Substack

Tuesday, March 20, 2018

The economic efficiency Vs social stability trade-off

The challenge posed by the rise of robots and resultant automation is well known. This is a nice summary of the evidence, 
Multiple studies suggest this is just the beginning and that there is more pain that lies ahead. A study by a real estate firm CBRE suggests 50% of occupations today will be gone by 2020. Then there is one by Oxford in 2013 that forecasts 47% of jobs will be automated by 2034. Yet another study has figured that only 13% of manufacturing job losses were due to trade. The rest has happened due to automation. And to make things worse, a McKinsey reckons 45% of knowledge work activity can be automated.
It is not a hyperbole to describe automation as the apotheosis of the search for economic efficiency. Robots make no mistakes, are more adaptable, can work 24X7, are much cheaper than labour, pose not threat of unionisation, and so on. Robots are super-efficient.

But this pursuit of efficiency in the modern economy sits with another trend - declining productivity in services sectors, those presumably most likely amenable to technological disruption and automation. 
Many of these services have seen increase in their share of the US labor force. Noah Smith has a very good article which captures the dilemma posed by this apparent weakness of the services sectors,  
The U.S. economy is sending more and more people into the sectors where productivity is either growing slowly or even falling... Is the stampede of American workers into unproductive industries really a bad thing?


Most economists would answer “yes” -- if construction, health care, education and the rest became more productive, workers would be freed up to go do other, more productive things, perhaps in industries that don’t even exist yet. But it’s also possible that some of these workers would otherwise just choose not to work -- to sit in their parents’ basements and play video games -- or to try to strike it rich in black-market sectors like drug sales. It’s also possible that the economic pressures of automation and trade, combined with the difficulty of retraining for new careers, would be sending some of these workers onto the welfare rolls instead of into new, better jobs.
And this conclusion is very sobering but rarely discussed in the mindless pursuit of efficiency,
So it’s possible that construction, health care, education and other industries are now functioning partly as giant make-work programs. Instead of giving a few people obviously useless jobs, this make-work system hides little bits of useless work in everybody’s job. That could be preserving the dignity of work for thousands, or even millions, of men and women standing around on construction sites, filling out paperwork in hospitals, or filing briefs for frivolous lawsuits. And that dignity, in turn, could be saving the U.S. from greater social unrest than it’s already experiencing.
In the efficiency and evidence maximising world-view that has gripped the worlds of business and academia respectively, the aforementioned would constitute an inefficient and therefore bad equilibrium. But when we step back and take a more comprehensive view, this may actually be a desirable situation. 

Change, especially by way of technological and social progress, is generally good. But such changes generally have an appropriate pace. Expedite the change and stresses will develop to disrupt the system, especially those with too many moving parts. There is no escape from the law of unintended consequences. Despite all its struggles, an organic evolution without too many mutations is the best response to such situations. The role of public policy should be confined to facilitating the process as well as mitigating the adverse consequences. 

This applies as much to a plunge towards automation and efficiency, as with the journey to become formal and shrink informality, or embrace digital technologies to reduce corruption. Press the pedal too early and too fast, and breakdowns or crashes are inevitable. 

No comments: