The fourth industrial revolution and the possibility of automation

In the history of the evolution of human society, technological progress has been a continuous process, and such progress has been associated with automation of various degrees and kinds. That, in turn, had significant implications for employment and the world of work. The world is currently witnessing the fourth industrial revolution, the basic characteristics of which include the use of robots, artificial intelligence (AI), nano technology, and biotechnology. A common perception in that respect is that this is going to threaten employment of human beings.

Even in Bangladesh, where the economy is still characterised by the existence of surplus labour, robots are making inroads. And if one takes a long-term perspective of several decades from now, one could imagine the following scenarios: (i) in factories producing textiles, garments, shoes, etc., instead of human beings, robots are performing major tasks, (ii) instead of the numerous retail stores of different types, there are huge stores where robots arrange merchandise on shelves, customers pick up their needed items and go out through automated check-out points, (iii) online retailers have replaced most of the retail stores and their warehouses are run primarily by robots, and so on. Although these may sound unrealistic at this point in time, as Bangladesh aspires to become a developed country in another couple of decades, one could perhaps imagine such scenarios.

If the above scenario becomes a reality even before millions of underemployed workers find good jobs characterised by high productivity and incomes, there could be serious problems of mass unemployment and underemployment. And it would be logical for policymakers to take steps to prevent such a scenario. But how realistic would it be to paint such a scenario for the future - even if one considers a period of several decades?

ARE THE LUDDITES COMING BACK? The issue of machines displacing labour is not new to human society; it dates back to the early 19th century when the so-called Luddites (in Britain) attacked weaving machines because they were thought to be causing destruction of jobs in textile factories. And the question has resurfaced in the wake of several reports published this year by influential institutions including renowned private companies like McKinsey (2017), PWC (2017), and international agencies like the United Nations (UN-DESA, 2017) and the World Bank (2017). In the context of the Fourth Industrial Revolution currently under way, these reports analyse activities and occupations that are "automatable" and develop scenarios of job losses if such automation does indeed take place. While most of these reports focus mainly on developed countries, the analysis is not limited to them. For example, the McKinsey report divides the countries covered by it into three categories: (i) advanced economies, (ii) emerging economies with ageing populations, and (iii) emerging economies with younger populations. The countries in the last category includes India, but Bangladesh is not included.  As if to repeat the attack of Luddites to destroy weaving machines, measures like taxing robots are being proposed in developed countries. No other than Bill Gates has made this proposal.

If the concern can be so serious in developed countries, for a country like Bangladesh, a development of the kind mentioned above can really spell doom. Shouldn't policy discourse take a serious view of it? However, before starting with a pessimistic and doomsday scenario, it is necessary to take a careful look at what one is talking about. In doing so, one should also distinguish between prospects that are likely to be faced by countries at different stages of development. At the risk of saying the obvious, the concern cannot be the same in USA, UK, China, Viet Nam and Bangladesh.

WHAT DOES THE HISTORY OF AUTOMATION TELL US? It may be useful to refresh our memory with the history of technological progress vis-à-vis employment, and a few facts may be worth recounting in that regard.

• Automation during the first industrial revolution was associated with an increase in jobs - not decline. Between the early 19th and early 20th century, the number of textile jobs increased.
•  Although jobs were lost in the steel and textile industries in countries like UK and USA during the 20th century, it's important to understand whether that was due to automation or globalisation leading to these industries moving offshore.
• The spread of information technology (IT) in recent decades has been associated with a rise in employment. Except during economic downturns, the US economy has not faced a problem of shortage of jobs. In November this year, 228,000 new jobs were created in that economy - much higher than the normal annual average.
•  Even in recent years, automation has not been associated with a decline in overall employment. The example of Amazon is often cited in this context where there has been a sharp increase in the number of robots used, but hiring of workers has also continued.
•  If one takes a longer-term view, one would see that fears of mass unemployment have, by and large, been proven unfounded. Employment-population ratio has increased during the twentieth century.

So, what happens when technological progress takes place and activities and occupations are automated? There can be two types of technological change: "automating technology" that can replace labour, and "labour augmenting technology" that can, by creating new tasks, create new jobs. For example, automation may be associated with new jobs in the spheres of supervision, repair and maintenance. The net impact on employment would depend on the relative strength and magnitude of the two effects.

Yes, the first and immediate impact may be the loss of jobs as machines may indeed replace some human jobs. But in addition to this immediate impact, technological progress leads to changes that may have a positive effect on employment. For example, one positive impact is often a rise in productivity leading to a decline in prices and a rise in the demand for products. That, in turn, leads to growth of output and employment.

Second, technology replaces certain tasks rather than complete occupations. Of course, new jobs that are created are likely to require different types and levels of education and skills compared to the jobs that may have been lost. We shall get back to this issue in a moment.

Third, automation, by raising the productivity of workers, creates a necessary condition for wage increases. Moreover, by reducing the drudgery of manual jobs, machines may lead to improvement in the quality of jobs. 

What is also important to note is that only in a small proportion of occupations, jobs are completely automated. Machines often work together with human beings - thus creating positive complementarity and raising productivity.

Of course, there would be winners and losers as automation creates differentiation in the labour markets with implications for relative wages and incomes. The nature of jobs is likely to change with greater demand for workers with higher levels of education and skills, thus creating conditions for accelerated wage increases in certain jobs. And that can unleash forces for a rise in inequality in income.

SOME HIGHLIGHTS ON THE POTENTIAL FOR AUTOMATION AND ITS IMPACT: It may be useful to highlight a few points on the potential for automation and its impact, especially on employment.

• A small proportion of all occupations (5.0 per cent, according to the Mckinsey report) can be automated entirely.
• 60 per cent of all occupations have at least 30 per cent of constituent activities that could be automated.
• Manufacturing, transport and storage, and retail trade are sectors where the potential for automation is high. At the other end lie educational services, management, health care, professionals, information (36 per cent), etc.
• At the country level, the Mckinsey report considers the potential for automation to be highest for Japan (55 per cent), followed by India (52 per cent), China (51 per cent), and USA (46 per cent). The PwC report estimates the potential for UK to be lower than USA.
• China and India showing higher potential for automation than USA would naturally raise a question of methodology used in the report. It could of course be argued that the potential for automation in a country would depend on the mix of products produced and the potential for automation in various sectors. But the country variation appears difficult to explain even considering such factors.
• As for the impact on employment, while some jobs will be lost, new job opportunities will be created in sectors (e. g., services) and occupations that are difficult to automate. So, it is difficult to predict whether the net impact on employment will be positive or negative.
• Regarding individual workers, it is the less educated who are likely to be more affected and those with higher and more specialised education who are likely to gain.
• Public policy will have an important role to play in ensuring that the potential gains from automation are shared more widely and the brunt of the negative effects can be minimized.

The writer is former Special Adviser, Employment Sector, International Labour Office, Geneva.

[email protected]