How to think about automation
Everyday, both in print and digital media, scary stories emerge about robots killing jobs. For now, such stories are familiar in developed countries, but they foreshadow an emerging premonition that the process of automation will soon affect millions of jobs in developing countries.
Automation has many dimensions, but the one that gets the most attention is the automation of tasks that are laborious, repetitive, dangerous, dirty, or dull. More often than not, these types of jobs pay less than what is considered a decent wage.
Once upon a time, rickshaws used to ply major streets of Dhaka. Now Dhaka's roads are dominated by motorised vehicles, mostly private cars. This structural transformation from manual to automatic transportation has not created anxiety over automation, because such shift came as a result of economic growth, which tends to benefit all segments of the population, however unequally that wealth may be distributed. Interestingly, in many parts of greater Dhaka, motorised rickshaws, despite the nostalgia that surrounds it, are becoming ubiquitous.
Unless we know the "direction of automation" or "what it is that is being automated," it is difficult to talk about automation. For example, the automation of government procurement system in Bangladesh has shown positive results in terms of cost saving and good governance, which paved the way for automating and digitising services of various other public services. But automation is often thwarted not because of the fear of losing jobs, but due to fear of losing side income through illegal means.
At a public seminar at East West University, a former top NBR official lamented that although the management was looking into ways of automating the tax collection process since the mid-1980s, the initiative never had a chance of becoming a reality due to fierce resistance from unscrupulous tax collectors.
As a labour-surplus country, a real concern for us is whether automation is replacing jobs of the Bangladeshi workers, at home and abroad. The worry intensified after the recent publication of the Labour Force Survey, which reports that during 2013-2017, industrial employment grew only 0.5 percent, despite a robust manufacturing output growth of around 10 percent. This phenomenon, which economists often refer to as "jobless growth," has caught many by surprise. This is because by conventional wisdom employment share of manufacturing should increase until a country becomes an upper-middle-income country. So according to this logic, being a low-middle-income country, manufacturing jobs should expand until Bangladesh reaches the saturation point.
But manufacturing job share is not increasing despite robust industrial output growth because like other developing countries in South Asia and sub-Saharan Africa, Bangladesh is importing "deindustrialisation" from advanced countries. Due to rapid technological progress and rising labour productivity in countries like Germany and the United States, the relative price of manufacturing has fallen significantly than the rest of the economy. This is why manufacturing job share has been declining for decades in advanced countries while manufacturing output kept growing.
In contrast, Bangladesh's falling manufacturing jobs is more of an outcome of low-cost imported machinery, and less due to productivity growth in manufacturing at home. If productivity growth were accelerating, both employment and output in manufacturing would have risen. The squeeze of manufacturing all over the world due to automation is also happening in Bangladesh. "The robots are coming for garment workers" read a recent headline in Wall Street Journal. This is a good step forward because if our manufacturing sector does not embrace the ongoing changes in the nature of production, we will soon be outcompeted by capital from neighbouring Asian countries.
While a gradual but sustained adoption of automation will make garment factories a somewhat quieter place than before, a critical question at this stage is: how will Bangladesh generate the desired economic growth when competition comes from relentless, scalable, and cheap computer software and robotics? Rising income inequality in advanced countries has generated a large pool of low-wage workers of their own. Under this scenario, even without mentioning the growing anti-trade sentiment, one wonders how long Bangladesh's trade competitiveness will continue to be determined by its lower wages.
Today, globalisation and technology are largely driven by the international supply chain. This makes it easier for domestic factories to be part of the global production network than in old days where an entire sector had to be developed. However, Bangladesh's engagement in buyer-driven, rather than producer-driven, international value chains limits its exposure to the international transfer of knowhow.
We need to realise that technology, of which automation is a byproduct, is nothing but people with technological ideas. Ultimately, it is the human factor that determines the extent to which a country can benefit from automation. The Arabs had walked over the sands for thousands of years, without realising oil beneath the desert sands until American engineers shared the knowhow of drilling and exploration. The hackers who stole USD 100 million from Bangladesh Bank had better knowhow than the IT personnel at the central bank. The upshot of this is that the best way to deal with the scope and threat of automation is by developing and nurturing human capital that is socially useful. Otherwise, unlike the USD 100 million cyber theft, an even bigger chunk of our economy could be hijacked by the force of automation from our neighbouring countries.
Syed Basher is an associate professor of economics and Pooja Agnes Gomes is a graduate student of economics at East West University (EWU). This article is based on Pooja Agnes Gomes' undergraduate research paper for the Department of Economics, EWU.
Comments