The Future before your Eyes

What happens as the few workers we see in these videos below are no longer needed? And, even more importantly, when middle class and professional jobs get hit by the same trend through AI and more sophisticated software?

Capitalism has both a supply-side and demand-side. As more and more work is done by machines or software, the relationship between aggregate demand growth and private sector employment growth will start to break down – or at the very least become very weak. Eventually, a government will not be able to stimulate aggregate demand as an effective solution to unemployment, because this will tend to cause more use of machines, not human beings.

The solution to this is obvious: governments need to do much more to manage the demand-side. They need to find economically and socially useful work that can still be done by human beings and also move towards a guaranteed basic income. If you want income above the guaranteed basic level, you will still need to do some work of value to human society and civilisation (e.g., science, medical and technology research, helping development in the third world, human social services etc.).

Would Capitalism necessarily be destroyed if Human Labour fell towards Zero?

In a word: no.

Why? Because an economy with more and more automation based on private enterprise and private capitalist production could still sell its output and obtain money profits, if a government managed the demand-side of the economy by providing a guaranteed income (with, say, taxes on consumption, property, and ownership of financial and real assets and returns from those assets, with the shortfall covered by central bank money creation). As long as the balance of payments functioned successfully, a type of capitalism could continue.

That is, such an economy would still be a variety or type of capitalism: it would not be a command economy or the type of socialism envisaged by Marx.

Any Marxist response to this depends on Marx’s definition of capitalism. If one wants to define capitalism merely as a system of private production based on employment of free human wage labourers, then of course capitalism would cease once employment of free human labour ceased.

But this is just playing with words: setting up a narrow analytic definition of capitalism (true by definition), and ignoring other obvious real world aspects of capitalist systems of production.

If capitalism is to be defined in any empirically-defensible sense, it would need to use the following criteria:

Capitalism is a system of production as follows:
(1) where the vast majority of all capital goods are owned privately and where there is a high degree of private property (in land, houses, private possessions, etc.) and rights to private property;

(2) where the vast majority of all decisions on investment and production of commodities are made by private agents (though this does not exclude certain public goods);

(3) where there exists a class of free human beings who work for a wage, either from the private or public sector (though mostly in the private sector).

Now if (3) fell and fell or even ceased to happen in an economy where production is increasingly done by machines, then it would still leave us with criteria (1) and (2).

The new system of nearly fully or fully automated production would in essence still be a type of capitalism, because it would still have traits (1) and (2), which clearly lie at the heart of what capitalism is.

In short, it would not be a system where all business is owned by the state or where the state plans all economic activity, and there is no necessary reason why a capitalist mode of production must end even if human wage labour falls towards zero.

Is a Shrinking Labour Force such a Problem?

It may not even be a problem at all.

That is, in light of the massive revolutionary effects of automation and robotics that is happening as we speak, as we can see here, here and here.

It is potentially so revolutionary it is rightly being dubbed the “fourth industrial revolution.”

The warning signs have been here for years: e.g., a 2013 Oxford study suggests that about 47% of human jobs will be lost to AI by 2030. First, working class and then middle class jobs will be hit.

When, for example, machines can do the work of people in fast food restaurants, what sort of world will result for low-skilled and semi-skilled workers? Even offshoring will become pointless, if production by machines costs less than third world labour.

And don’t trust halfwit neoclassical economists to provide sensible analysis here, because their response is almost always that only if you make wages become flexible and rely on laissez faire market theology, then this will eliminate involuntary unemployment. That is rubbish. There is no reason to think a fantasy-world self-equilibrating market will solve the problem of mass unemployment, because it does not exist.

At the very least, all this suggests that the endless hysteria about shrinking working-age labour forces in the West is grossly exaggerated. If anything, fewer workers mean less of a problem with mass unemployment in the future. More likely, there will be plenty of people to look after the elderly and do whatever necessary work is left to people.

Automation and Robots in the News

Interesting news on the rise of automation and robots both in the West and in China:

Paul Davidson, “More robots coming to U.S. factories,” USA Today, February 10, 2015.

Georgina Prodhan, “China to have most robots in world by 2017,” Reuters, February 6, 2015.

Paul Wiseman, “Robots Replacing Human Factory Workers at Fast Pace,” Cio Today, February 22, 2015.

While only about 10% of manufacturing tasks in American factories are automated today, some think that this will rise to about 25% in about 2025, a decade from now. But, more importantly, what will the percentage be in 2035 and 2050? Presumably much higher.

The upside of this will be increased productivity and what has been dubbed “reshoring,” or the return of manufacturing to the Western world. If this happens on a large enough scale in the long run, that in turn means lower trade deficits and a larger manufacturing sector and output for nations with mass consumer markets in North America and Europe.

But we can’t see the full effects of this today and the downsides, partly because robotics has been – for a long time – an overrated field with many robots being just expensive frivolous toys. All that is changing, as anyone reading the news can see.

A lot of neoclassical economists and those under their influence see no problem with mass automation, but that is only because neoclassical economists is deeply flawed and mistaken in its core principles.

Economies run on orthodox neoclassical theory are likely to have chronic problems of insufficient aggregate demand and mass structural unemployment as automation in production soars and even service and white collar work can be done by artificial intelligence (AI).

Market economies have no effective and reliable tendency to full employment equilibrium, and there is no necessary reason to think that the issue of structural unemployment will be solved by markets.

Worse still, with the fall in prices and factor input costs, possible general price deflation could put downward pressure on wages in the future, which means debt deflationary problems as goods prices, wages, nominal debt and asset prices are grossly distorted in relation to one another.

In the late 19th century from 1873 to 1896, for example, there was a period of chronic deflation, probably caused to a great extent by positive supply shocks and technological advancement, but the effects on investment, confidence and the economy at large were deleterious (the evidence can be seen here, here, and here).

Economic policies will need to change in the future to reflect the realities of production by mass automation.

Government welfare will have be reconsidered, not as some safety net, but as a basic human right in an age of automation: in essence, everyone will need a basic guaranteed income from the state.

If you wish to work in addition to this, as no doubt many – and probably most – people will, and you cannot find work in the private sector, the traditional policy of Keynesian fiscal stimulus will become weaker and weaker in its effects as more and more work is done by machines and artificial intelligence.

Eventually, governments will need direct mass employment programs to create economically and socially useful work, e.g., in social services where real people can never really be displaced, and in education, sciences, research and development, or other services.

Perhaps I just haven’t read the literature well enough, but I am surprised that Post Keynesian economists haven’t taken these issues more seriously in their academic writing.

Automation and Future Unemployment

The impact of cheaper and more sophisticated forms of automation has recently been examined in this report by two researchers from Oxford University’s Programme on the Impacts of Future Technology:

Carl Benedikt Frey and Michael A. Osborne, “The Future of Employment: How Susceptible are Jobs to Computerisation?,” September 17, 2013.

Some summaries of the report can be read here:

Josh Lieberman, “Half Of U.S. Jobs Are At Risk Of Computerization: Which 10 Are The Least Safe?,” Isciencetimes.com, September 25, 2013.

Mark Hoffman, “Study: Half of All US Jobs Could Be at Risk of Computerisation,” Scienceworldreport.com, September 27, 2013.

“Nearly half of US jobs could be at risk of computerization, Oxford Martin School study shows,” September 19, 2013

Richard Eskow, “The Robots Are Coming – Now What?,” September 26, 2013.

Michael Snyder, “Oxford Professors: Robots And Computers Could Take Half Our Jobs Within The Next 20 Years,” September 29, 2013.

Their main conclusion is that 47% of US jobs might be lost owing to automation in twenty years, with similar trends in many other nations.

Also, they conclude:

“Our model predicts that most workers in transportation and logistics occupations, together with the bulk of office and administrative support workers, and labour in production occupations, are at risk. These findings are consistent with recent technological developments documented in the literature. More surprisingly, we find that a substantial share of employment in service occupations where most US job growth has occurred over the past decades …, are highly susceptible to computerisation. Additional support for this finding is provided by the recent growth in the market for service robots ... and the gradually diminishment of the comparative advantage of human labour in tasks involving mobility and dexterity.”
Carl Benedikt Frey and Michael A. Osborne, “The Future of Employment: How Susceptible are Jobs to Computerisation?,” September 17, 2013. pp. 43–44.

And also of interest was this part of the paper:

“Our paper is motivated by John Maynard Keynes’s frequently cited prediction of widespread technological 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’ (Keynes, 1933, p. 3). Indeed, over the past decades, computers have substituted for a number of jobs, including the functions of bookkeepers, cashiers and telephone operators … . More recently, the poor performance of labour markets across advanced economies has intensified the debate about technological unemployment among economists. While there is ongoing disagreement about the driving forces behind the persistently high unemployment rates, a number of scholars have pointed at computercontrolled equipment as a possible explanation for recent jobless growth.”
Carl Benedikt Frey and Michael A. Osborne, “The Future of Employment: How Susceptible are Jobs to Computerisation?,” September 17, 2013. pp. 43–44.

While it is absurd to deny that the surge in unemployment across the Western world since 2008 has been fundamentally caused by the aggregate demand shocks stemming from the Great Recession, nevertheless there would appear to be underlying structural unemployment problems caused by automation as well.

Market economies have no tendency to full employment equilibrium, and there is no necessary reason to think that the issue of structural unemployment will be solved by magic market solutions.

And there are also other issues: with the fall in prices and factor input costs, possible commodity deflation could put downward pressure on wages in other industries, which means debt deflationary problems as goods prices, wages, nominal debt and asset prices are grossly distorted in relation to one another.

The solution to these problems is: full employment macroeconomic policies, maintaining of a basic level of income for all (especially the unemployed), and policies to detect and deal with debt deflation.


BIBLIOGRAPHY
Keynes, J. M. 1933 [1930]. “Economic Possibilities for our Grandchildren,” in John Maynard Keynes, Essays in Persuasion. Macmillan, London. 358–373.

Frey, Carl Benedikt and Michael A. Osborne, “The Future of Employment: How Susceptible are Jobs to Computerisation?,” September 17, 2013.

Will Hutton on the Future of Automation

A thought-provoking article by Will Hutton in the Guardian/Observer on the consequences of automation on employment:

Will Hutton, “Driverless Cars, Pilotless Planes … Will There be Jobs Left for a Human Being?,” Guardian, Sunday 19 May 2013.

It is strange that there is no mention of the need to maintain aggregate demand, but otherwise the speculations at the end make interesting reading.

I am inclined to think we are only seeing the the small effects of these processes now, and the really huge changes will be more a medium to long term problem, something that will become acute from 2030s or 2040s, though that is just a wild guess.

A final point about energy technology: I suspect many people think of nuclear fusion as the promise that bounced, but recent news about the international nuclear fusion project should give people pause. Unfortunately, at a conservative estimate commercial fusion reactors might not be around until the 2050s.

Rise of the Robots?

Yes, this is already the title of a recent post by Krugman here:

Paul Krugman, “Rise of the Robots,” December 8, 2012

I have to say I feel rather vindicated, since I made similar points over two years ago in this post:

“Automation and Robotics: The Future of Manufacturing?,” September 12, 2010.

Krugman notes that the strong trend towards industrial automation is good news for American manufacturing. Indeed, it is, and we can expect to see some return of manufacturing to the US and Western nations from East Asia and other developing, low wage countries. As Krugman says, robots “mean that labor costs don’t matter much, so you might as well locate in advanced countries with large markets and good infrastructure.” This means that the US should see a fall in its unbalanced trade deficit with other nations.

But what about the possible problems it will cause? Firstly, I want to qualify everything I say below with the clear statement: I do not oppose the trend to automation; I welcome it. Labour saving machinery was the key to the industrial revolution and the many increases in productivity growth we have seen over the past two centuries. The problem is that a new era of mass automation requires macroeconomic policies to deal with its perverse negatives consequences.

The media already reports that some economists think mass automation might not necessarily benefit everyone, but the analysis there is poor because it is based on the fantasy of neoclassical general equilibrium models.

In contrast, Krugman focuses on the implications for income distribution: away from workers to capital. First, the wealthy have a lower marginal propensity to consume (or alternatively a higher marginal propensity to save), and they are likely to buy financial assets on secondary markets with their extra money, not final goods and services.

Secondly, America and many Western nations already have a chronic unemployment problem, which, without proper fiscal expansion, will continue for years. What will happen when that persistent unemployment problem is exacerbated by large-scale structural unemployment owing to mass automation in industry?

Factors (1) and (2) above already strongly suggest a serious aggregate demand shortfall in the future. Keynesian demand management and income distribution are going to be more important than anyone ever dreamed. Government must step in and provide employment programs or funds to employ those who are unemployed and who seek employment. (In fact, structural unemployment induced by technology was already a problem in the 19th century and even as strong a defender of laissez faire as Jean-Baptiste Say – the inventor of Say’s law – advocated public works as a solution to such unemployment. See Appendix 1 below.)

No doubt additional jobs will be created in new private businesses, but it is unlikely to be enough. Free markets do not guarantee full employment, nor does Say’s law work. Employment in tradable goods and services in many countries will probably fall dramatically. Our employment future will probably be mainly in services, education, and most probably in employment programs funded by government or in government-sector jobs. There will probably be a great reduction in the hours that people need to work as well and more leisure.

It might well be that much of the government-funded labour force will be in education (e.g., universities), sciences, research and development, or other services. I suspect a much greater labour force working in basic sciences and applied R&D would mean a much more rapid advancement of science and technology too – a virtuous circle.

As we move down the route of radical automation in the course of this century, equally radical Keynesian demand management will be necessary to maintain demand for goods and services, income equality, and continuing rises in living standards.


Addendum
Some other points that occur to me as an afterthought:

(1) Eventually automation and structural unemployment will affect even services. Artificial intelligence is not far fetched, and it is undoubtedly the way of the future. Already when you ring some companies you can find voice recognition software doing the work of people.

(2) Perhaps we will see a very strong deflation in the prices of many industrial goods, especially those where at present high wages are a major factor input cost.

(3) Some more relevant news and commentary here:

Paul Krugman, “Robots and Robber Barons,” December 9, 2012.

Christopher Matthews, “Can Robots Bring Manufacturing Jobs Back to the U.S.?,” Time.com, September 27, 2012.


Will Knight, “This Robot Could Transform Manufacturing,” MIT Technology Review, September 18, 2012.

Vivek Wadhwa, “The End of Chinese Manufacturing and Rebirth of U.S. Industry,” Forbes.com, 23 July, 2012.

APPENDIX 1: SAY ADVOCATED PUBLIC WORKS AS A SOLUTION TO UNEMPLOYMENT INDUCED BY TECHNOLOGY

In his discussion of the introduction of labour saving machines, Say recognised that this would create short term unemployment, and in a footnote actually advocated public works spending by government:

“Without having recourse to local or temporary restrictions on the use of new methods or machinery which are invasions of the property of the inventors or fabricators a benevolent administration can make prevision for the employment of supplanted or inactive labour in the construction of works of public utility at the public expense as of canals, roads, churches or the like …” (Say 1832: 87).

This must come as a shock to Austrians and libertarians who so frequently cite Say’s law with approval.


BIBLIOGRAPHY

Say, J. B. 1832. A Treatise on Political Economy; or, the Production, Distribution, and Consumption of Wealth (4th edn; trans. C. R. Princep and C. C. Bibble), Grigg & Elliott, Philadelphia

Automation and Robotics: The Future of Manufacturing?

First, let me offer a caveat: this post contains some speculative musings of mine on the future of manufacturing. No doubt various criticisms of it could be made.

Countries like the US and the UK are badly in need of trade and industrial policies to rebuild manufacturing. Very large trade deficits are potentially unsustainable. Such deficits often make a country dependent on foreign investment for the capital account surpluses needed to pay for current account deficits.

The crucial factor now, however, is that technology must be used to increase manufacturing productivity and cut costs.

If we want to decrease the trade deficits of the US or the UK, I would suggest an industrial policy to domestically manufacture things imported from China and East Asia.

Strong use of automation and technology to increase productivity and to lower price is necessary. This process can be made faster and more efficient through public R&D programs, and state transfer of new technology to domestic manufacturers.

In an earlier post, I drew attention to a very interesting initiative in the US called the “Save Your Factory movement,” launched by a company called Fanuc Robotics America Inc.

There is an absolutely excellent analysis of this in a 2005 issue of Manufacturing Engineering magazine. It shows how automation can cut costs and even beat low wage countries like China::

Rick Schneider, “Robotic Automation can cut costs,” Manufacturing Engineering 135.6 (December 2005): 65–72.

The US federal government needs to take up these ideas and implement this sort of policy at a federal level – which would make it more effective.

Moreover, the article cited above points out that from 1995 to 2002 the global labour force actually lost 22 million manufacturing jobs because of labour-reducing productivity gains through automation and robotics.

I would argue that it is extremely likely that the 21st century will see manufacturing employment as a percentage of the world labour force decline to a level as low as agricultural employment in most developed nations (2 or 3%).

Will this be a bad thing? Not necessarily. If output massively increases, prices are much lower and Western current account deficits fall or go into surplus, this will be a very good thing, and we will have an abundance of cheap goods.

But we will have to face the fact that, because of automation and technology, employment in tradable goods and services in many countries will probably fall dramatically. Our employment future will probably be mainly in services, education, and most probably in government-sector jobs or employment programs funded by government. There will probably be a great reduction in the hours that people need to work as well.

No doubt additional jobs will be created in new private industries as well, but government can step in and provide employment for those who are unemployed. It might well be that much of the government-funded labour force will be in education (e.g., universities), research or other services. A much greater labour force working in basic sciences and applied R&D in physics, chemistry, geology, biology, genetics, engineering and medicine would mean a much more rapid advancement of science and technology too – a virtuous circle.

In other words, in the face of massive productivity and output gains and cost reductions in many goods and services through technology, the government must use policies for full employment to maintain demand for such goods. The point is that should production go down the route of radical automation in the course of this century, then equally radical Keynesian demand management will be necessary to maintain demand for goods and services and ensure continuing rises in living standards.