Motorists in Singapore today rarely see parking attendants in carparks. Most carparks are now electronic, and parking charges are automatically deducted from our cashcards. We only require human assistance when parking if a problem occurs.
But parking a car in 1970s Singapore was a labour intensive process. As Mr Mohd Mohaimin Abu Bakar, a URA parking attendant then, recalled: “We did all the work. We were given a set of lots to take care of. When a car parked, we collected payment from the driver and issued receipts. When the parking time paid for expired, we issued an advice note and placed it on the car.” Singapore’s parking system required one parking attendant for every dozen or so car park lots and hundreds of officers in back-office operations (Azhar Ghani, 2011)1.
Singapore’s vast improvement in parking productivity illustrates how computer-driven automation has increased efficiency and economic growth worldwide over the last few decades. The ERP parking system requires far fewer workers than the labour-intensive parking attendant system of the 1970s. While it is easier to sit in an air-conditioned office and talk to drivers through the intercom than to handle payments in the hot sun, there are also far fewer jobs.
How will computerisation and automation change the nature of work in Singapore, particularly for the Malay/Muslim community? For decades, workers have expressed concerns about being replaced by new technology. Crucially, technology” does not have to be computer-based to replace jobs. The parking workforce shrank dramatically as early as 1980, when parking coupons were introduced. But the capabilities of modern computers and robotics far exceed the labour-saving technologies of the past. Tasks formerly impossible to automate, ranging from driving to financial modelling, are now potentially automatable given sufficient resources and computing power.
The best international research suggests three learning points relevant for the Malay/Muslim community:
First, jobs which rely on routine and systematic tasks, whether based in the office or the workshop, are the most vulnerable to automation, because computers are particularly good at accurately performing routine tasks. In many cases, the technology already exists to automate these jobs, and it is only a matter of relative costs. As wages rise in Singapore and the costs of automation fall, more routine jobs will be outsourced.
Second, new advances in robotics and computing power have enabled complex, non-routine tasks, ranging from driving to accounting, to become automatable in the near future. Workers who hold these medium-skilled jobs need to understand that as technology improves, the more costly and common medium-skilled jobs – such as driving – will become automated first. However, even skilled jobs that require thinking and analysis, such as accounting, research, and financial investing, are also potentially automatable.
Third, jobs which involve social intelligence and human interaction are complemented by automation and hence will remain relevant, regardless of whether those jobs are high-skilled or low-skilled. Therefore, professionals of all types, as well as service workers, will continue to be in demand. However, workers in people-facing jobs will have to increasingly depend on technology to increase their productivity and to stay relevant.
We estimate that about 66% of Malay/Muslim workers in Singapore hold jobs today that are susceptible to automation and computerisation (Figure 1)2. Malay/Muslims are significantly more vulnerable than the national population (45% at risk) because they are over-represented in medium-skilled jobs that rely on routine tasks. 17% are in clerical support, 18% in sales and services, and 11% hold skilled manual jobs in plant and machinery operation. While these occupations have provided a gateway to the middle class for many Malay/Muslim families, they are also the most susceptible to computerisation.
A FRAMEWORK FOR ANALYSING THE RISKS OF AUTOMATION AND COMPUTERISATION
Predicting the effects of automation and computerisation on jobs requires an understanding of the relative capabilities and limitations of technology and how that affects the susceptibility of a job to automation. Figure 2 breaks down jobs along two dimensions: whether the job relies on manual labour versus cognitive skills, and whether the key tasks involved are routine versus non-routine. The first wave of automation and computerisation started in the advanced economies in the 1960s and continued through the early 2000s, affecting jobs which specialised in performing routine and systematic tasks – such as filing records accurately, tabulating numbers, or assembling products (Autor et al., 2003)3. As anyone who has struggled with their smartphone knows, computers are not ‘smart’ in the conventional sense, and simply execute orders given by programmers. Routine and systematic tasks were historically the easiest to programme and automate, and hence were automated first.
Although workers who specialised in routine tasks were rapidly displaced by early automation, higher skilled workers saw their productivity grow because computers made them more effective at their jobs. Workers in management and professional roles, which require deep skills such as analysis, problem-solving, and creativity, benefited because computerisation eliminated the most tedious parts of their jobs. As a result, while middle skill jobs rapidly disappeared from many developed economies, demand for highly skilled professionals continued to grow from the late 1990s through the 2010s (Autor, 2015)4.
AUTOMATION TODAY AND IN THE FUTURE
However, deep, complex skills alone are no longer a barrier to automation today. The sophistication and power of computer programmes has grown dramatically over time, driven by a rapid fall in the cost of computing hardware; the average smartphone today has more computing power than a supercomputer of the 1980s5. The widespread availability of low-cost computing power, combined with recent advances in machine learning and robotics, has led to automation of even tasks that require deep, complex skills (Frey and Osborne, 2013). For example, the investment bank JPMorgan Chase & Co recently developed a programme named “Contract Intelligence” or COIN, which automatically reviews loan contract agreements, replacing 360,000 hours of legal services each year (Son, February 2017)6. Such technological breakthroughs may improve the productivity of top managers and lawyers, who can now concentrate on strategic decision making, but obviously replaces an army of skilled junior legal officers in the process. The latest advances in automation are already being piloted and tested in Singapore. Autonomous vehicles are already accepting passengers on a trial basis at One-North (Siong, October 2016)7. Other traditional service sector jobs – such as that of security guards – are being reshaped by automation. Concorde Security has developed a mobile command and control unit that monitors up to 30 buildings at once, reducing the manpower requirement from one security guard per building, to only 3 trained employees in the control centre (Law, October 2015)8. Further developments in artificial intelligence are likely to enable automatic recognition of trusted occupants, reducing the need for security personnel further. Even the process of making popiah skin has been automated with local manufacturer Mr. Popiah reporting a tenfold improvement in productivity from replacing workers with machinery (Lee, February 2017)9.
THE WAY FORWARD FOR THE MALAY/MUSLIM COMMUNITY
Industries which are poised for fundamental change from automation will not disappear from Singapore. Taxis and ride-sharing will still be here when automated cars are commonplace, but the traditional taxi driver will disappear from Singapore, just as the parking attendant has. With 66% of the Malay/Muslim workforce employed in occupations at risk from disruptive technology and automation, employees and employers alike must invest in adopting and complementing technology to raise efficiency and productivity.
To remain competitive, employers need to play their role in identifying potential areas of skills development for workers, especially in areas where their skill-level and job requirements can complement newly-adopted forms of technology. A particularly promising area lies in using technology to help build human-to-human relationships. For example, customer service ambassadors and autonomous vehicle controllers will still be required even when a human taxi driver is no longer needed. Sales representatives can still add value through building customer relationships even when the physical task of logistics fulfilment has been automated from the robotic warehouse to drone delivery.
Workers cannot adapt to the coming changes in the nature of work unless they also take active responsibility for their own lifelong learning. As a starting point, more Malay/Muslims could take advantage of existing Government schemes to help upgrade themselves. Speaker of Parliament Halimah Yacob recently highlighted that out of 126,000 claimants of SkillsFuture Credits in 2016, only 8.4%, or only slightly over 10,000 are Malays – far below their proportion of the population (Hasleen Bachik, January 2017)10. The reasons for this striking figure must be examined. If the lack of knowledge of skills required for future jobs or re-training programmes is the cause, then employers, the Malay/Muslim organisations, as well as media platforms can aid in bridging the information gap. However, if employees are resisting change, more must be done to identify workers in occupations at high risk of automation, to proactively convince such workers to set their careers on a new path before it is too late. Seemingly secure jobs are no longer secure. We must help to shape the future of work in Singapore, rather than let the future shape us. ⬛
1 Mr. Mohd. Mohaimin Abu Bakar’s quote is taken directly from Azhar Ghani (2011). See Azhar, G. (2011). The Parking Coupon System: A Rear View Perspective. IPS Update.
2 Our estimates are based on research by Frey and Osborne (2013), who estimate the probability of automation for 702 U.S. occupational classifications. We map these U.S. occupational classifications onto the closest corresponding Singapore occupational category as tabulated by the Singapore Department of Statistics, and we compute the risk of automation for each Singapore occupational category by taking the simple average of Frey and Osborne’s (2013) estimates for the mapped U.S. occupational classifications. Our procedure provides only a rough estimate of the risk of automation as we do not have the data to directly link Singapore occupations one-to-one to each specific U.S. occupational classification. See Frey, C.B., Osborne, M.A. (2013). The Future of Employment: How Susceptible Are Jobs to Computerisation? Oxford Martin Programme on Technology and Employment.
3 Autor, D.H., Levy, F., Murnane, R.J. (2003). The Skill Content of Recent Technological Change: An Empirical Exploration. The Quarterly Journal of Economics. 1279-1333.
4 Autor, D. H. (2015). Why Are There Still So Many Jobs? The History and Future of Workplace Automation. Journal of Economic Perspectives. 3-30.
5 In 1985, the Cray-2 Supercomputer was the fastest computer in the world, was the size of a large washing machine, and required a special liquid cooling system to operate. By 2010, each iPhone 4 provided similar computing power at a fraction of the price and space – giving millions of consumers the same power that only governments could buy in the 1980s.
6 Son, H. (2017, February 28). JPMorgan Software Does in Seconds What Took Lawyers 360,000 Hours. Bloomberg.
7 Siong, O. (2016, October 18). Test Bed For Driverless Vehicles Ramped Up At One-North. Channel NewsAsia.
8 Law, F. (2015, October 1). S’pore-Made Surveillance Van Can Cover 30 Buildings. Channel NewsAsia.
9 Lee, P. (2017, February 19). Popiah Firm Seeks Help Putting Automation On The Menu. The Straits Times.
10 Hasleen, B. (2017, January 22). Halimah: Masih Ramai Belum Guna Kredit Skillsfuture. Berita Harian.
Dr Walter Edgar Theseira is a Senior Lecturer of Economics at the Singapore University of Social Sciences. He holds a PhD in Managerial Science and Applied Economics from the Wharton School of the University of Pennsylvania. His research interests are in applied microeconomics and behavioural economics. He is currently working on research in transport economics, consumer credit, labour economics, and innovation policy.
Nabilah Isa is a Research Associate at the Singapore University of Social Sciences. She holds a degree in Economics with a second major in Public Policy and Global Affairs from Nanyang Technological University. Her interests lie in areas of the social sector, especially in issues pertaining to the less-privileged.
