Is a universal basic income the answer to rampant automation?
- Ivo Vegter
- 27 Sep 2017 01:35 (South Africa)
The notion that automation, artificial intelligence and robotics will put many, if not all, of us out of work, is gaining traction. The idea is neatly (but wrongly, in my opinion) summed up in a video by the wonderful CGP Grey, entitled Humans Need Not Apply. Grey argues that this time it really is different. A similar view is held by Jerry Kaplan, an entrepreneur, futurist and technology investor, who adds that men will be harder hit by automation than women.
These arguments often cite a study by two researchers at Oxford, Carl Frey and Michael Osborne, which finds “47 percent of total US employment is in the high risk category, meaning that associated occupations are potentially automatable over some unspecified number of years, perhaps a decade or two”.
This paper only investigates the likelihood of labour disruption, however. It does not consider the effects of productivity gains, or the possibility of new jobs arising. As such, it considers only short-term disruption, and not long-term demand for labour.
Frey and Osborne prominently cite John Maynard Keynes, who made exactly the same prediction in 1933: “We are being afflicted with a new disease of which some readers may not yet have heard the name, but of which they will hear a great deal in the years to come – namely, technological unemployment. This means unemployment due to our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour.”
Before Keynes, it was Marx, in 1867: “Within the capitalist system all methods for raising the social productivity of labour are put into effect at the cost of the individual worker; all means for the development of production undergo a dialectical inversion so that they become means of domination and exploitation of the producers; ...they alienate from him the intellectual potentialities of the labour process in the same proportion as science is incorporated in it as an independent power...” (At least, I think that is the same warning, phrased in Marxist cant.)
Before Marx, it was Ned Ludd, who destroyed two mechanical knitting machines in 1779 and gave rise to the anti-automation Luddite movement of the 19th century.
In Grey’s video version of the story, he uses the example of horses, which became largely obsolete when the automobile was invented. But that he cited horses, rather than people, is telling. Technological progress in transportation did not cause long-term structural unemployment. We no longer have as many horse breeders, stable boys, coachmen, grooms and farriers, but this disruption in the labour force was temporary. As wrenching as such a change may be in the short term for individual people, to extrapolate widespread structural unemployment from it is not justified by the evidence. Sure, automated transport that does not need drivers will put millions of drivers out of work, but there is no reason to believe that they will not be able to find alternative employment.
The same happened during the dramatic change from a world in which 95% of all people worked in agriculture, to today’s world in which only 5% do. The other 90% aren’t sitting around in a daze wondering where their jobs went. They moved to different, more productive employment. Not only that, most of us rather enjoy being free of back-breaking farm work. On average, we have easier jobs, more leisure time, more money, and better living standards as a result.
Today’s automation is also not the first time white-collar workers are threatened. “Computer” used to be a job title for people who calculated things like insurance premiums and rocket trajectories. Today, it refers to machines, which put millions of clerks, bookkeepers, secretaries, office assistants and human computers out of work in the two or three decades around the 1970s. It caused massive disruption, and some of the displaced workforce was unable to cope with the change. However, it did not cause permanent, structural unemployment.
That isn’t to say that it might not be different this time, that the effects of short-term disruptions in the labour market might not be severe, or that automation cannot give rise to greater income inequality. However, the certainty of the doomsayers is unjustified. A recent survey of experts conducted by Pew Research demonstrates a wide variety of opinions on the matter.
Other research exists that contradicts the predictions of dramatic declines in employment. David Autor, a professor of economics at MIT, for example, writes that, “journalists and expert commentators overstate the extent of machine substitution for human labour and ignore the strong complementarities that increase productivity, raise earnings and augment demand for skilled labour”.
He bases this view on the belief that many human skills, especially subconscious ones, are not easily replicable by computers. He believes that improvements in machine learning, artificial intelligence and robotics are not as impressive and all-encompassing as tech celebrities and futurists would have you believe, and consequently, humans will continue to have economic relevance for the foreseeable future.
A Universal Basic Income (UBI) is often proposed as a policy response to the likelihood of either a short-term “creative destruction” of human jobs, or a long-term decline in employment. It seems to be more popular with the technologists that are bringing about all this automation than with economists. Perhaps this is because those who create automation technology want to avert a popular or regulatory backlash to what they’re selling.
The idea of a UBI is that instead of welfare payments and public services for the poor and unemployed, a new scheme should be created by which everyone – employed and unemployed alike – gets a set monthly payment from the government. If they’re unemployed and find a job, that wage will increase their total income, rather than replacing the government payout. A flat tax levied on all income other than the UBI can help fund such a scheme, as can cuts in existing social services.
On the face of it, a UBI has some benefits. One that will appeal to free market advocates (assuming they have resigned themselves to enforced charity towards the poor and unemployed in the first place) is that at least a UBI allows people to spend their welfare money as they wish. This will encourage more competition in traditionally government-controlled sectors of the economy, like low-cost housing, education and healthcare. As The Economist once wrote: “The most efficient way to spend money on the homeless might be to give it to them.”
Its other benefit is that it eliminates the “poverty trap” that occurs when unemployed people see little income difference between their unemployment benefits and a low-wage, full-time job. Because they keep a UBI even when employed, the incentive to work might actually be higher. Conversely, however, the permanence of a UBI, compared to the declining benefits paid under most unemployment schemes, counteracts this incentive. It is hard to say whether it will create better or worse incentives to work overall than existing systems.
There is little reason to believe that a UBI will be less costly to administer than existing welfare systems, or reduce fraud and wasteful spending in any way. It is also unclear whether it will increase or decrease the bargaining power of workers when it comes to wages. One one hand, it is easier to play hardball in wage negotiations if you know you won’t starve if you stand your ground and lose a job. On the other, it could drive down wages because there won’t be a “living wage” cudgel to beat companies with, and a UBI may increase competition for low-wage jobs.
A UBI also has significant downsides. One is that it doesn’t specifically address any of the concerns with automation. It doesn’t affect labour disruption, and frankly, doesn’t do much for the unemployed that ordinary unemployment insurance or welfare benefits wouldn’t do. It also spends a great deal on people who don’t really need it, which makes it a far more expensive system than existing welfare.
This also raises a philosophical problem. If it is morally wrong to take earnings from one person and give them to another, but people grudgingly accept it because it creates a safety net for the poor, will they continue to accept it if the recipients don’t even need the money? Would any charity be okay with helping everyone who turns up, whether or not they need the help?
A much bigger problem for UBI is affordability, however. Ed Dolan, a libertarian economist who advocates in favour of a UBI, has worked out what people would get if the US funds a UBI by cutting all entitlements, tax deductibles, housing subsidies and welfare services it currently pays except for healthcare (which poses specific challenges, including that no UBI will ever be enough to pay for the existing cost of US healthcare). It works out to a maximum of $4,452 per person per year, or $371 per month. In the US, the poverty threshold for a person living alone with no dependents is $11,490 per year (although it is lower for people in larger households). Either way, however, if you’re unemployed and receive only a UBI, you’re going to be extremely poor.
If even a rich country cannot afford a sufficiently generous UBI without dramatic tax increases on the rich, what chance does a country such as South Africa have? If South Africa were to spend its entire government budget of R1.56-trillion on a UBI, everyone would get R2,328 per month.
The largest state-sponsored experiment of a UBI started in January 2017 in Finland, but within six months it was a mess. If even a sophisticated developed country cannot get a pilot project right, what chance does a country such as South Africa have?
Either way, a UBI does not address the real concerns over job losses due to automation, even if the disruption is only temporary. If you do lose your job, a UBI probably won’t give you enough to pay your existing bills, or even just keep you fed, clothed and housed. It also won’t help you find a new job.
The likelihood of long-term structural unemployment due to automation is probably overstated. Either way, a UBI needs to stand on its own merits. As a solution to the risks of automation, whatever those might prove to be, it is simply the wrong tool for the job. DM