The rise of the computers and the shrinking of the middle class

Posted by a.hay on 8 May 2014 - 2:00pm

By John Van Reenen, Professor of Economics and Guy Michaels, Associate Professor of Economics, Centre for Economic Performance London School of Economics.

Over the last 30 years, inequality has risen as new technologies have massively increased the demand for highly skilled workers. Yet as our research shows, it is not simply a case of the more educated benefiting at the expense of the less educated. Rather, it is the middle-skilled who are losing out most.

In the UK today, the richest tenth of male earners receive almost four times as much as the poorest tenth, whereas they only earned twice as much 30 years ago. A good chunk of this increase in wage inequality is due to higher returns to education which has been accompanied by a massive increase in the proportion of the university educated. The inescapable conclusion is that the demand for more highly skilled workers has risen. This holds true across the developed world. The academic consensus is that this increase in skill demand is linked to technological progress, driving up the demand for workers who are able to deal with a more complex and challenging workplace.

New facts on inequality have recently emerged. In the United States, ‘upper half’ inequality – the difference between the richest tenth of the population and the middle – has risen continuously over the last three decades. But Yet at the same time, after increasing during the 1980s, ‘lower half’ inequality – the difference between the middle and the poorest tenth – has actually fallen. As high school graduates’ wages (the wages of those who leave school at age 18) have ceased to increase relative to those of high-school dropouts (those who leave at age 16) since the 1990s, it also seems that jobs in middle-skilled occupations have decreased relative to both high-skilled and low-skilled occupations across Europe and North America

What could account for this ‘polarisation’, in which the prospects of the middle-skilled have been declining? One clue is to be found by looking at robot competitions in Japan. Every year in Tokyo, the Robo-One competition rewards the robot that is best at doing tasks such as cleaning, playing football, dancing and punching other robots (really hard). What is remarkable about this competition is not so much the sophistication of Japanese technology, but how bad these robots are at doing things that humans find very easy.

This suggests that what new technologies are very good at doing is replacing repetitive, boring, routine tasks such as assembly line work, but tasks that require responding rapidly to unfamiliar situations (such as driving or cleaning) are not easy for robots to reproduce.  Yet it isn’t only this group that has been affected. ICT has also reduced the need for middle-educated workers carrying out routine tasks. Bank clerks, for example, have found demand for their services plummeting as a result of computerisation.

More educated workers making analytical, non-routine use of ICT – such as management consultants, advertising executives and doctors – have found their jobs made easier by ICT rather than threatened by it. ICT has increased the demand for less educated workers to carry out non-routine manual tasks – such as janitors and cab drivers – contrary to claims that low-skilled jobs are disappearing. Since the number of routine jobs in the traditional manufacturing sectors (such as car assembly) declined substantially in the 1970s, the subsequent growth of computerisation may have primarily increased demand for highly educated workers at the expense of those in the middle of the educational distribution, leaving the least educated (mainly working in non-routine manual jobs) largely unaffected.

Technology is replacing the middle-skilled

To look at whether technology is reducing demand for those in the middle, we have conducted a simple test using 25 years of data across 11 countries and all sectors of the economy. If our story is right, we would expect industries that had a faster growth of computerisation to have also had an increase in demand for college- educated workers relative to workers with middle levels of education, leaving the least skilled unaffected. Our analysis of the data reveals that this is indeed what has been going on.

After 1980, countries with faster upgrading of ICT (Finland, the Netherlands, the UK and the United States) also saw the most rapid increase in high-skilled workers. Across different countries, similar industries – for example, financial services, telecommunications and electrical equipment manufacturers – replaced middle-skilled workers with high-skilled workers at the fastest rate.

This finding is consistent with a technology-based explanation for polarisation, but not with alternative explanations that emphasise institutional changes, such as the decline of trade unions or the introduction of minimum wages.

Changes in skill demand across industries are strongly correlated with ICT upgrading. Industries that experienced more rapid ICT upgrading also increased the relative demand for high-skilled workers at the expense of middle-skilled workers, with little impact on low-skilled. The change in demand reflected an increase in both the wages and the hours worked by high-skilled workers relative to middle-skilled workers. We document this finding not only for the full sample of countries together, but also separately for the United States and for continental Europe.

There is also evidence of technology polarising the demand for skills not only through increased intensity of ICT. Industries that engage in more research and development (R&D) also show the same pattern of substitution of middle-skilled workers by high-skilled workers. Taken together, ICT upgrading and R&D account for about a quarter of skill upgrading since 1980.

What about trade?

An alternative explanation for the falling demand for non-college workers is globalisation. The idea is that increased trade with low-wage countries like China has depressed the wages and taken the jobs of the less skilled.

We found that the positive correlation between trade openness and the increased demand for high-skilled relative to middle-skilled disappears once we control for technological change. This could either mean no role for trade or a more subtle effect whereby trade has an indirect effect by inducing faster technical change.

In either case, unless one believes in a Luddite view of smashing machines, there is no reason to erect trade barriers to protect less skilled workers from the effects of China, India or other emerging economies. Instead, we must conclude that polarisation is not bad news for the least skilled – there will be jobs for them even in a high-tech world. For the middle classes, however, technology may be endangering their future labour market prospects