On the importance of inflation [UK edition]

[Surlethe]

From the BBC:

The TUC said north-west and south-west England had seen the sharpest cuts - 10.6% and 10.1% respectively.

It blamed wages not keeping pace with inflation and changes in employment, such as more part-time working.

Business leaders said pay restraint had been crucial in protecting jobs during tough economic times.

After adjusting for inflation, the TUC's analysis of official figures suggested that on the eve of the recession, workers across the UK were earning a total of £690bn.

Last year the overall pay packet was £638bn - £52bn (7.5%) lower.

The North West saw the sharpest cut in its overall pay packet between 2007 and 2012 - a fall of 10.6% or £7bn last year. The West Midlands and Scotland saw cuts of 9.7%, it said.

Wales saw a reduction of 8.1%, Northern Ireland 4.8% and London 3.9%, the smallest cut.

The TUC analysed figures from the Office for National Statistics' annual survey of hours and earnings and its labour force survey.

'Massive hit'

A modest increase in employment had failed to offset a "sharp" cut in wages in recent years, said the TUC.

General secretary Frances O'Grady said: "Over the last five years, people have taken a massive hit in their pay packets, while millions more have had to reduce their hours or take lower paid work. Many people have lost their jobs altogether."

She said that shrinking wages were hitting living standards, holding back businesses and damaging growth prospects.

"While economic growth is the key challenge facing the UK today, the years running up to the crash taught us that growth without wage gains just creates more unsustainable debt," she said.

"Employers and both local and central governments need to recognise the importance of decent wages in delivering sustainable economic growth. They can start by becoming living wage employers and being more transparent about their pay systems."

'Tough climate'

But Neil Carberry, director for employment and skills at the Confederation of British Industry, said: "Pay restraint, though tough for many, has been crucial in protecting jobs and keeping people in work in the toughest economic climate for decades.

"The alternative for many businesses would have been to make staff redundant, but instead we've seen around a million new private sector jobs created in the last three years.

"The national minimum wage is already set at the right level, with tens of thousands of firms paying more if they can afford to do so."

Mike Cherry, national policy chairman at the Federation of Small Businesses, said: "At a time when economic growth is elusive and business confidence still fragile, it is understandable that many companies have been putting the annual pay rise on hold. In many low-skilled sectors, businesses already run on very thin margins, which means there is limited scope to increase wages even at the best of times, when the economy is performing strongly.

"Small employers will pay their employees more if they can afford it."

He said the government had an important role to play through greater action to reduce company overheads, such as business rates and fuel costs, and freeing up cash-flow by ending late payment practices by big companies to their suppliers.

"This will make small companies more profitable and enable them to pay their staff more," he said.

The TUC study coincides with its new pay campaign, Britain Needs A Pay Rise.

The tl;dr here is that a study by the Trade Union Congress (TUC) found that total UK compensation has fallen 10% since 2007, after adjusting for inflation, while employment is rising. The TUC press release is here. The TUC's takeaway is to campaign for higher wages, but my interpretation is a bit different.

First, note that their data start in 2007, right before the crash. Second, setting aside the fact that perhaps the starting point of the data skew the measurement, the key to understanding this, I think, is to remember that nominal wages are sticky. Since employment has been growing, and the unemployment rate has been holding steady at about 8%, this fall in real wages is not the result of a demand shortfall. If it were the result of a demand shortfall, spending and, crucially, employment would have fallen. Instead, employment is rising. How so?

I would say that high inflation is grinding down wages and making high employment possible, but that would contribute to a grievous misunderstanding of inflation. Instead, the price of labor is falling and nominal wages are sticky, so either inflation has to be high to allow the relative price change to continue, or unemployment has to rise. That's right: the UK's high inflation is preventing high unemployment. If inflation had been zero, a 10% fall in wages over 5 years would have resulted in a ~10% jump in unemployment, to 15% instead of 8%.

So in my reading, taking for granted bias issues from setting the start date of the study in 2007, right before an NGDP collapse, these numbers indicate that the UK is lucky to have a central bank willing to generate relatively high inflation to help the labor market balance itself, because the alternative is grinding unemployment.

[aerius]

I would say that high inflation is grinding down wages and making high employment possible, but that would contribute to a grievous misunderstanding of inflation. Instead, the price of labor is falling and nominal wages are sticky, so either inflation has to be high to allow the relative price change to continue, or unemployment has to rise. That's right: the UK's high inflation is preventing high unemployment. If inflation had been zero, a 10% fall in wages over 5 years would have resulted in a ~10% jump in unemployment, to 15% instead of 8%.

In other words you're lowering wages the backdoor way to make higher employment possible while keeping the wage slaves nicely pacified. Because if you knocked down nominal wages by the same percentage while keeping inflation low, even the people, dumb as they are, might just "get it" and become really pissed and start rioting. So we'll just rob them blind via inflation because the people don't understand this shit, and probably never will until it gets as bad as Zimbabwe.

[Guardsman Bass]

It's not going to get that bad. And if a slightly higher inflation rate stimulates aggregate demand and growth in the short run, then the nominal wages can always be raised again when the growth leads to greater employment and a tighter labor market.

[Irbis]

So in my reading, taking for granted bias issues from setting the start date of the study in 2007, right before an NGDP collapse, these numbers indicate that the UK is lucky to have a central bank willing to generate relatively high inflation to help the labor market balance itself, because the alternative is grinding unemployment.

Yeah, it's "lucky" to have central bank willing to dance to The City's tune and to shaft everyone with savings. Oh, and people paying taxes for their retirement pension, too, it will nominally stay the same but they will only be able to afford peanuts for it when they retire. So lucky

Meanwhile, Apple UK pays 0.16% tax from 6 bln £ profit. Run in your wheels, little inflation-shaved hamsters, after all, profits muss sein!

[madd0ct0r]

does this count as part of 'the great levelling' ?

[energiewende]

So in my reading, taking for granted bias issues from setting the start date of the study in 2007, right before an NGDP collapse, these numbers indicate that the UK is lucky to have a central bank willing to generate relatively high inflation to help the labor market balance itself, because the alternative is grinding unemployment.

Yeah, it's "lucky" to have central bank willing to dance to The City's tune and to shaft everyone with savings. Oh, and people paying taxes for their retirement pension, too, it will nominally stay the same but they will only be able to afford peanuts for it when they retire. So lucky

Meanwhile, Apple UK pays 0.16% tax from 6 bln £ profit. Run in your wheels, little inflation-shaved hamsters, after all, profits muss sein!

You are wrong about this. The wage cut is necessary because the GDP per capita has reduced. There is less money to go around. The only way to keep wages high is to make people unemployed which hurt the poor most and causes a deeper recession. This is what has happened in Spain because of its membership of the Euro. It has inflation rates that are optimal for Germany; inflation rates optimal for Spain would be much higher. As a result Spanish unemployment is about 4x that of the UK. In fact in pretty much every developed country that isn't afflicted by the Euro inflation has kept unemployment reasonably low. That is why we haven't seen a repeat of the Great Depression even though the GDP hit was comparable to the 1939 crash.

The TUC "report" is essentially propaganda, however it is possible that low inflation really would have been good for them. Most of their members are now state employees who are unlikely to be fired whatever happens, so it's possible their members could absorb the higher wages without suffering personally from increased unemployment. But they would be benefiting at the expense of the rest of society, so this isn't a good argument for keeping inflation low.

[Jub]

You are wrong about this. The wage cut is necessary because the GDP per capita has reduced. There is less money to go around. The only way to keep wages high is to make people unemployed which hurt the poor most and causes a deeper recession. This is what has happened in Spain because of its membership of the Euro. It has inflation rates that are optimal for Germany; inflation rates optimal for Spain would be much higher. As a result Spanish unemployment is about 4x that of the UK. In fact in pretty much every developed country that isn't afflicted by the Euro inflation has kept unemployment reasonably low. That is why we haven't seen a repeat of the Great Depression even though the GDP hit was comparable to the 1939 crash.

The TUC "report" is essentially propaganda, however it is possible that low inflation really would have been good for them. Most of their members are now state employees who are unlikely to be fired whatever happens, so it's possible their members could absorb the higher wages without suffering personally from increased unemployment. But they would be benefiting at the expense of the rest of society, so this isn't a good argument for keeping inflation low.

Why not simply pass laws closing loopholes in corporate tax law and then use the extra cash flow to keep the citizens afloat until things recover? If you fear companies pulling out over it you can pass strict laws on what can be sold by companies that dodge in this way. If you fear job loss set up some state owned factories and replace some of the goods you just lost with the tax grab.

[energiewende]

Because that wouldn't reduce unemployment. It's deflationary in fact, so it would likely increase it.

Given wage stickiness there is no free lunch from low inflation in a recession. Total compensation is going to drop anyway just because the economy is smaller; the only question is whether it will be distributed more equitably (high inflation/low unemployment) or less equitably (low inflation/high unemployment).

As for establishing Gosplan, I don't think I need to say why that is a bad idea.

[Jub]

Because that wouldn't reduce unemployment. It's deflationary in fact, so it would likely increase it.

Given wage stickiness there is no free lunch from low inflation in a recession. Total compensation is going to drop anyway just because the economy is smaller; the only question is whether it will be distributed more equitably (high inflation/low unemployment) or less equitably (low inflation/high unemployment).

As for establishing Gosplan, I don't think I need to say why that is a bad idea.

So taxing companies to the rates that they would pay if they weren't all nailing every loophole and/or cooking the books is a bad thing now? That is and always will be bullshit. It will hurt for a little bit, but the government will have increased income and can use that to start up other programs. Unemployment doesn't matter when the state can afford to keep people in roughly the same conditions with social programs.

Really this is where the world is heading anyway with everything getting more automated. The more automation we get the less workers we'll need in service jobs, and service jobs are where most people work their entire lives. Thus the goal shouldn't be maintaining employment levels, but instead maintaining and increasing standards of living for those that are made replaceable in the work force.

This also isn't anything like Gosplan. In the near term state run factories can make substandard crap and it won't matter because the whole thing is being funded by Apple and other companies paying 10% tax instead of less than 0.2%. In the long term the nation transitions to higher levels of unemployment as people are made redundant by technology.

[energiewende]

It's neither good or bad, it's just beside the point. Raising taxes won't reduce the negative effects of wage stickiness on employment.

Replace 'service' with 'industry' and go back to 1950. Replace 'industry' with 'agriculture' and go back to 1850. There's no long run trend of increased unemployment.

It means that instead of working for Apple, as people would if we inflated to counteract wage stickiness, labour is instead wasted with people working to make "sub-standard crap". This is opportunity cost of uncreated value that will affect the whole of society.

[Jub]

It's neither good or bad, it's just beside the point. Raising taxes won't reduce the negative effects of wage stickiness on employment.

Replace 'service' with 'industry' and go back to 1950. Replace 'industry' with 'agriculture' and go back to 1850. There's no long run trend of increased unemployment.

It means that instead of working for Apple, as people would if we inflated to counteract wage stickiness, labour is instead wasted with people working to make "sub-standard crap". This is opportunity cost of uncreated value that will affect the whole of society.

You obviously missed the bit where we transition out of having those people make anything. Plus in the past you could only automate so much. You could cut 90% of fast food workers if you built a burger factory, you could do the same for many retail stores if you made them vending machines, construction won't be safe for long either. Where do the entry level people go when we cut the bottom of the job pool out?

[Starglider]

So taxing companies to the rates that they would pay if they weren't all nailing every loophole and/or cooking the books is a bad thing now?

Taxing productive activity (manufacturing and non-situated services) reduces investment and hence employment, as no one wants to operate a business in a high-cost environment.
Taxing retail businesses is just another sales tax (on top of the very high existing UK sales tax) which will be passed on to consumers.

It will hurt for a little bit, but the government will have increased income and can use that to start up other programs.

The government will spend the money on pensions, civil service pay increases, civil service pensions, subsidies to banks and miscellaneous corruption.

Unemployment doesn't matter when the state can afford to keep people in roughly the same conditions with social programs.

"The problem with socialism is that you eventually run out of other people's money." It is easier for productive people, industry and capital to leave than ever before. The tax base will evaporate as France is currently discovering. You cannot create an economy from subsidised substandard companies funded (ultimately) by taxes on imports.

[Surlethe]

I would say that high inflation is grinding down wages and making high employment possible, but that would contribute to a grievous misunderstanding of inflation. Instead, the price of labor is falling and nominal wages are sticky, so either inflation has to be high to allow the relative price change to continue, or unemployment has to rise. That's right: the UK's high inflation is preventing high unemployment. If inflation had been zero, a 10% fall in wages over 5 years would have resulted in a ~10% jump in unemployment, to 15% instead of 8%.

In other words you're lowering wages the backdoor way to make higher employment possible while keeping the wage slaves nicely pacified. Because if you knocked down nominal wages by the same percentage while keeping inflation low, even the people, dumb as they are, might just "get it" and become really pissed and start rioting. So we'll just rob them blind via inflation because the people don't understand this shit, and probably never will until it gets as bad as Zimbabwe.

We've been over inflation before. The basic point I'm making is: UK wages are falling because of structural reasons whether you want them to or not. Labor is less valuable than it used to be; its price must fall relative to other goods and services. Society gets to choose whether that happens by throwing some unlucky, low-product people out of work, or by more equitably distributing the wage cuts through slightly higher inflation.

And don't get too scared of inflation; it's not going to get as bad as Zimbabwe. MV = PQ.

Yeah, it's "lucky" to have central bank willing to dance to The City's tune and to shaft everyone with savings. Oh, and people paying taxes for their retirement pension, too, it will nominally stay the same but they will only be able to afford peanuts for it when they retire. So lucky

Savers have social responsibility too, you know. If the economy isn't as productive, everyone should feel it, lenders and borrowers alike. Higher inflation caused by supply problems lets savers and borrowers share the pain instead of throwing the marginal workers under the bus. Basically, holding the UK's supply side issues constant, you get to choose: are you going to fuck over a few of the least productive workers and let everyone else off the hook, or are you going to try to distribute the suffering across all of society?

[Surlethe]

It's not going to get that bad. And if a slightly higher inflation rate stimulates aggregate demand and growth in the short run, then the nominal wages can always be raised again when the growth leads to greater employment and a tighter labor market.

It's not so clear that the UK's problem is mostly AD. Maybe some - unemployment is still at 8% - but nothing like the US. The evidence for this is that UK total employment is rising and past where it was in 2007, while inflation has been between 2-4% since 2007.

Also, I think we should be careful about causation. Higher inflation doesn't stimulate AD and growth in the short run; rather, stimulated AD and growth in the short run cause higher inflation. What stimulates short-run AD and growth is higher expected NGDP. (From the equation of exchange, MV = PQ; higher expected PQ induces an increase in V, which pushes up current PQ.) Whether that short-run AD stimulation ends up as inflation or production depends on whether the economy is producing at capacity or below capacity.

[Guardsman Bass]

It's not so clear that the UK's problem is mostly AD. Maybe some - unemployment is still at 8% - but nothing like the US. The evidence for this is that UK total employment is rising and past where it was in 2007, while inflation has been between 2-4% since 2007.

NGDP in the UK is still well below levels in 2008-2009, so I'm not convinced it's structural yet. That said, the case is definitely stronger for it than in the US, and there's also the combination of shrinking unemployment with low productivity.

Also, I think we should be careful about causation. Higher inflation doesn't stimulate AD and growth in the short run; rather, stimulated AD and growth in the short run cause higher inflation. What stimulates short-run AD and growth is higher expected NGDP. (From the equation of exchange, MV = PQ; higher expected PQ induces an increase in V, which pushes up current PQ.) Whether that short-run AD stimulation ends up as inflation or production depends on whether the economy is producing at capacity or below capacity.

True, I should have been referring to stimulating NGDP by increasing the money supply.

So taxing companies to the rates that they would pay if they weren't all nailing every loophole and/or cooking the books is a bad thing now? That is and always will be bullshit. It will hurt for a little bit, but the government will have increased income and can use that to start up other programs. Unemployment doesn't matter when the state can afford to keep people in roughly the same conditions with social programs.

Issues about exactly how much would be saved if you could somehow close every "loophole", monetary policy is far faster and far more capable of being done at the necessary scale. Compare how quickly the Federal Reserve could do quantitative easing compared to how difficult it was to even get a stimulus package passed in the US Congress, and you get the idea.

Really this is where the world is heading anyway with everything getting more automated. The more automation we get the less workers we'll need in service jobs, and service jobs are where most people work their entire lives. Thus the goal shouldn't be maintaining employment levels, but instead maintaining and increasing standards of living for those that are made replaceable in the work force.

If we manage nominal GDP growth to the "full employment" level, then we'll probably find replacement jobs for the ones lost. You can't imagine them now, but then workers in the 1920s didn't imagine computer programmers. Moreover, I suspect a lot of jobs will actually stick around even if they get complemented by new technology, because the use and degree of automation is controlled by the opportunity cost of using it (meaning that it's the most efficient way of doing something). You brought up vending machines, but there are tons of places where we could be using that technology now and don't, because the opportunity cost of the resources spent setting it up and operating it is more expensive than just using workers at cash registers.

Now, if we go through a few more business cycles and the unemployment rate remains stubbornly high despite strong growth, then I'll be convinced. But as for the US, I'm not convinced we're in that situation.

[energiewende]

You obviously missed the bit where we transition out of having those people make anything. Plus in the past you could only automate so much. You could cut 90% of fast food workers if you built a burger factory, you could do the same for many retail stores if you made them vending machines, construction won't be safe for long either. Where do the entry level people go when we cut the bottom of the job pool out?

Vending machines and food processing plants are not new technology. Why don't people buy TV dinners from a mile long vending machine rather than going to store?

These are especially terrible examples, but suppose the whole retail sector does disappear - what do you think happened when the agricultural sector as good as disappeared? Mechanisation is nothing new. It's at least 300 years old, and so far has had next to no effect on employment rates. This is no surprise to economics...

[Simon_Jester]

Vending machines and food processing plants are not new technology. Why don't people buy TV dinners from a mile long vending machine rather than going to store?

These are especially terrible examples, but suppose the whole retail sector does disappear - what do you think happened when the agricultural sector as good as disappeared? Mechanisation is nothing new. It's at least 300 years old, and so far has had next to no effect on employment rates. This is no surprise to economics...

Congratulations, you've identified a trend! Now, to prove that you can really think, and not just repeat maxims, let's take it to the next level: analyzing the trend.

Mechanization displaces jobs. New jobs are created... where? In new sectors, where people do something a machine cannot do, or cannot do well.

If you were pushed out of the labor force by mechanization in 1800 or 1900 or even 1980, the odds were that you could find another job which required no unusual credentials or education to do, right away. A farmhand displaced by mechanical harvesters could get a job digging ditches; his son displaced from the ditch digging job by steam shovels could get a job throwing switches on a railroad; his son displaced by electrical signaling and control could get a job on an assembly line; his son displaced by better factory automation could get a job as a store clerk.

But now that guy's son, displaced by an automated online filing system, needs a substantial education in some specialized field to be a competitive player in the workforce.

This creates a new obstacle to workers seeking to re-enter the labor market and sell their labor: sure, jobs exist in theory, that hasn't changed. But the barrier to entry is higher. And it's damn hard for an unemployable random person to find "venture capital" to improve their standing as a seller of labor, except on predatory terms.

[Guardsman Bass]

Mechanization displaces jobs. New jobs are created... where? In new sectors, where people do something a machine cannot do, or cannot do well.

No, they're created in places where it makes more economic sense to use people than machines. Having it be in tasks that automation is simply unable to do is just the most prominent example of that, but you can still have people employed even in a situation where machines could literally do anything that human beings can. And as I pointed out, we almost never see exactly where that's going to be until we're almost there - nobody in the 1920s saw a computer programmer coming, and a farmer in the late 1700s wouldn't have foreseen the possibility of someone working in a factory manufacturing cars in the 1920s. It's one of the reasons why futurism is almost always a joke when it comes to predictions.

Hence my point about being skeptical that the unemployment in the current business cycle is anything other than cyclical - recessions always drag the "structural problem" crowd out of the woodwork to complain about how this shows the economy has deeper problems. If it sticks around for another decade or two despite strong growth, then we might have a problem.

If you were pushed out of the labor force by mechanization in 1800 or 1900 or even 1980, the odds were that you could find another job which required no unusual credentials or education to do, right away. A farmhand displaced by mechanical harvesters could get a job digging ditches; his son displaced from the ditch digging job by steam shovels could get a job throwing switches on a railroad; his son displaced by electrical signaling and control could get a job on an assembly line; his son displaced by better factory automation could get a job as a store clerk.

That just proves my point. The farmhand displaced by mechanical threshers would never foresee that his son was going to get a job throwing switches on the railroad, until it actually happens.

[Zaune]

There's an angle to this that Simon didn't cover. A non-trivial percentage of workers couldn't find replacement work back then, because the whole point of most noteworthy technological innovations in history has been to do more work with less manpower. But that wasn't an insurmountable problem in those days because there was somewhere to emigrate to if your local supply of jobs had dried up. That's not really an option for my generation.

[energiewende]

Vending machines and food processing plants are not new technology. Why don't people buy TV dinners from a mile long vending machine rather than going to store?

These are especially terrible examples, but suppose the whole retail sector does disappear - what do you think happened when the agricultural sector as good as disappeared? Mechanisation is nothing new. It's at least 300 years old, and so far has had next to no effect on employment rates. This is no surprise to economics...

Congratulations, you've identified a trend! Now, to prove that you can really think, and not just repeat maxims, let's take it to the next level: analyzing the trend.

Mechanization displaces jobs. New jobs are created... where? In new sectors, where people do something a machine cannot do, or cannot do well.

If you were pushed out of the labor force by mechanization in 1800 or 1900 or even 1980, the odds were that you could find another job which required no unusual credentials or education to do, right away. A farmhand displaced by mechanical harvesters could get a job digging ditches; his son displaced from the ditch digging job by steam shovels could get a job throwing switches on a railroad; his son displaced by electrical signaling and control could get a job on an assembly line; his son displaced by better factory automation could get a job as a store clerk.

But now that guy's son, displaced by an automated online filing system, needs a substantial education in some specialized field to be a competitive player in the workforce.

This creates a new obstacle to workers seeking to re-enter the labor market and sell their labor: sure, jobs exist in theory, that hasn't changed. But the barrier to entry is higher. And it's damn hard for an unemployable random person to find "venture capital" to improve their standing as a seller of labor, except on predatory terms.

There seems indeed to be a widespread belief that number of workers and number of jobs are entirely decoupled quantities. Agriculture may go out, but then railways come in, so the accidental discovery of technologies to build railways just perfectly compensates the accidental discovery of largely different technologies that increased agricultural productivity. I call it the Tightrope Theory of Employment - a wobble one way must be compensated by a wobble the other way, or else society will plunge to its doom - but how sensible a belief is it really that these two entirely decoupled mechanisms have produced identical numbers for centuries, even as the workforce and technology change at rapid and often erratic rates? I think that to ask the question is to answer it.

This is no surprise to economists, because in fact the two numbers aren't decoupled at all. The number of jobs expands to fill the number of job-seekers (at least, in the long run) regardless of technology. Labour is just like any other commodity on the market - it will sell for some price - and mechanisation, by increasing the productivity of labour, increases that price. The worst case in the 19th century is that you get a lot more people making arts & crafts with old fashioned methods after being displaced from the fields by mechanised agriculture.

This is true regardless of education; to the extent it is useful, it's provided by employers. To the extent it isn't, it's a sorting mechanism and the responsibility for it being required lies with the government for subsidising it. This also ensures that most people have it - if they didn't, it wouldn't be required for jobs.

[Jub]

My thinking is the same as Simon_Jester's. There undoubtedly will be new jobs created, but educational needs continue to rise and the public education system isn't up to the task of training people. Thus people need to scrounge up the capital to go to post secondary to get the same type of entry level job that you need a simple diploma (or less) for now. This wouldn't be such an issue if we had free post secondary or added a few years to public school, but even then there are people that wouldn't be happy doing a job that required so much extra knowledge and it marginalizes those that don't/couldn't finish school even more than the current system.

Yeah, there will need to be more maintenance techs if we increase automation to the levels that I can see coming, but that might replace 1 job in 10. Think off all the jobs that would be lost if something like a Kinect could scan you and allow you to basically try on clothing from the comfort of home, now imagine if the two gas station employees are cut down to one guy at all hours, and the fast food places basically only have a host or two in the lobby and a tech that comes by once a week or so. Construction jobs are at risk too, you could replace an entire curb, gutter, and sidewalk crew with a few automated trucks and a supervisor, the same goes for a machine to knock up a house frame. These sorts of mindless entry level and hands on jobs will go away eventually. What sort of job requiring the same skill and commitment do you see replacing those types of jobs?

Now you might argue that the factories and supply chains to fuel these automatons might pick up the slack. Producing the new machines will likely create either few jobs, or jobs that only the developing world can compete at cost wise. Warehouse and distribution jobs already exist but need only a fraction of the employees that even a Walmart does. The trucking/delivery end of things might be most automated within our lifetime for the simple reason that a machine doesn't need mandatory sleep breaks and carries less risk than a cab full of immigrants.

You claim we can't predict the future, and I can agree, but I don't see what will be replacing the bottom of the heap labour market once a machine is cheaper. Instead I see countries failing to adapt to the new economic reality this will create.

There seems indeed to be a widespread belief that number of workers and number of jobs are entirely decoupled quantities. Agriculture may go out, but then railways come in, so the accidental discovery of technologies to build railways just perfectly compensates the accidental discovery of largely different technologies that increased agricultural productivity. I call it the Tightrope Theory of Employment - a wobble one way must be compensated by a wobble the other way, or else society will plunge to its doom - but how sensible a belief is it really that these two entirely decoupled mechanisms have produced identical numbers for centuries, even as the workforce and technology change at rapid and often erratic rates? I think that to ask the question is to answer it.

This is no surprise to economists, because in fact the two numbers aren't decoupled at all. The number of jobs expands to fill the number of job-seekers (at least, in the long run) regardless of technology. Labour is just like any other commodity on the market - it will sell for some price - and mechanisation, by increasing the productivity of labour, increases that price. The worst case in the 19th century is that you get a lot more people making arts & crafts with old fashioned methods after being displaced from the fields by mechanised agriculture. This is true regardless of education; to the extent it is useful, it's provided by employers. To the extent it isn't, it's a sorting mechanism and the responsibility for it being required lies with the government for subsidising it. This also ensures that most people have it - if they didn't, it wouldn't be required for jobs.

Never before have we had the threat of widespread automation replacing so many sectors. When the cost of buying a machine dips below the cost of training a man, and maybe even a bit before that, you will start to see companies make the switch. The only way an employee for the same role can be attractive at that stage is to work inhumanly long hours at below the cost of maintaining the machine. Even then the employer will value constancy and reliability and the machine will trump even the best person on those areas.

When work gets to the point where a man can't beat the machine in cost, value to the company, and results a lot of people will find the skills they have obsolete. Some might have the means, desire, and the mental capacity to retrain; many won't. Where does this leave the 90 IQ stockboy for life, or the 40 something who thought she could work until retirement as a cashier? Do we even have the capacity to retrain all of them even if education were free and all involved were able?

[aerius]

We've been over inflation before. The basic point I'm making is: UK wages are falling because of structural reasons whether you want them to or not. Labor is less valuable than it used to be; its price must fall relative to other goods and services. Society gets to choose whether that happens by throwing some unlucky, low-product people out of work, or by more equitably distributing the wage cuts through slightly higher inflation.

Or we can be honest about it and just say that everyone can either take a 10% paycut or lose their jobs and keep the inflation rate at 0%. But no, we do it the sneaky way through jacking up inflation because people won't "get it" and will remain nicely pacified until they're well & truly fucked.

BTW, I hear that money printing to target inflation is working great in Japan, they've had 2 big crashes in their markets, yo-yo'd the Yen, and gone on their longest streak to date of trade deficits since they started their unlimited monetization.

[Simon_Jester]

Mechanization displaces jobs. New jobs are created... where? In new sectors, where people do something a machine cannot do, or cannot do well.

No, they're created in places where it makes more economic sense to use people than machines. Having it be in tasks that automation is simply unable to do is just the most prominent example of that, but you can still have people employed even in a situation where machines could literally do anything that human beings can.

OK, strictly this is true. Would it be acceptable to my point to modify that to say "where people do something a machine cannot do, or cannot do well, or cannot do cost-effectively?"

In the developed world, labor has gotten very expensive; life isn't as cheap as it used to be. So I'd be worried about relying on sectors where we know right now we could build a machine to do the job, but also know it wouldn't be cost-effective. That could change fast when someone figures out how to build a better mouse trap.

Hence my point about being skeptical that the unemployment in the current business cycle is anything other than cyclical - recessions always drag the "structural problem" crowd out of the woodwork to complain about how this shows the economy has deeper problems. If it sticks around for another decade or two despite strong growth, then we might have a problem.

Well, personally I think that the rise of the computer has fundamentally changed the dynamic, but not necessarily in a way we can't solve. However, we may have to make institutional changes before we can solve it.

That just proves my point. The farmhand displaced by mechanical threshers would never foresee that his son was going to get a job throwing switches on the railroad, until it actually happens.

But to make that possible, we had to, well... build those railroads. We had to create infrastructure that could keep people profitably engaged doing the right kind of work. I think that's happening again. This time the necessary infrastructure is a comprehensive education system and social support network that can keep people from dropping out and becoming a permanent underclass. Because with the rise of greater automation, that underclass is really unemployable, compared to past decades when someone would have hired them as navvies to dig canals or something.

There seems indeed to be a widespread belief that number of workers and number of jobs are entirely decoupled quantities. Agriculture may go out, but then railways come in, so the accidental discovery of technologies to build railways just perfectly compensates the accidental discovery of largely different technologies that increased agricultural productivity. I call it the Tightrope Theory of Employment - a wobble one way must be compensated by a wobble the other way, or else society will plunge to its doom - but how sensible a belief is it really that these two entirely decoupled mechanisms have produced identical numbers for centuries, even as the workforce and technology change at rapid and often erratic rates? I think that to ask the question is to answer it.

The problem is that the coupling isn't a simple relationship. And it isn't a nice simple, antiseptic thing either. Altering the number of jobs (labor demand) can have huge effects on the suppliers of labor... i.e. everyone in the damn country.

Conditions for labor suppliers (everyone) were pretty good from (roughly) WWII to 2000. Changes in technology and social policy could alter that, and this is a matter of concern for people affected by the labor market (everyone).

Civilization won't come to an end if we wind up in a situation where 30% of the population is employed doing absurd scut-work (or nothing at all), and where the median standard of living declines to 1925 levels. But that's hardly desirable; it's a huge waste of human potential and technological potential.

This is true regardless of education; to the extent it is useful, it's provided by employers. To the extent it isn't, it's a sorting mechanism and the responsibility for it being required lies with the government for subsidising it. This also ensures that most people have it - if they didn't, it wouldn't be required for jobs.

But when you say that you're entirely ignoring what the jobs are. Normal humans care whether we have an 1850 or a 1990 standard of living, you see.

An industrial economy, let alone a post-industrial economy, requires a lot of skilled, intelligent workers who can solve problems independently and who understand technical subjects. You don't get that without education. Sure, if the state stops subsidizing education the resulting generation of ignoramuses will be employable... but doing what? Doing pretty much the same things their equally ignorant great-great-grandparents were doing: manual labor of minimal productivity. Things someone would happily design a machine for, if only there were enough educated people to design, build, and maintain the necessary machines.

My thinking is the same as Simon_Jester's. There undoubtedly will be new jobs created, but educational needs continue to rise and the public education system isn't up to the task of training people. Thus people need to scrounge up the capital to go to post secondary to get the same type of entry level job that you need a simple diploma (or less) for now. This wouldn't be such an issue if we had free post secondary or added a few years to public school, but even then there are people that wouldn't be happy doing a job that required so much extra knowledge and it marginalizes those that don't/couldn't finish school even more than the current system.

Now THAT is actually a problem energiewende references: credential inflation. Especially for young people today, "do you have a bachelor's degree" is being used as a proxy for "are you intelligent, disciplined, and obedient enough to be a model employee?" If the answer is "no," or if there's anything else wrong with you, other people get bumped up the desirability list.

Given the way modern job applications work, this makes people with bad credentials (or other problems like long term unemployment) very unlikely to get hired.

This only works because we pressure everyone very hard to go to a university, and assume that the people who don't make it are failures. Adjust the system and we adjust the problem.

[energiewende]

Those dates you cherrypicked are just the times of major US recessions. Recessions can temporarily increase unemployment but recessions aren't caused by indefinite technological shifts; generally they're caused by bad monetary policy which is the point of this thread. The fact that unemployment before and after the Great Depression, and today, are all hardly different just torpedoes your thesis completely. To the extent I am not really interest in answering all the tangents.

Now you segue into what is a totally unrelated claim - that if the state doesn't fun 4 year college degrees the GDP per capita will start to reduce rather than historically increase. All I can say to that is that if you really think historical US GDP per capita growth - which is slower since the 70s than before - is only being propped up by millions of psychology, poli sci, history, english, and other non-vocational graduates who have less than 10% chance of ever working in their field of alleged expertise, I have a bridge to sell you.

[Guardsman Bass]

BTW, I hear that money printing to target inflation is working great in Japan, they've had 2 big crashes in their markets, yo-yo'd the Yen, and gone on their longest streak to date of trade deficits since they started their unlimited monetization.

They're still better off than where they were before Abe starting pushing for a higher inflation target.

OK, strictly this is true. Would it be acceptable to my point to modify that to say "where people do something a machine cannot do, or cannot do well, or cannot do cost-effectively?"

In the developed world, labor has gotten very expensive; life isn't as cheap as it used to be. So I'd be worried about relying on sectors where we know right now we could build a machine to do the job, but also know it wouldn't be cost-effective. That could change fast when someone figures out how to build a better mouse trap.

"Cost-Effective" is the key word there, but you have to be careful - the critical factor in determining whether automation or humans-plus-machines is used comes down to which is more cost-effective/efficient compared to the alternative at a particular task. Like I said earlier, you can see some of this with vending machines, where we could be using them to a vastly higher degree instead of using cashiers at registers, but we don't.

You're right that conditions can change, and using automation for a particular task might have a lower opportunity cost for a particular task compared to humans. But that's been true for most of the industrial revolution, as I pointed out earlier, and we keep finding ways to integrate humans into the production of goods and services. And so far, we haven't seen a link between greater unemployment and greater use of robots in the labor force - Germany has more than twice the "robot density" of the US, but with a lower unemployment rate.

But to make that possible, we had to, well... build those railroads. We had to create infrastructure that could keep people profitably engaged doing the right kind of work. I think that's happening again. This time the necessary infrastructure is a comprehensive education system and social support network that can keep people from dropping out and becoming a permanent underclass. Because with the rise of greater automation, that underclass is really unemployable, compared to past decades when someone would have hired them as navvies to dig canals or something.

I'm not convinced they're unemployable, particularly if we truly are heading into a period with smarter machines (one side-effect of which might be that lower-trained workers can make up for lower training with extra smart machine assistance). If you look at the last true "boom" period in the US (the late 1990s), you see that tons of otherwise "low-skilled" workers were drawn into the economy. The labor market was tight enough that companies re-trained them, and found things for them to do. Others have made that point better than me.

But in the mean-time, we are responding to this by getting more education, at least in younger cohorts of workers.

Civilization won't come to an end if we wind up in a situation where 30% of the population is employed doing absurd scut-work (or nothing at all), and where the median standard of living declines to 1925 levels. But that's hardly desirable; it's a huge waste of human potential and technological potential.

Depends on the "scut-work" - it's not like we actually enjoyed doing repetitive tasks in factories for eight hours a day, and the future alternative might be better. I will concede that while I don't think employment will be a problem per se, pay could be an issue. Even then, we've got a labor force in most rich countries that is actually shrinking in size if not for immigration, which will tend to drive up wages as the available supply of labor relative to demand shrinks.