I spend a lot of my time thinking about the future of blue-collared work. As the co-founder of Nimbus, we build technology to make it easy for our clients to run their office. Unlike other on-demand startups, we do not employ people as independent contractors nor do we seek to replace our workers with robots. On the contrary, we invest a lot in proper employment, hiring, training and paying our blue-collar workers well to perform both scheduled and on-demand work from cleaning to aircon servicing and pantry stock up.
However, our good jobs strategy seems to fly in the face of the prevailing wisdom that robotics and automation will eventually replace all routine and mundane work. There is a passage in Erik Brynjolfsson and Andrew McAfee’s book Race Against the Machines that best encapsulates what I call the ‘Futurist’ view of work:
“Rapid and accelerating digitization is likely to bring economic rather than environmental disruption, stemming from the fact that as computers get more powerful, companies have less need for some kinds of workers. Technological progress is going to leave behind some people, perhaps even a lot of people, as it races ahead. As we’ll demonstrate, there’s never been a better time to be a worker with special skills or the right education, because these people can use technology to create and capture value. However, there’s never been a worse time to be a worker with only ‘ordinary’ skills and abilities to offer, because computers, robots, and other digital technologies are acquiring these skills and abilities at an extraordinary rate.”
I think the view is misguided – arguments about the future of work are too often framed in binary terms: either robot or human worker, machine learning or human memory. The reality is even if machines do take over some form of human activities in blue-collar work, it does not necessarily spell the end of the jobs in that line of work. The answer is often a bit more nuanced. I submit humans possess a lot of tacit knowledge that’s very hard to capture and automate. Overall, I think there are reasons to be bullish about the overall demand for blue-collar work. I shall outline my rationale below.
The Robot Apocalypse View
If we take a step-back to unpack this argument, Futurists who believe in books like The Rise of Robots argue typically make two critical assumptions:
(1) ‘Low-skilled’ jobs like cleaning and plumbing are routine and mundane, and it’s easy to build a robot to perform all of these tasks
(2) Robots are cheaper in the long run and less prone to human error which encourages a substitution effect from labor intensiveness to one which is more machinery intensive
Therefore, it follows from (1) and (2) blue-collared work like cleaning would soon be substituted by machines.
Robot Apocalypse Does Not Correspond With History
While the logic of the apocalypse view described above seems convincing, the first problem it faces is it flies in the face of empirical evidence that there are still many jobs available out there.
In his Ted Talk, MIT professor David Autor highlighted that many technological innovations over the last 200 years that were explicitly designed to replace human labor- Tractors, Assembly lines, Computers you name it- have succeeded. However, the percentage of US adults employed in the labor market is higher now in 2016 than it was during the industrial revolution.
Let’s first take the example of the ATM. Since the introduction of ATMs, which was widely touted to replace bank tellers, the number of human bank tellers employed in the United States has also roughly doubled, from a quarter of a million in 1970 to about a half a million today. Research by Basker (2015) showed that while barcode scanners have reduced a cashiers’ check out times by “18-19%, the absolute number of cashiers has grown at an annual rate of 2.1% between the 1980s through to 2013 since scanners were widely deployed during the 1980s.”
And the same phenomena can be seen in blue-collar work. In Singapore, Census data suggests there are more cleaners and laborers in 2000 than there were in 1990. Employment insights platform, Skillta also shows the demand for cleaners in Singapore is consistently increasing week-over-week. In absolute terms, there seem to be more cleaners and blue-collar workers than ever before.
In other words, there are still plenty of jobs around.
Now, this is not to say that machines have never succeeded in eradicating jobs. Examples of the lift operators or agricultural work in the US immediately comes to mind. But widespread unemployment due to technology has never materialized before, and it certainly does not look to be the case for blue-collared work. In fact, rather than automation accelerating job redundancy, I predict it will slow down. It is because so far, when it comes to automation, we have been picking the low hanging fruit first and this has given us an overly optimistic idea of how easy is it is to automate a job fully.
If machines can increasingly do the work for firms, why doesn’t this make labor and skills redundant?
Jobs are not as routine and mundane as the Futurist think. Tacit Knowledge cannot be automated.
I wonder how many people who make predictions of cleaners ever did a cleaning shift before to see what is it a cleaner exactly does. Cleaners don’t just clean as an autonomous iRobot ‘cleans.’ They have to take stock of missing pantry items. They determine what constitutes trash or valuable office items or where best dispose what kinds of trash. They water the office plants. They also have to analyse when best to clear the trash depending on an office’s workflow and meeting schedule. When talking to employees in an office, they act as customer representatives. When working with the office manager, they serve as a workspace consultant. They learn what the best way to clean specific areas. Many of these tasks require very ad-hoc and human judgment. These tacit knowledge are difficult to transfer.
Tyler Cowen gives a great example of another blue-collar occupation- truck drivers. In his commentary, he points out truck drivers don’t just drive trucks. They act as security for the load, they also “secure loads, including determining what to load first and last … They deal with the government and others at weighing stations… they also do all the bookwork, preventative maintenance, taxes, etc.”
In other words, the Futurists analyses of machines substituting routine work tends to overlook the complexity of blue-collar jobs at the granular level, severely underestimating the importance of local, tacit knowledge.
At Nimbus, when we deploy autonomous vacuum cleaners or robots to take over some of the activities of our workers, this does not spell the end of our worker’s jobs In fact, it boosts the demand for our workers to perform complementary tasks, like fault-reporting of office space, conducting pantry stock-taking or socialising with new employees in an office environment to make them feel welcome.
In the services sector, where workers are often over utilised, the adoption of more machinery and robotics can alleviate the strain and increase the leverage of workers to become more productive. This often leads to higher salaries, more exciting work and just as many jobs as ever before. In this sector, machines are tools to enhance human productivity, not agents of creative destruction.
Task displacement does not equal job redundancy.
Substitution effect is only one part of the picture. Cheaper relative costs can indirectly create even more jobs than before.
Furthermore, suppose even if a machine can be so efficient and automate many of the tasks of a single worker, job creation can still happen due to a firm’s increasing bullishness to expand their output. It could indirectly create more employment.
It was the case in the earlier example of the ATM. As David Autor showed, even while the number of bank tellers per branch fell by a third because it became much cheaper to run and open new bank branches, banks became more bullish about their business, and the net result was more branches, more bank tellers. And these bank tellers started to do different works which the machine couldn’t. For example, they began to do less of the cash-handling tasks and focused more on forging relationships with customers, troubling shooting banking issues and upselling more banking products and services. In this case, task displacement by a machine has occurred, and it led to more employment. Task displacement does not necessitate human replacement.
In the future, no doubt there will be new and better autonomous robots, more software and technology to automate office maintenance and cleaning. But at the end of the day (quite literally speaking), I am willing to bet someone will still have to service the space, top up the chemicals, water for the robotic cleaning machines, clear the trash, check for dust in non-obvious spots or welcome new employees with a friendly smile.
In my mind, the future of work and technology is symbiotic, not mutually exclusive.
Innovative business models that aim to co-opt and build technology that empowers their workers, rather than replace them are the smarter bet.
This article was published on Nimbus blog on 8 March 2018 by Daniel Thong, co-founder of Nimbus, who believes in changing the perception of blue collared workers in Singapore.