Automation was supposed to threaten our jobs, yet the unemployment rate hovers near all-time lows. We’re in our eighth year of uninterrupted jobs growth, the longest streak in our history. An adage repeated without evidence since the earliest days of capitalism – every “job-destroying” innovation creates more jobs that it destroys – seems to be vindicated again.
A wave of automation that swept across manufacturing, especially automobile manufacturing, seems to have stalled. Other industries poised for disruption by computer and robotic technologies like fast food, retail, accounting, law and even radiology, have seen little impact from automation as yet. A backlog of astonishing robotic technology has been introduced, but not yet adopted. Leaving aside for a moment the question of whether automation will destroy our employment culture (addressed at length here), an economic puzzle confounds us: why does automation seemed to have slowed?
One explanation seems the most persuasive. Capital owners have amassed extraordinary market power. In short, companies have little reason to invest their capital in disruptive innovations with uncertain outcomes. A company that can earn a steady small profit from changing nothing will continue to do so indefinitely, even if presented with a speculative opportunity to multiply that profit many times over with an investment in technology. At present, companies feel little need to take a chance on automation.
Many of the forces that would cut into the profits of a stagnant, dumb company have been tamed. Labor unions have collapsed, crippling the capacity of workers to demand a greater share of corporate earnings. Since the Bush II Administration, monopoly regulation has largely disappeared. Tax rates keep dropping. Regulations that curb pollution, securities fraud, and collusion have been consistently weakened. Adjusted for inflation, the federal minimum wage hasn’t budged in fifty years. Low-skilled, median-wage work has largely disappeared while skyrocketing costs for health care and housing have strained working families.
Amazon has both the capital and the technology to automate away most of the work in its fulfillment centers, but with wages for full-time employees stuck around $13/hour, why bother? To cope with seasonal spikes in demand, the company has developed an entire culture of low-paid itinerant workers, many of whom are over 60 and essentially homeless, living in their vehicles. They’ve even given this program a name – Camperforce. A migrant labor pool reanimated from the pages a Steinbeck novel is given a catchy, tech-era gloss, buying time for Amazon to work out the kinks in its automation plans. Meanwhile, investors pocket capital that would otherwise have flowed into faster and riskier innovation.
Compare job growth with wage growth, and the forces stalling automation become obvious. Yes, we’re in a cycle of extraordinary employment growth, but with none of the inflationary pressure we should be seeing. Wage growth, which has been expanding slightly in recent years, remains stalled for median earners. Almost all income growth over the past thirty years has flowed to the highest earners. When the wealthy have no incentive to invest in disruptive innovations, those innovations do not get deployed.
An economy that can sustain full employment (an unemployment rate below roughly 5%) for sustained periods of time without inflationary pressures is badly broken. An unemployment rate below 3.9% signals an employment market in deep trouble.
Over the past twenty years America has been converting to a rentier economy, enabled by hyper-low tax rates and a disappearing government. The wealthy are sitting on their capital, enjoying adequate returns from their depreciating assets. That won’t last forever. The rest of the world isn’t sitting still. Competitive pressures from other economies, or another financial crisis sparked (again) by asset inflation will eventually force a break. When that break comes, the machine of automation will crank back up.
The question for the future is who will benefit from the continued destruction of jobs. America still holds a lead in technology, built up from a century of investments in education, but that lead is shrinking. A powerful rentier class is flexing its muscles in ways not seen here since the Gilded Age and they have nothing to gain from disruptive innovation.
20th century innovations like a minimum wage and unionization won’t make a dent in this problem. Disrupt the political conditions that are making capital owners secure, and jobs will start to disappear. For example, technology to automate the fast food business is already in place. With a $7.25 minimum wage, owners have little need to spend capital on those machines, and on the expensive engineers and designers needed to deploy them. Impose a fair minimum wage and eliminate the regulatory conditions that make the rentier economy work, and those capital owners will have to start investing and competing again. They won’t survive true competition by wasting money on $15/hour order-takers.
Does that mean we should fight the expansion of automation and hold on to those cheap jobs? Should we prefer horse wagons over cars? Leeches over prescription drugs? Never bet against technology. If innovation is going to displace jobs, we should embrace it anyway, using social and political adaptations to cushion the impact of job loss rather than slowing technical growth.
We are trapped in a push-pull between low unemployment with low pay and higher wages with disappearing jobs. That dynamic will remain as long we retain our cultural and economic dependence on mass employment. Smarter and faster tools for production demand smarter and faster mechanisms for distributing the bounty of those tools. The tighter we cling to cling to an outdated model of employment the more spectacular the impact when that model collapses.
I realize this completely off-topic, but given today’s events, I must say something. I really really hope that China starts selling military technology to Iran, and Russia sells more, with both countries establishing a public military presence in that country, in the form of “military advisors”.
China imports in excess of 600,000 barrels of Iranian oil daily. No way, no how, they are going to start messing with that.
Would love to see if the U.S. and their pet pitbull Israel would actually roll the dice and strike military anti-aircraft sites manned by Russians and Chinese. Then again, this madman might actually do it.
That, or actually sell nukes to Iran. No country that has nukes has ever been invaded, for obvious reasons.
Dins, you gotta rein this in. Seriously. Your head’s in a scary place.
Trying to avoid World War III here, Dinsdale, not jumpstarting it.
Re- the Robots and Automation
Everybody seems to think that “Blue Collar” jobs are at risk
But the real risk is to “White Collar” jobs – advances in computers and software mean that a computer will be able to replace almost all workers who shift information
Jobs that actually shift things in the real world – do physical work – are much much harder to automate
You can easily make a mechanism to fry hamburgers – but the rest of the McDonald’s staff tasks are a lot more difficult to automate
Have a look at DARPA’s various robotic challenges – opening a door is difficult
Now think about getting the food from the delivery van – storing it – bringing it out – opening the boxes – preparing it
Cleaning the floor – cleaning the fridge, cleaning the cooker
You could rebuild the whole “assembly line” – but it would still need to be cleaned and maintained
It’s easier to make a replacement for a Bank Manager than for a Janitor!
Hey Duncan – Blue collar workers are not replaced directly with machines. The environments in which they work are redesigned to allow for easy automation. Sure it’s difficult for a robot to turn a knob and open a door. You don’t need new technology to fix that; you get rid of the door knob.
Warehouse men were blue collar workers in the most classic sense. Look: https://www.youtube.com/watch?v=cLVCGEmkJs0 So were auto workers. Look at modern automobile plants – like the one that built our Jeep – and there’s hardly a human in sight. The pace of this change is poised to accelerate if human labor becomes too expensive.
Of course there are maintenance technicians behind the scenes to work on the machines – but they are many, many times fewer in number than the humans they replaced.
Of course there are white collar jobs that are vulnerable – in the future. Internists, travel agents, real estate agents, all come to mind. But their numbers are small by comparison. Self-driving trucks threaten over 3,500,000 jobs *in the US alone*. The automation of most white collar jobs is fairly far in the future. Jobs in engineering and the sciences are for now, pretty safe. As are those in the arts. Only in the distant future will the financial futures of companies be entrusted solely to machines. (And I’m not talking about Excel or SAP). In quality control, machines currently do the bulk of detailed inspections, but the systems and broader concepts are a different story.
If you have a lot of “work” concentrated in one place – a big busy warehouse or a car assembly factory then it’s worth automating
I suspect that is about 5% of today’s “Blue collar” work – if that
The other 95% is out in the world that we live in where it’s not one doorknob but every doorknob!
Driving is physically an incredibly simple operation – and from a physical point of view is about a thousand times easier to automate than simply cleaning a toilet
– Computer decision wise it’s the other way around
We are simply a long long way away from making anything that can do the jobs that any human can do easily
“Automation of white collar jobs is far in the future”
Only if we decide NOT to automate them – in terms of what we CAN do the vast majority of “white collar” jobs could easily be automated
Much much easier to automate a Bank Manager’s job or a Shop Manager than it is to automate the guy who loads the shelves
It’s easier “Technically” – but we may choose not to do so –
I think that we will automate these positions first
It’s easier, cheaper, and it saves more money
Duncan – There are about as many truckers in the US as fast food workers. All of them could be be displaced, (with near-term exceptions for janitorial staff, etc.), soon. These two sectors *alone* account for 5% of the workforce. Agriculture saw the first wave. I don’t guess I need mention the statistics. Household robots can already vacuum your house and mow the garden. And I bet you don’t do the dishes by hand either. And washing machines and dryers; what’s left for a domestic, (like Rosie from the Jetsons), to do?
The point is that to see where we are going, we need to where we’ve come from. I think you grossly underestimate the number of unskilled jobs that cannot be automated in the near term. Those jobs will be the first on the block. Some other white collar jobs will come later, for sure. How much later is anyone’s guess.
Those are all perfect examples
Vacuum the house – those Roomba things are useless – they need more physical work than it takes to actually vacuum the place – but I will admit that we could make one that would save effort
Wash the dishes – I’ve got to load it and unload it
Mow the garden – same
You can automate “easy” jobs – like anything that involves information or something like driving where the interface is easy
But as soon as you have got to move actual physical objects in the real world it becomes nearly impossible
Frying the burgers or washing the dishes is the easy part – and is only part of most tasks
We simply can’t make a “Rosie” – and we are quite a long way from being able to
You’re right Duncan, we cannot make a Rosie. We don’t need to. Time was when domestic service comprised a fair fraction of the workforce. Today, the number is tiny. Why? Because living at the level and standards once reserved for the elite, now takes far less labor.
I had a gardener at the other house. He’d come once a week and mow, trim the shrubbery, and tidy the walks and such. One guy, and it would take him an hour. Without modern mowers, power trimmers, and leaf blowers, three guys would have been required to do it in that time. And old chap, you might recall such devices were pretty much unavailable when we were kids.
When we were kids, most all airliners had a flight crew of three. The FE was automated off the flight deck. Today, even the Airburst 380 flies with a crew of two. And yeah, the FE was a pretty much “blue collar” position.
The trajectory for unskilled and semi-skilled labor is established, and has been for a very long time. We are not reaching some asymptote, only to proceed towards it at an ever decreasing rate. And we’re not proceeding apace with Rosies or Gorts, either.
Ah! I forgot garbage collection. Now there’s something that moves stuff around for ya. Remember garbage trucks with about three guys hanging off the back, ready to spring into action grabbing cans and emptying them into the truck? Today a truck, driven by a single person, maneuvers forks into slots in a large bin, and into the truck it goes.
We needn’t continue with these examples. Twenty years before hydraulic forked garbage trucks, and dumpster trash bins, nobody saw this coming. The idea of automating trash collection was a distant pipe dream. Trash robots, scurrying around the streets! Yeah, right.
The very character of society changes to accommodate fewer laborers. The act of taking your trash a few steps to a larger bin does it. Buying milk at the grocery killed the milkman, (and presumably prevented manifold pregnancies!)
Of course, societal change takes time. Automation is faster acting, and can react to changes in economics quicker, as necessary. As Ryan said below, there is a huge wave of automation being held back by lower wages. At this point, the dikes are a bit leaky, but the pumps are keeping up. Have a labor action among the Pumpkeepers and Windmillmans Union, and it’s game over for those in the lowlands.
You are right – those jobs went – but NOT because of automation
Because of small motors and batteries and things like that – and that process WILL continue
But automation – as in weak AI is – a different process – and it will NOT take jobs from the Blue Collar workers – the White Collar workers are the ones to be worried
So this has become a damned interesting discussion! Too bad the pack bas moved on. I like to hear what they have to offer. Maybe Chris will do a piece on AI and all its possible incarnations.
To our conversation, I think we need to clear up some definitional stuff before proceeding. You mention “weak AI”. That’s a great start, but we need to flesh that out. Is Alexa “weak AI”? (Let’s shorten that to WAI.) Are there current examples? Is say, DeepMind, WAI?
I’d also offer that GIA, (general AI), be avoided here, as we then step into completely speculative realms where not Bostrom nor Tegmark can help us much. There be dragons…
We haven’t moved on. Just watching.
Narrow AI is here. DeepMind is the best player of the most complex board game on the planet. I suppose if I were a white collar, professional Go player, I’d be worried. But even with all its complexity, Go is still far narrower in scope than say, electrical engineering. Or accounting. Or the law. Frankly, I can’t think of a white collar job than is as narrow as Go other than maybe internal medicine, maybe radiology, and… a few more?
(Hey, Chris. Hope you’re doing well!)
Accounting and law are pretty narrow. Narrow enough in fact that the first wave of the simplest automation has made much of both professions redundant. 90% of tax preparers would be out of business if Congress just allowed the IRS to give us access to its own automation tools (Quicken has been blocking this). And there’s no technological reason why most common legal work can’t be replaced by technology.
Next on the chopping block – the doctors who read MRIs and CT scans https://www.enlitic.com/index.html
And once those processes no longer need to be structured for reading by human eyes, imagine what kind of insights we can potentially get from otherwise unreadable scans – and imagine how cheap they’ll be.
Tax prep has really been gone for years. Very few people are involved with actually filling in forms. The collection of the information necessary for filing is, at least for businesses, largely automated. But stragegies for minimizing tax through the myriad of possibilities like restructuring, relationships between related entities, debt management, and tons of others are, so far, way too general and complex for AI.
Radiology is indeed something else again. Humans are very good at pattern recognition, but machines with the ability to learn have made great strides, in many cases surpassing their human counterparts. So yes, as I mentioned, radiology is certainly on the block. And giant steps in diagnostic accuracy and sensitivity are in the offing as a result.
It’s harder, (for me anyway), to generalize regarding the law. Take the IP subset, for example. I can see how searching for prior art might be automated. Likewise, case precident. But litigating ownership, royalties, or even writing picket claims? I dunno.
But I agree that most of the really elementary stuff has already been automated.
I don’t think there’s much of a mystery here. Raise wages for unskilled labor, and the robots will come. The technology is here. Many, many years ago, we invested in a company that built automated POS systems for McDonalds. They were very successful, and ultimately became the terminals you use for United Airlines check-in. Raise the minimum wage to $15, or some other arbitrary nonsense, and employment at Mickey D’s will plummet. Count on it.
The Longshoremans union “innovated” containerized freight, and pretty much ended the reign of the Port of New York. There are many, many more examples. Labor is a commodity. As commodities become too expensive, they are replaced with alternatives. And for this concept, there is no alternative.
How ironic that the people’s frustration at income inequality and stagnant wages might have them elect Democrats who’ll enact every possibility under the sun to address them given the choice, only to have automation come crashing down like a tidal wave as a result.
Ironic indeed, Ryan. It’s a classic example of Merton’s notion of unintended consequences.
I am not a good enough historian to dabble in this concept, but here goes anyway. I really need Chris to tackle this subject and either confirm or destroy my argument.
I am wondering if this latest age of robber barons is different than any others.
When the industrial revolution started in England, workers trashed the looms to protect their jobs, but it seemed that in the long run more jobs were created and workers needed, though I am not sure the quality of life of the workers improved much.
Then in the Gilded Age of the U.S., a few amassed insane fortunes. Rockefeller with Standard Oil apparently had wealth equivalent to 1.5% of the U.S. GDP when he died. Through decades of bloody, deadly fights, unions rose and workers got something closer to their fair share.
But in both those cases, humans were needed for manual labor. Wages rose out of the demand for that scare commodity. In the case today, how are wages going to ever rise if the demand for that commodity drops?
The owners of the capital, and robots, win. Everyone else loses. And BTW, I am actually looking forward to when robots start driving long haul trucks and a million men lose their jobs with no skills to fit into a high knowledge economy. The violence that will ensue will make the Luddit uprising look like a garden party, and hopefully, that violence will be targeted at the correct people.
Violence is almost never directed at the correct people.
I don’t agree with Din in practice – the French and Russian Revolutions come to mind. Violence necessitates strongmen leaders, sacrificing individual good for the greater cause, tribe mentality, centralization, authoritarianism, etc. Liberal systems take time to evolve and if we try to make progress through violence we’re most likely to reset and end up with feudalism/extreme conservatism/totalitarianism.
But the theory here is interesting. Production is a function of capital and labor. Technology increases the relative importance of capital, and the declining relevance of labor drives down wages. So, if we correct for low wages via government intervention (which we obviously should) then the common man’s standard of living is no longer based on what they can produce; it’s based on the political force of voting and the phantom threat of riots or revolt.
The disruption of the robber baron age led to the social safety net, inspired in no small part by violence in Russia and here at home. I don’t think this new wave of job replacement is quite so much larger than the industrial revolution, it’s just a continuation of a long process. Capitalism invented the job. Now capitalism is finding replacements for that prior innovation. Now it’s up to our culture to take all of those necessary innovations like salaries, health insurance, and retirement that we previously attached to the concept of a job, and find a new ways to deliver them.
You are completely wrong about the unemployment rate
The idea is that “Full Employment” is when everybody who wants a job has one so that if you want a worker you have to “buy” him from another company
Then there was the idea that people shift jobs and take a few weeks as a break between jobs – so a 4% rate was full employment
Makes sense – shift jobs every 3 years (150 weeks) and take 6 weeks off as a break
But think about it – in the USA 90% of workers live from one paycheck to the next and leaving a job means that you are vulnerable to any health issues
Effectively NOBODY moves jobs and takes a six week break
That 4% is millions of people who want a job and can’t get one! – and that does not include all of the people who have given up
When unemployment gets down to 1% – THEN you may be at “Full Employment”
Why do we say that “full employment” is somewhere around 4.5%?
Economists use the term “frictional unemployment” to describe why unemployment can never reach zero.
For starters, take a look at this stat from BLS, # of unemployed persons per job opening, https://www.bls.gov/charts/job-openings-and-labor-turnover/unemp-per-job-opening.htm. Today, that stat is near a record low, almost a statistical impossibility of 1.1 unemployed people per opening. It’s been below 1.5 for almost three years. That means almost no one gets hired in this economy without being poached from another company.
Let’s say, you’re an accountant working in an economy running at full employment. The small firm you work for breaks up, perhaps due to the death of a partner and related issues. You get the news that your job is disappearing on Friday. You have no concerns about being able to find a new position, but you won’t be in your new job on Monday. You have three months of severance and your spouse has a job.
Lots of firms in your area may be hiring, but they won’t all be a good fit. Even if you were very diligent about finding a new position instead of taking a few weeks off, you’ll probably spend at least a month, maybe two, interviewing and filtering down choices. During that time you will be unemployed from a statistical perspective.
The same frictional dynamics apply for lower wage workers. Your construction company completes a project and no longer needs your particular specialty, the one for which you earn the most money. Perhaps you could find a lower earning position in a few weeks doing general carpentry, but if you hold out a little longer you expect you could get a higher earning job in your electrical or plumbing niche. That guy might be unemployed in an economy at full employment for a couple of months.
Then there are tougher cases. Coal miners are losing jobs in their fields that are not being replaced. There is no other comparably-paid work in the lousy places they live, and there won’t be anytime soon. There are comparably-paid blue collar jobs out there, but they aren’t qualified for them (he’s not a welder or machinist) and they’d have to move. Any economy will always have some percentage of people in “frictional” circumstances like this, just the grist in the economic mill.
And finally you have some percentage of the workforce who are just, ahem…difficult to employ. They’re drunk or irresponsible or violent or corrupt or whatever. Statistically, that group should be large enough to keep the unemployment rate at 3-3.5% all by themselves.
So, “full-employment” signifies an economy below about 4.5% unemployment, and it is a sign of considerable economic strain. This is not just unsustainable, it is a harbinger of a looming crash.
Workforce participation rates would seem to be a big question mark here. The participation rate would make a huge difference in how unemployment affects society, certainly.
Chris, how does the equation factor in people that hold down multiple jobs?
Example: The Oklahoma teachers that wait on tables as well.
How does the BLS work with that? What would happen if every single person made a living wage from a single job?
My kid wants to get a job to earn some money to spend back at school in the fall. He has no interest in supporting a family. Hell, he doesn’t fully appreciate what it takes to support himself. If all legal work paid enough to support a family, he wouldn’t be able to get a job. Blanket assumptions about what should and shouldn’t be in the world often create unintended consequences.
As for people who hold down multiple jobs, I’m not aware of a formal statistic for that. We’re also missing reliable statistics on the number of people who are self-employed. Current methods (measuring the # of 1099 filers) are probably inaccurate.
Not that this is going to clear anything up, but here’s a bit with a chart comparing numbers of nonworking people to numbers of job openings.
A couple of other reasons for a “full employment” number that isn’t at 100%. I’m not sure if it’s still the case but auto manufacturers used to shut down production for the model year changeover. Everyone that wasn’t needed for the retooling was laid off. It was understood that they could collect unemployment during this period, and they did. Everyone knew they would be called back in a month or two. Also in northern climates, when it gets too cold to dig or otherwise work outside, the workers collect unemployment and add to the unemployed rolls.
All that said, I’ve always been convinced that the full employment rate was calculated to make sure there was always someone to hire to take the job you valued. And to keep the worker from getting too uppity.
Hi Chris the problem with that is that it’s bollocks
You simply don’t “leave” a job until you have another one lined up – because if it takes a month to get another one then you will be 5 years making up for the loss
If there are very few applicants it’s because the wages are too low!
The UK – back before the Thatcher disaster had 30 years of unemployment between 1% and 2%
I’m calling bullshit on those post-war numbers. The official, semi-consistent unemployment statistics for the UK start in 1971 for some reason. Not sure why. It appears they used more informal methods previously.
There’s a paper explaining the gap in unemployment rates in the US and the UK during that period which also explains the different methods used. Unfortunately, it’s only available on JSTOR. If somebody has a JSTOR account maybe they can shed some light.
When “those jobs” disappear due to automation or whatever, we’re freed up to do other things. This is an opportunity, because there’s no end of useful things to do. The challenge is to revamp our economic system so that necessary resources are put towards doing those other useful things.
If the economic system can’t make those changes on its own, the changes will have to be forced onto it, because economic systems exist to serve people, not the other way around.
When you have a flood of Immigrants (many illegally hired) willing to work in industries like fast food for $7.25/hr. There is NO supply and demand market force to raise it. Enforce illegal hiring laws, implement E-Verify and impose SERIOUS fines on every business which benefits from illegal labor and the wages will rise NATURALLY. Until then there is little need for expensive automated technology.
Bart, think about what you’re saying for a moment. Even if what you’re saying *were* true (it’s not, but for the sake of argument, let’s run with it), the fact remains that we’ve an economy at effectively full employment right now. If conventional wisdom held true, you’d expect to see rising wages because of the tight labor market, and yet they remain modest at best. Why?
Conventional wisdom *always* lends itself to being flipped on its head when the right circumstances come along, and as Chris mentioned, we’re living in a time where a comparative few have amassed an extraordinary amount of market power – power which, at least for the moment, is letting them do more or less as they please. Why rock the boat when the status quo (low-wage earners with effectively no real power and a government content to just cut taxes and nothing else) is working just fine for them?
Speaking of that boat, it’s more like a leaky dinghy with the Mother of all Typhoons on the horizon, and barely anyone has life jackets. Why’s that, you might ask? Well…
Honestly, the uncertainty in our situation should really give you pause as to just how potentially dire our situation is – the reason for that being just how tenuous the status quo is, and the recognition that the longer it goes on, the greater the shock to the system. We’ve a truly extraordinary number of factors (economic and political) coming together to form a Perfect Storm, and I doubt there’s so much as a single individual that rightfully grasps what’s going to happen when it finally hits.
I certainly hope you didn’t think I am advocating the continuing of the “Status Quo”, The massive influx of low-skilled workers definitely afffects the normal market pressure to raise wages. I find it to be working against their own interests for organized labor to continue to support it while those with the power to gain from it do as well. That is why applying “Conventional Wisdom” is what is wrong and simply claiming my assertion is without any support isn’t. Forcing the economy to change to serve the people against the forces of the free market is EXACTLY why we are in the current situation of why the income inequality gap has now reached its biggest gap in history.
You assertion’s effectively that it’s all immigrants’ fault that wages have stagnated, and my response to that is that that’s nonsense. In fact, some studies have found that immigration actually causes wages to *rise*, not fall. Why? Because it prompts Americans to go back to school, get a higher education, and find newer and better jobs.
So let’s get real here and face the facts. Stagnant wages are the result of a variety of economic and political forces, and it’s going to take a comprehensive strategy to get them rising again. We’re not just going to kick people out and have everything be A-OK again. That’s just not how this is going to work.
In in my field Utilities (power and water) automation has been well entrenched for many decades. You still need humans to shift through the reams of information interpenetrating them and still need the mobility of human hands and intelligence to deal with the unexpected which is a main component of most jobs. Take Homer Simpson and invert that show 180 degrees. Most power plant workers are mentally gifted and scarce. Which is why people from all over the USA and the world were my co-workers in the plant I retired from. I lived in the world you described. Where a few high skilled workers make a very good living and the less skilled did not. I think those people will take the green energy revolution in stride. The office help and unskilled labor were poorly paid. I really think a Universal Income is going to be needed to keep social unrest in checked. And we can I believe afford it as our productivity seems only limited to human ingenuity.