6

Globalisation or automation: what killed the jobs?

[Globalization] has lifted more people out of poverty around the world than probably any other advance in recent history. It has certainly helped expand our prosperity here at home … So there is no escaping this, and what we have to do in America today is figure out how to best harness the force of globalization to ensure a continuing quality of life and standard of living.

Hillary Clinton, speech in Silicon Valley, 1 May 2016

Surprise, surprise. Workers in Britain, many of whom have seen a decline in their standard of living while the very rich in their country have become much richer, have turned their backs on the European Union and a globalized economy that is failing them and their children … That increasingly globalized economy, established and maintained by the world’s economic elite, is failing people everywhere … We need to fundamentally reject our ‘free trade’ policies and move to fair trade.

Bernie Sanders, New York Times, 27 June 2016

Globalization has made the financial elite who donate to politicians very, very wealthy … But it has left millions of our workers with nothing but poverty and heartache.

Donald Trump, speech in Monessen, Pennsylvania, 28 June 2016

Hillary Clinton was trimming her sails by midway through the 2016 election campaign, because her views on free trade were clearly not playing well with many traditionally Democratic voters, but she could neither abandon her fundamental faith in globalisation nor deny her long connection with the policy. (It was her husband, Bill, who signed the North American Free Trade Agreement in 1993, although it was negotiated by his predecessor, Republican president George H.W. Bush.)

Bernie Sanders, who gave Mrs Clinton a good run for the presidential nomination, had figured out long before that many if not most Democratic voters were looking for a better answer to their economic plight than more of the same. He was dead wrong about Britain, where globalisation barely got mentioned in the Brexit campaign, but his rhetoric resonated strongly with Americans who were sick of being told how wonderful globalisation was while their own incomes stagnated and ‘blue-collar’ jobs disappeared. Trump was saying the same thing but in a punchier way, with promises of a wall on the Mexican border and a 45 per cent tariff on imports from China.

It’s hardly surprising that globalisation got so much emphasis in the US election, because both of the competing explanations for what had gone wrong were off-limits for at least two of the leading three contenders. None of them wanted to talk about automation, because if you have no plausible solution to a problem, it’s safer not to bring it up at all. Clinton and Trump were also uncomfortable condemning the neo-liberal ideology that both of their parties had espoused unquestioningly for at least a quarter-century: Sanders, as a self-declared ‘democratic socialist’, was the only one willing to talk explicitly about the damage that neo-liberalism has inflicted on working-class living standards. By default, therefore, globalisation became the alleged prime culprit for the profound unhappiness and disaffection of the ‘white working class’. That was bound to favour a Trump victory — and when Trump says ‘globalisation’, he mainly means free trade.

I discussed the role of free trade in driving down American wages in the previous chapter. Here I look at it as an actual killer of jobs, and that is a different question. Was there a tidal wave of cheap imports that drove American companies making the same goods out of business? This certainly did happen in the 1970s and 1980, particularly to the American textile industry: only exclusive branded clothing is now made in the United States, and not much of that. The other 98 per cent of the clothes that Americans wear are made abroad. But the direct impact of free trade on American jobs has been quite limited, because the United States is the least trade-dependent of all the developed economies.

In 2015, trade accounted for only 27 per cent of US Gross Domestic Product, compared with an average of 56 per cent for all OECD countries and 84 per cent for the German economy.1 More than in any other developed country, the American economy is driven primarily by domestic factors, with imports and exports playing a relatively minor role. Free trade has not been the great job-killer — and neither has unchecked immigration, another of Mr Trump’s favourite themes.

The United States admits about one million immigrants each year. Illegal immigration was once a major problem, but it has been reduced by nine-tenths over the past fifteen years. In 2015 the total was just one hundred and seventy thousand people, and it has since fallen further. Even counting the illegals, the United States receives only one new immigrant annually for every 282 people already in the country.

Canada, the country most similar to the United States in its economy and demography, admits three hundred thousand immigrants a year. Since its population is just over one-tenth of that of the United States, it is admitting immigrants at more than twice the American rate: one new immigrant annually for every 116 people already in the country. Yet there is no anti-immigrant outcry in Canada, no fear that the country is being overrun, and no labour-union opposition to the immigration policy. To the extent that American workers buy the idea that immigrants are depressing their wages or outright stealing their jobs, they are scapegoating the wrong people. It is the neo-liberal economic order (including free trade) that is depressing their wages, to be sure, but their job losses are due to one or the other of two radically different phenomena: computers, or the specific aspect of globalisation known as ‘offshoring’.

Back in the late 1970s, when ‘offshoring’ began in a significant way, there was a direct substitution of foreign labour for American labour in industries such as textiles and furniture-making, and job losses were easily measurable. As the phenomenon gained speed through the 1980s and 1990s, there were still clearly identifiable cases where a factory shut down in the United States and re-opened again in Mexico or even China in order to exploit the far lower wages that prevailed in those developing countries. And in the grand old tradition of buccaneering capitalism, some masters of industry publicly flaunted their disdain for mere human considerations such as the jobs of fellow Americans.

Ideally, you’d have every plant you own on a barge to move with currencies and changes in the economy.

Jack Welch, CEO, General Electric, 1981–2001

But as we move forward through the 1990s, the picture gets muddier. Lots of IT jobs are being created abroad to serve American business (accountancy, billing services, call centres, and the like), but it is not clear that they would otherwise have become or remained jobs in the United States. Would those jobs have been done by Americans if there had been no foreign alternative, or would the companies just have invested in more automation because that would still be cheaper than paying Americans to do them? Some whole factories were still moving abroad, too, but once again there may have been a trade-off between more expensive automation at home and just moving the factory unchanged to a place with lower labour costs. What this means is that all estimates of how many jobs were ‘lost’ are inherently unreliable, but at a rough guess the United States lost a couple of million jobs to offshoring in the later 20th century out of a workforce (in 1979) of 104.6 million. Many more jobs have disappeared in the early 21st century, particularly in manufacturing, but few of them appear to have been exported.

In 1979 there were 19.6 million manufacturing jobs in the United States. Contrary to popular belief, that number only declined slightly in the next twenty years, remaining around 17.5 million right through the 1990s. Then it collapsed, and by 2009 (just after the financial crisis of 2008) it was down to 12.4 million. It dipped even further at the depth of the Great Recession in 2010–12, but had recovered to 12.4 million by 2017.2 So something very big happened to American manufacturing in the first decade of this century, but it wasn’t offshoring.

The key to this puzzle is the fact that American manufacturing jobs have declined by one-third in the past forty years, but American industrial production is now approximately twice what it was in 1979.3 This can only be explained by much higher productivity per worker, and the presumption must be that the higher productivity is mostly due to automation, which transformed the way work was done in factories in the first decade of this century.

As Ball State University economics professor Michael Hicks observed in 2015, ‘Had we kept 2000 levels of productivity and applied them to 2010 levels of production, we would have required 20.9 million manufacturing workers. Instead, we employed only 12.1 million.’4

That’s where the jobs went, or most of them. Out of 7 million manufacturing jobs lost in the United States in the past forty years, at least 5 million were lost to automation. Then the manufacturers moved a lot of the remaining industrial jobs not abroad, but south and west to the ‘Sun Belt’, where unions are weak and the sun always shines, leaving behind the Rust Belt — which eventually gave us Donald Trump. The people whose incomes had been relentlessly squeezed for decades, and the people whose jobs had simply disappeared, finally struck back in 2016. Their aim wasn’t very good, but their anger was entirely justified. They have been patient for a very long time.

Every time I drive through Detroit, it seems, they are tearing another skyscraper down: it’s been empty for a decade or two (or three), the water has got in, bits are falling off, and they are dismantling it before it collapses into the street. That saddens me, because Detroit has the best Art Deco skyscrapers in the world, but I have no desire to engage in what is rightly condemned as ‘ruin porn’. It’s just a big city that has been having a hard time for a long time (as great cities sometimes do), but I choose it because it is a useful thread to lead us through the next part of the story: it is Motor City, the heartland of American manufacturing since the time of Henry Ford.

The initial downturn in Detroit’s fortunes, after two generations of soaring prosperity — it still has one of the finest art galleries in America — was driven not by automation but by inequality: the injustices suffered by hundreds of thousands of African–Americans who had migrated north to work in the factories during the Second World War.

People who have a stake in their society, protect that society, but when they don’t have it, they unconsciously want to destroy it.

Martin Luther King Jr.

The burning of much of Detroit’s inner city in 1967 marked the beginning of a period of continuous decline that was steeper than in any other major American industrial city: Detroit’s population has fallen by 60 per cent since 1950. Part of it was ‘white flight’ after the riots, and part was just a general drift to the suburbs; but from the mid-1970s on, wages ceased to keep up with rises in productivity for American workers, and from the early 1990s a great many jobs began to disappear in the automobile industry. Much the same was happening in other industrial cities, but the damage was most extreme in what was becoming known as the Rust Belt, and in 2016 the voters of that region delivered their verdict on the management of the American economy over the past generation. They elected Donald Trump.

No need to panic. If it’s just a case of falling wages, that can be fixed, though Trump is probably not the person for the job. Wages can be tied to productivity again, income inequalities can shrink as well as expand, the anger can be soothed, and the system can be stabilised. It’s what democratic politics does best, in fact. There’s a kind of self-correcting feedback mechanism built into democracy: it operates over long time scales, which leaves a lot of room for ugly accidents, but it usually gets there in the end. Americans are beginning to digest the lessons of 2016, and it’s likely that the platforms of the major parties will have taken account of the new political and economic realities by the next presidential elections in 2020. The anger level would then begin to go down — if it were not for the continuing rise of automation. Car-making was one of the first major industries to be remade by it, but automation is not finished with the Motor City.

For the past three generations, manufacturing vehicles has been the largest single industrial occupation on the planet. The global motor industry produced 95 million cars and commercial vehicles in 2016, up from 66 million only ten years before, and it directly or indirectly employed some 50 million people. If car-making were a country, it would have the world’s sixth-biggest economy. And it’s already pretty automated.

Assembly-line work was a prime target for early computer takeover, because it broke down the task of building a car, for example, into several hundred much simpler tasks that could each be done by a person who didn’t require specialised skills. Some of the people on the assembly line were just attaching door handles and the like to each car as it trundled down the track, one on the left side and one on the right. Writing the program that would enable a robot arm to do that work instead was cheap and easy. Other jobs on the line required more skill, but with a little more investment you could write programs for them, too. In a few years the human population of that line dropped by 80 or 90 per cent — and most of the humans who remained were basically monitoring the machines doing the actual job. The same thing was also happening in all the other, smaller factories back up the supply chain where the headlight assemblies, the little electric motors, and all the other components that come together on the final assembly line were made.

Automation is an almost irresistible temptation for management because, from an employer’s point of view, robots are much preferable to human workers, even if they cost the same or more. The robots don’t take lunch breaks and toilet breaks, they’ll work all day and all night, and they don’t have to be paid overtime (indeed, they don’t have to be paid at all), they don’t demand sick leave or annual holidays, they don’t join trade unions, and they don’t sue you if they get mangled in some industrial accident. Not only that, but they are accurate, tireless workers (no more ‘Monday morning cars’), and they generally do work out cheaper than humans in the end, despite their large up-front capital cost. What’s not to love?

This wave of automation largely accounted for the huge fall in employment in manufacturing in the United States in the first decade of this century. Comparable numbers of new jobs were not created elsewhere, so the long-term unemployment rate went up permanently (although, as noted previously, much of this new unemployment was not regularly recorded). Even when people eventually did find new jobs, the number of months they spent without work while they were looking doubled in that decade. But then there was a lull.

The number of Americans working in manufacturing stabilised for most of the ‘teen’ years — and some of the factories that had been offshored in earlier years even began to be ‘reshored’ back to the United States, because so few people were now needed in the factories that it didn’t matter all that much that American workers had to be paid more than Mexicans or Chinese. The convenience of having the production done at home, in a more predictable legal and political environment, outweighed the cost in higher wages. But the next round of job losses is about to begin. Let’s stay with the car industry.

France and the United Kingdom recently announced that they will ban the sales of petrol- and diesel-engined cars from 2040. From then on, only zero-emission vehicles may be sold. India says it will institute a similar ban by 2030. China, the world’s largest producer of cars — 28 million vehicles last year, more than the United States, Japan, and Germany combined — is also planning to declare a ban soon, but is still working on the cut-off date. And the European Commission is planning to impose a minimum annual quota of electric vehicles (EVs) for all European car producers.

So if you were looking for a safe place for a long-term investment, you probably wouldn’t choose the oil industry. Just over half of the 98 million barrels of oil produced in the world each day go directly to making petrol, used almost exclusively in motor vehicles. Another 15 per cent go to make ‘distillate fuel oil’, of which at least half is diesel fuel. So around 58 per cent of total world oil production is being used in vehicles now. There may be almost none in 35 years’ time. Britain, for example, is planning to allow only zero-emission vehicles on the road (apart from a few specially licensed vintage cars) by 2050, only ten years after the ban on selling new cars with internal-combustion engines comes into effect. In practice, if these deadlines are observed, the cars on sale will be almost entirely EVs by the mid-2030s — and what’s left of the oil industry will have a very different shape.

Countries that export most of their oil, such as Russia and Saudi Arabia, will find their incomes crashing for two reasons: sheer lack of demand, and very low prices ($40 per barrel or less) due to the huge glut of productive capacity. This may have major political consequences. Countries with some oil production of their own, such as the United States and China, will probably stop importing oil entirely. The US federal government will remain in the last ditch federally so long as Donald Trump is president — he’s even trying to revive the coal industry — but eight states have already signed an agreement to have 3.5 million zero-emissions vehicles on the road by 2025. Millions of jobs will be lost in the oil industry as global production falls by at least half, although significant numbers of new jobs may be created in the renewable-power industries as electricity demand ramps up. But none of this has directly to do with computers and automation.

Some jobs will be lost in the car industry because electric vehicles are much simpler to manufacture, having many fewer moving parts than the old internal-combustion engines, but that’s a small price to pay for the environmental benefits the EVs will bring with them. In any case, it’s not really an automation issue, either. What will kill far more jobs, in the car factories and far beyond them, is the advent of self-driving cars, which most certainly is an automation issue.

Driverless vehicles will end up being ownerless vehicles. They will become public utilities, summoned when they are required for the specific trip you have in mind at the moment. Urban car clubs and peer-to-peer rentals are one precursor of this phenomenon; Uber and Lyft, in their different ways, are another.

Privately owned cars are parked an average of 95 per cent of the time. This figure varies little from one city or country to another, and illustrates why private car ownership will become a dispensable luxury. The difficulty in the past was gaining immediate access to a car for as long as you needed it at a reasonable cost unless you actually owned one, but the combination of the smart phone and the self-driving vehicle will solve that problem.

This, rather than a cheaper taxi service, is the real goal of Uber’s business model, but once reliable self-driving cars are widely available, Uber will find itself deluged with competition. Private ownership will decline steeply: a recent KPMG survey of car-industry executives found that 59 per cent of CEOs believe that more than half of today’s car-owners will no longer want to own a car by 2025.5 In fact, there are hardly ever more than a quarter of privately owned cars being driven at the same time, so the total number of cars on the road worldwide may eventually crash to around one-quarter of the current number. That would eliminate most of the remaining jobs in the car-making industry.

The change goes wider than this, of course. Conventional buses and taxis will virtually disappear, taking millions of driving jobs with them. (There are a million taxi, Uber, and bus drivers in the United States alone.) Long-distance truckers and van drivers (another 3.5 million in the US) will also find work increasingly scarce: Daimler, Volvo, Uber, and Baidu are already road-testing the first self-driving 18-wheelers. And the changes still aren’t finished.

About a quarter of the average central city in North America (less in Europe and Asia) is devoted to surface parking lots and multi-storey garages. They are part of the 95-per cent-parked problem. The car doesn’t just take you downtown; it has to stay there the whole time you do, so it must be parked somewhere. Once people realise that most of this land is now available for redevelopment, it will get a lot easier and cheaper to live downtown: less commuting, more community. There may be quite a lot of new jobs in construction for a while as a result, but they won’t be remotely comparable to the number being lost in the auto industry as a whole, and they won’t last.

The same calculations for job losses apply elsewhere. Klaus Schwab, who runs the World Economic Forum (the annual shindig for the ultra-rich at Davos in Switzerland each January), illustrates the phenomenon vividly in his book The Fourth Industrial Revolution. (Never mind the ‘Fourth’ — it’s just a conceit.) In 1990, Schwab points out, the three biggest companies in Detroit, all car-makers, had a market capitalisation of $36 billion, annual revenues of $250 billion, and 1.2 million employees. A quarter-century later, in 2014, the three biggest companies in Silicon Valley had a much higher market capitalisation ($1.09 trillion), earned approximately the same revenues ($247 billion), but only employed about one-tenth as many people (137,000).6

At different speeds and in different ways, this kind of thing is happening to a great many occupations, and the casualties will not be restricted to manual workers. Indeed, many manual jobs are quite safe at the moment: it will be a long time before we have autonomous robots capable of operating independently in the real world and doing forestry or fisheries work — or, for that matter, on-site plumbing and carpentry. Whereas white-collar jobs that require considerable education and intelligence but involve a lot of repetition, such as researching legal precedents or keeping medical records, are being rapidly eroded by automation right now. Take the case of Fukuoka Mutual Life Insurance, which is currently replacing the people who calculate its payments on insurance claims with an artificial intelligence (AI) system based on IBM’s Watson Explorer.

It’s a complicated job, reading medical histories and reports, checking certificates, calculating the length of hospital stays and the cost of surgical procedures, and spitting out the sum to be reimbursed to the client — and at Fukuoka Mutual it had to be done 132,000 times a year. That kept thirty-four employees busy in full-time jobs, but they were all laid off in March 2017, replaced by a computer system that cost $1.8 million to install but will save the company $1.2 million a year in wages.7 If any job includes a lot of routine work, however complicated, it can be automated. If there are enough people doing that job to justify the expense of writing the software that will replace them, it will be automated eventually.

Workers of the future will need to be highly adaptable and juggle three or more different roles at a time … There will be constant new areas of work people need to stay on top of.

Anand Chopra-McGowan, head of enterprise new markets, General Assembly

The idea of continuous training is optimistic — I imagine there will be one-day training blitzes where people learn new skills quickly, and then are employed for a month when they’re needed.

Mark Spelman, co-head, Future of the Internet Initiative, member of the executive committee, World Economic Forum

We can’t all be knowledge workers. So there will be a lot of unemployment — and perhaps no impetus to help these people. There will be a division between the few jobs that need humans, and those that can be automated.

Dan Collier, CEO Elevate

Every day, somewhere on the planet, you will find a panel of academics, IT experts, and business people delivering ‘wisdom’ of this sort. In this case, it was a gathering called together by Deloitte UK, which provides ‘audit, consulting, financial advisory and tax services’, to ponder the Great Questions of the Day.8 We’ve heard it all before, and our kids are already living in the ‘gig’ economy. The discussion needs to move on to questions such as, ‘What kind of politics will we have with 50 per cent unemployment?’ or ‘Who’s going to buy your products if you automate everybody out of a job?’

Actually, those are pretty easy questions. The answer to ‘what kind of politics?’ is that by then you will either have created a society that is serious about redistributing income in a way that we haven’t seen for decades, or you will be living in a society that is on the brink of revolutionary violence (or has already fallen over the edge). And the answer to ‘Who’s going to buy your products?’ is: nobody. People who have been automated out of their jobs have no money to spend on them. In the name of increasing efficiency and squeezing the maximum return out of your capital, you are destroying the consumer base on which a successful capitalism depends.

The owner, the employees, and the buying public are all one and the same, and unless an industry can so manage itself as to keep wages high and prices low it destroys itself, for otherwise it limits the number of its customers. One’s own employees ought to be one’s own best customers.

Henry Ford, 19269

Ford practised what he preached: in 1914, he spontaneously doubled the wages of all his assembly-line workers to $5 a day and cut their work-day from nine hours to eight — and by the end of the decade, he had tripled his production of Model Ts to a million a year. But we aren’t going to solve the coming Great Unemployment by such dramatic gestures. You can argue about how fast the change is going to come, debate which occupations will be hit first or hardest, quibble about each and every detail in the predictions, but the change will be huge and it is already on the way. Carl Benedikt Frey, one of the economists at Oxford University’s Oxford Martin School whose analysis of which trades and professions were most vulnerable to automation in the United States was quoted in Chapter One, is profoundly pessimistic about the rate at which jobs will vanish. (He and his colleague Michael Osborne saw 47 per cent of all American jobs as ‘potentially automatable’ in twenty years.)

‘My reading of the evidence so far is that there will be less job-creating and ever greater labour-saving,’ Frey said in a recent interview. ‘If we look at the creation of new occupations by decades, they accounted for 8.2 per cent of new jobs in the 1980s, 4.4 per cent in the 1990s, and 0.5 per cent in the 2000s. It is not necessarily true that we will have a jobless future. But I struggle to use my imagination to see which industries will emerge to balance the loss of jobs.’10 So we are probably looking at a future twenty years down the road that includes a very high rate of unemployment: not necessarily a 50 per cent jobless rate, for some new jobs will have been created, but almost certainly 35 or 40 per cent. We could be surprised by good luck and find that the problem turns out to be smaller than this, but it would be unwise to plan for that. Less than half such an unemployment rate gave us Donald Trump, so what can we do to avoid even worse outcomes in the future?

Well, consider what they did to avoid even worse future outcomes after the Second World War. The 1930s had ended extremely badly, and it was clear to most people that the rise of the fascist dictators had been greatly facilitated, if not actually caused, by the impoverishment of the working class and especially the lower middle class after the great stock-market crash of 1929. Unemployment rates in the major industrial countries surged to unheard-of levels — a peak of 25 per cent in the United States and 27 per cent in Canada in 1933, nearly 30 per cent in Germany in 1932 (just before Adolf Hitler came to power), and 29 per cent in Australia the same year — and radical politics of both the right and the left flourished almost everywhere. So once the Second World War was over, people in the democratic countries began to think about creating political and economic systems that weren’t so vulnerable to popular rage. They created versions of the welfare state.

Most Western countries, both in Europe and overseas, already had old-age pensions, but the full set of state-provided services — from a comprehensive national health service free at the point of delivery, free education, and subsidised housing for the poor, to unemployment insurance, disability benefits, child benefits, and legal aid — were only instituted in most countries in the period 1945–1970. Even the United States developed a limited welfare state (‘social security’), beginning with Franklin D. Roosevelt’s New Deal in the 1930s, but it is significantly less ambitious than elsewhere — which may partly explain why popular anger is greater in the United States, and the political consequences more extreme. Of particular note is the fact that the United States is the only major developed country without a uniform and comprehensive national healthcare programme. Even if ‘Obamacare’ survives, some 25 million of the poorest Americans are still not covered, and most of the money passes through insurance companies rather than being administered directly by the government. As a result, American spending on health is the highest in the world — 17.6 per cent of Gross Domestic Product, compared to 11.4–12.0 per cent for France, Germany, the Netherlands, and Canada, and 9.1–9.6 per cent for the United Kingdom, Australia, New Zealand, Japan, Italy, and Spain — but health outcomes are roughly similar, regardless of spending.

It is not possible to say with confidence that the creation of more or less comprehensive welfare states was the key factor that kept Western politics off the radical road in the first half of the period since the Second World War, but it certainly helped to keep income inequality from getting out of hand. (It should also be recalled that during the Cold War the Western countries were competing with the communist-ruled states of Europe, all of which were full welfare states.) Equally, it cannot be proved that the switch to neo-liberal policies in the late 1970s, particularly in the ‘Anglosphere’, was enabled by the fact the lessons of the 1930s had been forgotten by then, but 40 years is a pretty plausible time-span for society-wide historical memory loss. At any rate, the forgetting was done, the income inequalities grew again, and here we are on Dover Beach with Donald Trump.

Compared to the people who built the welfare states 60–75 years ago, the political task that faces us now is not really that daunting. We aren’t trying to rebuild after the biggest war ever, with half the cities of Europe and Asia bombed flat. We aren’t all trapped in a new and potentially terminal global confrontation like the Cold War. We’re very much richer than we were last time around. We just have to shrink the income inequalities that are radicalising our politics already, and come up with some solutions for the fact that a great many jobs are going to disappear no matter what we do, causing further destabilisation as they go. And, yes, it probably will involve expanding the welfare state further.

Within a couple of months of Trump’s election, his canary-in-a-coalmine function had already produced its first and probably most important response: the revival of the old idea of a universal basic income (UBI). And its first high-profile outing, remarkably, was at the Davos meeting in January 2017, where several speakers were brought in to discuss the idea.

Well, not so remarkably perhaps, because the very rich are good at spotting emerging threats to their wealth and power, and the clear message of the US election was that the proles (and a lot of the old middle class) are angry and getting angrier. This will only get worse as automation advances, and at the same time the pool of potential customers for the goods and services whose sale keeps the whole capitalist economy afloat is shrinking as the jobs disappear. UBI is a tool that might cool the anger of people like Trump’s supporters by shrinking the income gap, and at the same time put some purchasing power in their pockets. It is therefore gaining support on both the right and the left, and it is now part of any discussion of how social justice and political stability might be preserved in a democratic society where capital is replacing labour (for that is precisely what automation really does).

One cautionary note. Many of the proposals that are being made, and almost all of the pilot programmes that are being run in various countries at the moment, are about a basic income, but not a universal basic income. Typically, such basic incomes are just attempts to streamline the delivery of various state benefits such as unemployment payments, child benefits, rent subsidies, and the like into a single payment that is less encumbered with restrictions and conditions, cheaper to administer, and perhaps more generous. At best, these basic incomes are provided to the working poor as well as to the workless, and do not involve the usual docking of welfare payments as earned income rises, but they are payable only to people below a certain income level. President Richard Nixon’s original 1969 proposal for a basic income for poor Americans was the first time such an idea reached the highest political level, but nowadays such proposals are commonplace.

A properly universal basic income, by contrast, goes to every citizen, rich or poor, as a matter of right with no strings attached. It should be enough to sustain a modest although not lavish lifestyle, and it does not decrease if the recipient is also in paid work. In the eyes of those who have recently fallen in love with the idea of UBI, the attraction is mainly that it may be able to prevent the social and political disaster that they fear automation will otherwise bring. It would not only keep the growing mass of people who have lost their jobs fed and clothed. Because everybody gets the universal basic income as a right, the argument goes, it would not stigmatise people in the same way that unemployment or welfare payments do. They would not feel humiliated by it, and so it would not breed the same misery, anger, and political extremism. And if the anger does not overflow, maybe democratic political systems can survive automation.

On the other hand, we should be aware that most people currently in the workforce, when asked to rank their preferences on a list that includes a higher minimum wage, laws that force firms to share some of their profits with employees, direct government subsidies to wages, and a universal basic income, will put UBI in last place. This may be mostly just that they are more comfortable with the familiar options, but it may also be because they are very reluctant to break the link between work and compensation. It makes the shift from being economically necessary to being surplus to economic requirements too open and too obvious. One sympathises, but if the jobs are really going to vanish, such preferences will have to give way to reality. And if the subsistence farmers of two centuries ago could make the transition to urban workers in industry and commerce who were paid in cash, today’s urban workforce can probably make a gradual transition to a more leisured existence without suffering a collective nervous breakdown.

We are probably in the early stage of a global economic transformation comparable to the industrial revolution, when entire populations went from being overwhelmingly rural to overwhelmingly urban in only two generations. This time, the transformation is from a full-employment economy to an economy of abundance that only requires a fraction of the population to work. But that does not necessarily or even probably mean a division of society into two castes, one of which does all the work while the other lazes around. It’s much more likely to be a society where most people work part-time or drop in and out of work for periods of months or years, although there will obviously be some at the extremes who work all of the time or none of it.

The industrial revolution was an angry, turbulent time, with urban uprisings and class warfare. We’ll be lucky if the damage this time is limited to demagogues like Donald Trump, who pander to the fear and anger of the newly displaced — and not just the displaced of the old working class, but the growing numbers of middle-class people who are also being displaced by machines. They are not ‘right wing’ in the traditional sense, although many have become more socially conservative, and some openly racist, as their panic rises. ‘Populist’ is a much better word: they hate the changes and the ‘elites’ who seem untouched by them, and they want their old jobs and their self-respect back. But the old jobs are not coming back, and even populist politics cannot resurrect them.

Besides, many of these people actually hated their jobs, from which they were only free for two weeks (in the US and Japan) or at most five weeks (in Europe) a year. The real task will be to find ways to provide a growing minority of our fellow citizens with money and self-respect without those jobs. They may become the majority in the end, just as UBI may be the answer in the end, but in the meantime it will have to be the old patches: higher minimum wages, stronger trade unions, perhaps a legal linkage between productivity gains and wage rises, and unemployment and welfare payments that do not automatically shrink or vanish when the recipient finds some low-paid casual work (some form of basic income, perhaps, though not the universal kind).

What we are going through now is not a disaster; it’s a process. Last time, it took over a century of mass misery and occasional mass bloodshed to get through it, but at the end most people were living much longer, healthier, more interesting lives than their peasant ancestors. We should try to do it a lot better and quicker this time.