Labor Statistics (2016), ‘Current Employment Statistics, Employment Level – All Industries Self Employed, Unincorporated, and Total Non-Farm, All Employees’.
[i] This is based on the share of wage and salary income by quartile.