Dean Baker: Confusing the Confusion
In a recent piece on Project Syndicate Dean Baker warns us not to be confused between the impact of worsening demographics on labour demand and the horror stories of robots stealing our jobs.
I have to admit there are times when I imagine those at the more extreme end of the latter camp sounding a little like this.
But I think Baker himself confuses two distinct issues in his account. He is right to say that the demographic situation poses a challenge for Western economies as declining shares of the total population will be of working-age and therefore paying into the system to support a growing numbers of retirees. What I am less sure of is whether this need necessarily equates to a labour demand story.
What I mean is this: Why should the growing demands of the elderly on the state necessarily suggest we have a labour shortage rather than a fiscal problem? That is, why should paying for the elderly be assumed to be a problem of too few workers rather than simply (mis)allocation of capital?
Jonathan Portes, director of the National Institute of Economic and Social Research (NIESR), put it as follows:
“The medium- to long-term problem for public finances is predominantly one of healthcare and pensions. To my mind, I actually believe that thinking of those as a fiscal problem is probably the wrong way to start.
We are getting older and richer, which means that we will be spending a lot more on healthcare in 20 years time and that’s what we should be doing. There is no model you can write down of preferences under which that would not be the right thing to do. That has to be paid for and ultimately it has to be paid out of our pockets. And then there’s a choice over whether that funding comes from taxation, private insurance, some sort of social insurance or direct out-of-pocket payments. You have to come up with a system that is both politically acceptable and that has as much equity and efficiency you can achieve.”
Viewing longer life as a bug in the system to be overcome rather than a positive feature to be celebrated is part of the problem. Even economists, who are not overly prone to bouts of unfettered exuberance, should be able to applaud the success of medical science in achieving this feat – rather than decrying its long run impact on government finances.
Baker is pleasantly philosophical on the demographic issue noting “if the benefits of productivity growth are widely shared they will swamp the negative effects of a declining ratio of workers to retirees” as they have done since the early 1960s. I would add that, as Portes says, it is well within the power of governments to adjust the tax rate to reflect demographic changes as well as undertaking structural reforms of state healthcare and pensions to prepare for them.
On robots, however, I think Baker is plain wrong on what the more sophisticated arguments are for the implications of technological progress on the workforce. It is certainly true that labour saving technology has helped significantly improve productivity and that this has been, in the most part, to the benefit of most people in developed countries.
The shift from factory production-line work to a services-focused economy, for example, can mean both higher wages but also work that requires more cognitive engagement. In that way it has, for many, encouraged the use of fundamentally human skills such as creativity and emotional sensibilities that can enhance both the environment of work itself and the experience of the end customers.
Yet to separate the story of automation from labour’s falling share of output, as Baker attempts to do, seems to me to miss one of the most important trends of recent decades. While the initial productivity boost helped drive the creation of new, high-skilled, better paid jobs it also lead to an increase in low-skilled, low paid services jobs where people are treated much like the robots on the production line.This trend has become evident in the job creation statistics as shown in MIT’s David Autor’s 2010 paper on The Polarization of Job Opportunities in the U.S. Labor Market:
While the pace of high-skilled job creation easily outstripped those in low-skilled professions between 1979-1999, this trend has inverted since then. Most of these low-skilled jobs involve caring for others such as “food service workers, security guards, janitors and gardeners, cleaners, home health aides, child care workers, hairdressers and beauticians, and recreation occupations” – e.g. jobs where robots as yet cannot compete.
So unlike Baker I do not believe that the “problem is an institutional structure that systemically redistributes income upward”. Factory-owners, and before them wealthy landowners, relied on labourers to produce the goods and services that they desired or sold. That created a bond of necessity between them, and laid the foundation for the labour union movements that were to come. Due to improvements in automation, however, increasingly today’s producers don’t need an equivalent amount of labour to achieve those ends – and see no reason to redistribute the spoils.This to me suggests the lack of real wage growth may not simply be a consequence of high unemployment but the chronic undervaluing of “soft” human skills that focus less on products and more on personal interaction. It also indicative of an inability to adequately compensate people for the risks inherent in highly flexible labour markets since the reforms of the 1970s despite the increased efficiency worker flexibility offers. Until policymakers and the wider public learn not to fetishize the production-line jobs for life of the past we may indeed have much to fear from the robots.
Redistribution and the Hollow Middle Class - Tomas Hirst
The Financialisation of Labour - Frances Coppola
The changing nature of work - Frances Coppola
UpdateBaker has responded on Twitter:
"I see Autor's story as more handwaving than evidence: http://emlab.berkeley.edu/users/webfac/moretti/e251_s13/mishel.pdf"
And my response:
"Thanks. Leaving Autor aside for a moment, still think inflation staying restrained with unemp below 5% in 90s suggests our understanding of changes in labour's bargaining power very limited - as Roger Farmer's work suggests: http://www.pieria.co.uk/articles/is_the_natural_rate_of_unemployment_an_out-of-date_concept"