In the autumn of 2019, the Massachusetts Institute of Technology (“MIT”) published a report entitled, “The Work of the Future: Shaping Technology and Institutions.” This report comes amidst an increasing landscape of technological buzzwords in the 2010s, featuring things like Artificial Intelligence, Machine Learning, Internet of Things, Mass Digitalisation and my personal favourite (rather meaningless) buzzword, “The Fourth Industrial Revolution.”
It is in this landscape of technological buzzwords that many, myself included, have spilt printer ink on a whole variety of issues ranging from technology as a solution to all our societal woes, to how technology can both increase and decrease inequality, to the rise of the gig economy and the ‘precariat’ class, and, perhaps most of all, to what such rapid advances in technology may mean for the future of work and jobs.
To a very large extent, the impact of technology on society is something that can be studied by turning to history. After all, in the lingo of the World Economic Forum’s (rather meaningless) “Fourth Industrial Revolution”, if we are really in the fourth Industrial Revolution, we can look at the first, second and third Industrial Revolutions to see what happened.
Indeed, the MIT report poses exactly this point – “In prior eras, mechanisation and automation eliminated much undesirable work, while creating substantial new and more desirable work, and simultaneously raising productivity and enabling higher living standards. Does the current era of digital technologies possess these same virtues – or is it different this time?”
According to the report, there are two main reasons to believe why this current era is different. First, in earlier Industrial Revolutions – mechanisation, electrification, computerisation – tended to create strong earning opportunities for both blue- and white-collar workers particularly a strong and vibrant middle-class, the current trend of mass digitalisation tends to polarise the labour force. Digital automation today tends to displace middle-skilled workers performing routine cognitive tasks – call centre workers, back office administrators and so on – thereby splicing the labour force into high-skilled workers and low-skilled workers.
Indeed, the report states that, in the United States, in 1970, middle-skill occupations accounted for 38% of employment. In 2016, this number has fallen to 23%, as a consequence of both rising automation and international trade. The report argues that, “…unlike the era of equitable growth that preceded it, the digital era has catalysed…the simultaneous growth of high-education, high-wage and low-education, low-wage jobs at the expense of middle-skill jobs.”
The second reason why this current era of digital automation is different is due to a change in the nature of the technologies we see rapidly spreading today. To understand this, we need to observe a fact – in the United States and the European Union, productivity growth has been very sluggish since the mid-2000s, a reversal of the trend from 1975 to 2005, and of the peak trend from the first three decades after World War II.
This fact is something that should concern us deeply. If we are seeing such rapid technological advancement and we believe that this progress should make us more productive as a society, why is productivity growth slowing? How do we reconcile this anaemic growth with the technical complexities of the new technologies around us?
Well, at risk of over-generalisation, there are essentially two types of technologies or automation vis-à-vis their impact on productivity. The first is automation that substitutes human labour with machines. If the long-term cost of machines is cheaper than that of humans, we simply make the humans redundant and have machines take up the production. Examples for these include auto-pay parking machines, tax preparation software, automated call centres, chatbots and so on.
The second is automation that may complement workers. This type of automation does not displace workers, rather they provide tools that help workers do more with less, i.e. become more productive. Computers are an excellent example of this – consider how much more productive NASA would have been had it had today’s computing power instead of depending on human computers as it did in the 1950s and 1960s. For economists, statistical analysis software such as STATA allows economists to run regressions far more easily than programming new codes every time.
The distinction – vis-à-vis impact on productivity growth – between the two types of technologies is as follows. While both types of technologies may raise productivity, complementary technologies tend to increase earnings because they make workers more effective in their existing job tasks, allowing those workers to focus on more value-added work. Simply substituting a worker with a machine may not allow for that shift in production capabilities. As the report puts it, “…not all innovations that raise productivity displace workers, and not all innovations that displace workers substantially raise productivity.”
Indeed, the economists Daron Acemoglu and Pascal Restrepo label automation that disrupt employment without a corresponding boost in productivity as ‘so-so’ technologies. The report then goes on to argue that some of today’s digital innovations are ‘so-so’ in nature and may therefore explain the paradox of anaemic growth despite the wondrous technical complexities of today’s new technologies. To be clear, not all technologies are ‘either-or’ – some technologies may both substitute some workers and complement others, and there are certainly digital innovations today that are complementary in nature but if they are biased towards high-skilled workers, then we face the same problem of labour market polarisation as described above.
There are three lessons here that I believe Malaysia should apply as it considers its technology policy moving into the 2020s, assuming of course that the government intends to produce a clear technology policy, the last of which was the National Policy on Science, Technology and Innovation in 2013. The first lesson is that the pursuit of technology for technology’s sake is silly. What we care about is how technology ultimately improves societal welfare – be it in terms of worker’s earnings, firm productivity, and economic growth. Figuring out the appropriate outcomes and KPIs for technology is essential.
Secondly, as we consider what forms of technology we want to support and push for, we need to be more nuanced and figure out if a particular technology is going to be substituting or complementing workers, or both. Besides the economic benefits or costs associated with either forms of technology, there are societal benefits and costs as well, particularly with increased polarisation.
Thirdly, and finally, technology is not some knight-in-shining-armour saviour to all our problems. There are things where technology can be the chief driving force, and there are things where technology can only be a supporting force. As we enter into the 2020s, Malaysia still faces issues of corruption, of discrimination, of entrenched inequality, of cronyism, of over-obedience, and much more. These issues do not go away easily and we should not be fooled into thinking that more rapid technological advancement will make these issues go away. These issues cannot just be hand waved away by technology and startups alone, they require all hands on deck – firms, government, institutions, NGOs, and citizens.