Commentary: The optimist's case for automation

Commentary: The optimist's case for automation

Automation holds considerable promise for reversing the slowdown in workforce growth in developed economies like Singapore. The conditions needed for successful adoption are falling into place, argues McKinsey’s Diaan-Yi Lin.

SINGAPORE: Singapore faces a serious gap in labour productivity. Singapore’s labour productivity grew by an average of 2.2 per cent per year from 2000 to 2013, according to estimates by the McKinsey Global Institute (MGI).

Yet for Singapore to maintain historical levels of growth, the city would need to increase its labour productivity by an average of 5.8 per cent per year from 2013 to 2030 - a rate nearly three times as high as in the previous 13 years.

In many countries, increasing labour productivity usually requires employment to shift from less productive sectors, such as agriculture, to more productive ones, like manufacturing.

For Singapore, however, this model might not be workable. Its economy has been restructured to some extent, and its labour market is very tight. Employment shifts among sectors actually reduced Singapore’s economic growth from 2006 to 2012.

Further labour productivity gains need to take place within sectors. In this regard, automation technologies - which include artificial intelligence (AI), machine learning and robotics - hold considerable promise. Automation could raise productivity growth by 0.8 to 1.4 per cent annually in 20 large countries, according to MGI estimates.

To take advantage of this opportunity, executives and policy makers in Singapore need to understand the nature of today’s automation technologies and the potential productivity improvements that they offer to Singapore and other Asian economies.


For decades, organisations have used computers and machines to streamline and perform the work of humans. Recent advances in computing technology, programming techniques and data collection are making it possible for machines to do more cognitive work, such as finding patterns in data and reaching decisions.

When my McKinsey colleagues in China conducted a survey on AI, respondents identified more than 100 ways that AI might affect their industries. Innovation seems likely to accelerate, also because investment is pouring into new applications for AI.

Artificial intelligence can improve health care by analyzing data from apps, smartphones and
Healthcare is one area where investments are pouring into artificial intelligence and the development of apps, smartphones and wearable technology. (Photo: AFP/Patrick Hertzog)


A common fear is that automation will destroy jobs. But the situation is more nuanced. Tasks like collecting data or doing predictable physical labour can be automated readily. This is not the case for activities that involve social, emotional and cognitive skills, such as dealing with customers and managing workers.

By looking at the potential for automating activities, we found that just 5 per cent of occupations could be fully automated with currently demonstrated technologies. Many more could be partly automated: Some 60 per cent of jobs could have 30 per cent of their activities automated.

These jobs span the pay scales and ranks of organisations, all the way up to those working in companies’ “C-suite” leadership roles: Activities consuming more than an estimated 20 per cent of a CEO’s working time could be automated using current technologies.

Visitor has a picture taken wtih an operational robot policeman at the opening of the 4th Gulf Info
A visitor has a picture taken with an operational robot policeman at the opening of the 4th Gulf Information Security Expo and Conference in Dubai, United Arab Emirates on May 2. (Photo: REUTERS/Stringer)


Although automation will influence jobs in every sector and country, it will make more of a difference in some places than in others.

The potential for automation is concentrated in four countries with large populations, high wages, or both: China, India, Japan, and the US. These countries account for just over half of the wages and almost two-thirds of the work associated with automatable activities.

Of the 11 Asian countries that MGI studied, Singapore actually has the lowest proportion of work that can be automated with current technologies (44 per cent, which is admittedly still high).

In Singapore, much of the work that can be automated using existing technologies is in the city’s larger industries: Manufacturing (equivalent to 213,800 jobs), administrative and support services (134,200 jobs), retail (124,900 jobs), and construction (120,000 jobs). Two smaller sectors have particularly high percentages of automatable work: accommodation and food services (60 per cent) and transportation and warehousing (59 per cent).

sef spacehub 1 (1)
Integrated Construction and Prefabrication Hubs in Singapore allow for more efficient use of land and highly mechanised and automated systems. (Photo: Calvin Hui)


Some think that automation will happen rapidly, but it appears likely that the adoption of automation will take decades. The pace and extent of automation’s effect on work activities depends on five factors.

First, whether a demonstrated technology can be turned into a commercial product or service quickly. Second, whether the costs of development and deployment, which have to be covered in advance, can be eventually recouped.

Third, dynamics in the labour market, including demographics, wage levels, and worker training, which can help workers adapt to new technologies. Fourth, the type and distribution of economic benefits such as increased productivity, improved safety, lower labour costs and higher product quality, which determines whether companies have a strong incentive to adopt automation technologies.

Last, regulatory and social acceptance, related to issues such as safety and liability, data privacy and security, and possible increases in unemployment levels.

driverless vehicle NTU
Since 2016, Nanyang Technological University has put driverless vehicles like this minibus to the test. Such test-bedding and experimentation are helping to accelerate the rate of technology adoption.

A projected 50 per cent of all work activities could be automated by around 2035 if these five factors favour the rapid development and adoption of automation technologies.

Should automation develop more slowly, the same level of automation might not occur until 2075.


The productivity boost from automation in the world’s 20 largest economies could be equivalent to adding 1.1 billion to 2.3 billion full-time workers when we reach 2065, based on MGI's estimates.

This could increase growth by 0.8 to 1.4 per cent of global GDP annually. Such gains would offset some of the slowdown in workforce growth that is happening in many advanced and some emerging economies - a demographic trend that could cut economic growth nearly in half.

Previous periods of structural economic change created winners and losers, but not in a zero-sum way. In the US, for example, manufacturing employment fell from 25 per cent in 1950 to less than 10 per cent in 2010, but new jobs replaced the ones that disappeared and society was better off on the whole when the transition was complete.

As automation progresses, economic growth will increase most if workers who are affected by automation continue working at the same levels of productivity. Meeting this condition will require concerted action in the private and public sectors.

Healthcare robots
A robot at work at Changi General Hospital. (Photo: Vimita Mohandas)

Business leaders could find ways to redeploy the displaced, either within their own organisations or elsewhere. Policy makers should develop measures to help workers develop new skills and to promote the creation of new jobs.

Singapore is well-positioned to help its workers enter the age of automation, thanks to efforts like the SkillsFuture programme, which helps workers pay for the training they’ll need to keep up with the demands of the digital economy.

The Government also has institutions in place, like the Smart Nation Programme Office, that could assist with tracking the progress of automation and devising new initiatives to help companies deploy advanced technologies. Tying spending and incentives to investments in new technologies more closely could be one approach.

I am optimistic that Singapore and other Asian economies have the human and technological capital, as well as the international outlook, to capitalise on the opportunities created by automation while limiting the downside.

Diaan-Yi Lin is Managing Partner of McKinsey & Company in Singapore.

This is part of Channel NewsAsia's commentaries on robots and automation.

Watch our commentary on whether robots will steal your job:

"They work much faster than humans. They never take annual leave, and they don't need a retirement plan." Will robots one day replace human beings completely in the workplace? Why do we need so many of them? Toni Waterman explores.

Read also a commentary on the introduction of robots in hotels here.

Source: CNA/sl