The readymade garments (RMG) industry is vital to Bangladesh economy. It contributes 84 percent of the total exports—nearly 13 percent of the GDP—and employs some four million workers. The direct and indirect impacts of the RMG sector in the country are profound. Because it employs so much of unskilled labour, the sector looks particularly vulnerable to automation: will machines replace the workers and unleash mass unemployment? The fear is not totally unfounded since robots are already common in manufacturing. In textiles, smart sewing machines, jacquard machines, smart sensors, big data, 3D printing, etc. are already in place. The automation of cutting, drawing patterns, spreading, and relaxation of fabrics has quietly begun, while some administrative tasks like maintaining attendance or working hours, which were previously carried out manually, have now become digitised.
However, the widespread perception that industrial robots will take over and automate everything is questionable. Globally, the share of industrial robots used in textile and apparel is minuscule at best. For example, of the 1.63m industrial robots used in 2015, only 1,580 were in textiles and apparel. The hype generated by the prospect of sewbot as a potential disruption in the apparel industry is unwarranted. Much textile automation is about smoothing supply chains and reducing environmental strains rather than about replacing abundant cheap labour. As pointed out by The Economist in 2017, the unpredictable formations of fabric make it very difficult for a robot to keep track of what it is handling and where to apply itself. What is remarkable is that despite spinning being the first process to succumb to industrialisation in the early 19th century, textile and apparel activities still have to be guided by hand. The actual impact of automation is an empirical question.
In new research, we carried out a field survey to understand unemployment in the RMG sector by interviewing garment workers who have recently become jobless. The survey approached nearly 200 respondents (including garment officials) across ten areas in Dhaka. Regrettably, the last portion of the survey overlapped with the Covid-19 outbreak, and prematurely ended our survey.
To date, the early adopters of automation were the large apparel factories. Typically, these factories employ a large number of workers, operate from their own premises, and receive steady export orders from the buyers. By boosting efficiency, new technologies are permitting big factories to focus on greater product diversification. Despite automation, large factories have often kept displaced workers as more labourers are still needed to produce a much higher volume of output.
By comparison, medium and small units could not afford expensive automation. Today, they mostly survive on sub-contracting and, with limited bargaining power over western retailers, they often take orders at depressed prices. With the improved productivity and product diversity of other garment exporting countries in Asia, international buyers now enjoy multiple sourcing options. Some predict that with the passage of time many small factories will be forced to close down or consolidate, while some medium factories will continue to operate manually.
To get an idea of how automation impacts garments manufacturing, consider the initial stage—"pattern design". Before automation, pattern design needed 10-12 workers to complete a task; but after automation, with computers and 3D printers, the same task now requires only 1-2 workers. Similarly, the "spreading" section formerly needed 10-12 workers, but after the automation it requires only 2-3 workers. The "cutting" section has been partly automated. Previously, cutting required 100-120 workers, but now 60-70 workers suffice. In making sweaters, an operator can supervise six jacquard machines at a time, greatly reducing the need for workers to perform repetitive manual tasks. Hence, a good number of sweater factories were forced to terminate workers, whose jobs are being replaced by jacquard machines. Our analysis shows that it is certainly possible for automation to cause much unemployment, but has it actually done so?
Interestingly, garment workers displaced by automation were often not laid off but reallocated to other sections to boost production. Some factories are also providing on-the-job training so that employees become familiar with technology, yet others hoard labour with the expectation that they can be used on a just-in-time basis. The picture that emerges from our survey is that, if before the automation 50 workers were needed to produce 5,000 pieces a day, after automation a factory requires 30 workers to produce the same output while the remaining workers continue to work with the factory in a new line. As a consequence, with automated processes, 50 workers can now produce an estimated 8,000 pieces a day.
While automating an apparel factory can undeniably be expensive, several owners argued that it takes only 1-2 years to recover the purchase cost. In fact, factories that invested in high-tech manufacturing were able to secure new orders to fill up the extra capacity. Sadly, the small and medium factories could not exploit the scope of automation like their large counterparts. Automation is disrupting the entire sector on an ongoing basis. The small and medium factories are struggling to operate at break-even point by any means; the ghosts of Rana Plaza and the wave of automation weigh them down. Many have already shed redundant workers and kept only the skilled, multitasking, and experienced workers. Quite a lot of factories were closed down after encountering huge losses.
Another finding that came out of the survey was that the increase in the minimum wage has increased the unemployment of RMG workers. Factories that embraced automation and did not face difficulty in getting steady orders still laid off unskilled workers in the wake of a higher minimum wage. The combination of falling prices and higher minimum wage led to worker layoffs in some factories as they restructured to meet intensified global competition. Factories that were unable to compete or adapt to new technology laid off unskilled workers foreseeing a decline in economic activity.
In Bangladesh's RMG industry, automation has hitherto been a secondary reason for job loss. Unemployment rose more due to factors such as squeezed profit margins, higher minimum wages, and a slowdown in export orders. Automation has affected both skilled and unskilled workers and both high- and low-paying occupations. However, low-paid workers disproportionately bear the brunt of automation's impact as they lack the basic knowledge to operate advanced machinery and also are not often chosen for training to operate such machines.
The most worrying development is the prejudice that female workers are either physically or mentally incapable of handling machines and equipment. This prevents top management from encouraging female workers to receive the necessary training to upgrade their current skills. When the shortage of skilled workers is holding back RMG potential, the tendency to deprive female workers of upper-end work is short-sighted and wrong. While new technologies are creating new job opportunities, they should be geared towards females, since a plethora of research show that compared to men, women workers are less mobile and therefore less likely to switch jobs. Prejudice may be depriving Bangladesh of the profit-maximising option.
The clear conclusion that emerged from our survey is that it is not automation but rather falling export orders, rising wages and decreasing global prices that are behind the rising unemployment in the RMG sector. If these trends continue, textile manufacturers will have no option but to embrace automation and cut jobs in order to revive the business.
Jobaida Behtarin, Salim Rashid and Syed Basher are, respectively, a research associate, university professor, and professor of economics at East West University.