120 years ago American factories electrified their operations, triggering the Second Industrial Revolution in which steam engines were replaced with motors. This general purpose technology (GPT) – a technology that can affect an entire economy, usually at a national or global level – created new advantages for factories and prompted the invention of new work processes, allowing for increased growth and productivity. Innovation researcher Erik Brynjolfsson outlines three major GPT since the 18th century: the steam engine, electricity, and the internet, and along with Andrew McAfee, has coined this era The New Machine Age and produced a highly praised book of the same name. In this New Machine Era, they identify growing decoupling of productivity and employment: productivity is growing, but employment is decreasing. Correspondingly, wealth is increasing, but work is decreasing.
Ideas about brain emulations, or ems, set forth by economics professor Robin Hanson, provide one view of the future relationship between employment and increased technology, focusing on the role of robots. He suggests that in the future, truly smart robots will be created by emulating the brains of the ‘best humans.’ Those robots can then be multiplied to perform a certain job, which may lead to humans being displaced within a century. The professor further suggests that a new emulation economy will be very competitive, wages will fall to the cost of making new machines, and in doing so, fall to subsistence level.
As per usual, somewhat contradicting statistics have emerged as to the economic impact of increased use of robots. A paper from the Centre of Economic Performance at the LSE, ‘Robots at Work’, uses data from industries in 17 developed countries, concluding that ‘robot densification’ has a statistically negligible effect on total hours worked, while increasing both labour productivity and value added. This would suggest that robots are not, in fact, decreasing humans’ potential for employment. However, the paper also states that the use of robots reduced the hours of low-skilled workers, as well as (to a lesser extent) middle-skilled workers. This corresponds to the findings of Brynjolfsson and McAfee: of increased decoupling of productivity and employment, particularly for low/middle-skilled workers. This is also in line with professor Hanson’s predictions, depicting the future displacement of humans as not just a wild prediction but rather an actual possibility that we may be progressing towards.
In the past, the two industrial revolutions have given rise to structural unemployment – temporary unemployment caused by a mismatch between workers’ skills and the skills’ demanded by producers. However, with the arrival of robots, many fear that the unemployment will become permanent, technological unemployment. Despite debate regarding the imminence, gravity, or even existence, of the threat posed by this development, it does prompt us to rethink our current economic system (perhaps, at this point in time, only in theoretical terms).
Authors Gary Marchant, Yvonne Stevens and James Hennessy in the Journal of Evolution and Technology propose multiple policy options to tackle this problem, one of which is particularly striking as falling into the trap we must all be careful of: employing “legal interventions to protect jobs that might otherwise be lost due to technology, innovation, and development.” Such measures would, without a doubt, do more harm than good as they would stifle innovation, keep prices up, and, as twisted as it sounds, make efficiency illegal. Marchant et al. do recognise this, stressing their “counter-productive” nature and the unsuccessful track record of “stopping progress” strategies.
Another approach would be to re-evaluate the welfare system by introducing the Negative Income Tax (or Basic Income), which could not only address key problems of today, but also those that are underway with the changing labour market and the prospect of technological unemployment.
The current system creates perverse social incentives and features a piecemeal provision of benefits, consisting of working tax credits, means-tested housing benefits, and a number of other smaller benefits including council tax relief and skills training. On the other side of this is the Department for Work and Pensions, employing 34,000 people and spending 5.3 billion pounds on administration (in 2011-12 – 6.1 billion in today’s money). The system also disincentivises work due to high marginal withdrawal rates. Despite attempts to relieve these problems through Universal Credit, its design as a single monthly payment to the highest earner of the household has negative implications for women and children, in addition to the fact that it has not come close to reaching its target of 1 million claimants and has cost £612 million in administration costs. Hence, we are in need of a cheaper, administratively simpler and less dis-incentivising system – namely, a Negative Income Tax.
The Negative Income Tax is a policy that replaces all major means-tested cash benefits and guarantees a universal basic income by paying individuals directly. As a person earns more, the benefit is steadily withdrawn, ideally at a rate that will always make increasing work and reducing leisure worthwhile. An example of a NIT would be a £5,000 basic payment at a 50% marginal withdrawal rate (for every additional pound earned, the worker will receive 50p less in NIT payments). Someone with an income of zero would receive an NIT payment of £5,000, or just under £100/week. If they took a job that paid £5,000/year, they would receive a top-up of £2,500/year; if they took a job that paid £7,500, they would receive a top-up of £1,250/year. Once they reached £10,000/year, they would receive nothing in NIT and the citizen would become a net tax contributor. This policy enjoys substantial support from free marketeers and has been advocated by Nobel prize winning economists Milton Friedman and Friedrich Hayek.
The economic benefits of a NIT in this day and age are thus substantial, slashing bureaucratic costs, in addition to the political and social benefits of decreasing the size of government and promoting individual liberty. However, as unemployment rises, the political and social benefits brought about by a NIT will become even more important for society as a whole. The fact of the matter is: as more people become unemployed due to robot’s impact on the labour market, the strain on the welfare system will increase. That will, in effect, expose its inefficiencies and disenfranchising characteristics. Hence, a greater proportion of the population will be told how to live their lives through the government’s welfare schemes, which, as Sam Bowman puts it, “is a bad idea because the government is extremely ignorant, and that ignorance does not change just because the person being told what to do is on benefits.”
With a NIT, on the other hand, these problems can be counteracted. First of all, it would become the chief means by which consumption, the sole end and purpose of all production according to Adam Smith, can be kept alive, without the dictation of the government as a middleman. Similarly, those who fear that the ‘capitalist exploiters’ who own the robots will own all the wealth will be happy to realise that only 3% of the benefit of Schumpeterian profits (profits arising from innovative activity) goes to the entrepreneur: the other 97% goes to the consumer. An NIT would thus ensure that also the unemployed would benefit from these advancements by giving them the means to continue to be autonomous consumers. Ideally, the increased availability of robots, and increasingly perfect information due to other technologies, will make more and more markets perfectly competitive, again, much to the benefit of the consumer as opposed to the capitalist robot-owners. Furthermore, with the eventual material abundance that increased robot productivity will bring about, the basic income will be increasingly fundable, as prices will fall significantly and the ‘real’ basic income will increase.
With the rise of technological unemployment, the transformation of society, and diminishing of the labour market, the changes to the welfare system in response to these changing circumstances will become decisive. Do we want to move towards an ever bigger government, as more people become dependent on the welfare system, losing their power as consumers as their needs are taken care of by a paternalistic state welfare scheme? Or, do we want an administratively simple system that gives the unemployed the freedom to enjoy the material abundance and fruits of centuries of intellectual labour, reflected in lower prices brought about by the robots’ increased productivity? I think the latter is infinitely more attractive.
However, accurately predicting the future is tricky, so the future effect of robots remains an intellectual game. As the study at LSE showed, ‘robot densification’ has had a statistically negligible effect on total hours worked in 17 developed countries. Furthermore, using the UK’s 30 million jobs as an example, 10% of jobs are destroyed every year, yet the economy also creates about 3 million. In fact, the majority of jobs are created by small, new companies – not the large ones that would employ robots. Hence, another solution would be to encourage small businesses, through deregulation, to enter markets and thus generate jobs for which they would employ humans (that is, assuming that robots have not become more accessible than human labour). Furthermore, as we have seen with the invention of past GPTs, the increased productivity has increased wealth, which has, in return, generated more jobs. Hence, we may simply be looking at prospective structural unemployment that will, with time, be solved with the creation of jobs unthinkable to us today, just as the idea of a computer scientist was in the 19th century.
As mentioned earlier, professor Hanson predicts that robots will displace humans in most jobs within a century. However, history suggests that new, some unimaginable, opportunities for employment will arise following such a development. Whichever prediction one chooses to believe, two things can be concluded. With the rise of robots, the lower input costs will make products cheaper and we will all become richer. It might even, as Keynes prophesised, solve the fundamental economic problem of scarcity that has characterised our endeavours as humans. Second, the effects of this technological progress on employment may be (and according to Brynjolfsson and McAfee already has been) negative, whether that be permanently or until new jobs emerge. Some version of a Negative Income Tax might be the ideal way to protect those affected by this transition, enjoy our progression towards material abundance and free ourselves to make our own decisions in such a world.
By Halla Mathiesen