A good survey at The Economist on the coming tsunami on unskilled labor, for which no government is prepared. They’re right that the most important thing is to reform K through post-grad education, root and branch, but there are a lot of entrenched interests that will continue to fight that.
18 thoughts on “Jobs And Technology”
Comments are closed.
Industrial-era schools were built to turn most kids into mindless production line drones, and a small fraction into middle managers. A system based on forcing kids to spend the best part of the first twenty years of their lives doing what others tell them is about the worst possible preparation for the post-industrial future.
Reform rarely fixes something that’s broken by design. And there are far too many entrenched interests–not just teaching unions, but the whole political structure that depends on schools indoctrinating lofo voters to support them–to make radical reform possible.
Agree that reform is unlikely. Alternative to reform? Incremental sidestepping and replacement. A bit like girdling a tree. It doesn’t die or fall over right away, but eventually it does.
The problem is, you don’t have time for slow reform; kids coming out of school in fifteen years will have been crippled by industrial-era schooling today.
The obvious option for sidestepping is something like school vouchers, which allow parents to simply remove their kids from government schools. But teaching unions will fight that to the death.
There’s another key element in addition to education: regulation/taxation. Right now every aspect of “the system” from public education (all the way through labor laws, taxes, and business regulations) is aligned with the model of individuals serving as worker-cogs in the industrial machine. The system is setup in opposition to new businesses and self-employment. Part of the reason why startups and self-employment (consulting/contracting) tends to be focused primarily in the software realm is because the margins and RoI there are often so high that they can tolerate otherwise huge regulatory and tax disincentives.
But how much is the economy and especially individual workers suffering because of lost opportunities due to this onerous system? Many people with just a high school level of education nevertheless manage to acquire lots of unique and valuable skills. Yet often times those skills go to waste, economically speaking, because of the difficulty of startinga business. Fortunately the internet and things like crowd funding have helped but there is still a huge weight of regulation keeping countless businesses from sprouting up. And that, ultimately, is the real reason why so much of the work force ends up in the “unskilled labor” bucket. Not just because education sucks, but also because it’s so hard to fight against the system which funnels people into that bucket as a matter of implicit policy.
This illustrates some of the strange cognitive dissonance surrounding regulation of business. Anti-business advocates wouldn’t trust a business to run a lemonade stand, but they will trust it to overcome all the obstacles that they throw in its path and fulfill all the policy goals on their wish list.
Exactly. It’s the same sort of cognitive dissonance that leads folks who have the most distrust of the government and authority figures to shift power and wealth into those hands. Think of who, politically, would be more likely to agree with the statement “fuck the police”, yet think of who, politically, would hold the position that only the police should have guns or more that the police should have a monopoly entirely on lawful use of force in general. Shockingly it tends to be the same folks.
Logically it doesn’t hold up. Some sort of fantasy ideology is required for these ideas to exist in the same skull.
The denial of tired Luddite tropes without any understanding of why they’re wrong is hardly encouraging.
Let’s take a single example: bank tellers. We’re to believe the ATM put them out of work. What actually happened is that more people got bank accounts because they were using them in different ways, and automation was required to service the new demand which was of a different kind.
Computer use didn’t “replace typists” either, it was fundamentally different.
When someone blindly repeats a myth of Marxism – cottage industry weavers were displaced by the Industrial Revolution – without correcting it (they weren’t displaced at all, they just became so niche that we stopping thinking of them as the industry) then it doesn’t matter what the rest of their article says. Their goal is to get you to accept the myth as fact.
While there’s some truth in that, I can’t see how anyone can deny that many jobs that existed a few decades ago have almost vanished today. I was watching documentaries from the 50s some time back showing the men who used to work on the London docks loading and unloading cargoes before container ships, or used to work for the Post Office in London, moving letters around before computerized sorting offices and e-mail. Sure, jobs still exist in those fields, but far less than sixty years ago.
Similarly, I read a book about people developing new technology during WWII in the UK (I forget now whether it was a new aircraft or weapon), and at one point they needed a prototype built, so they just got in a car and drove to one of the big industrial cities, where they visited a few dozen small engineering firms until they found people who could build all the parts for them. Today? Good luck… those old factories have probably been turned into apartment blocks.
In my own line of work, aircraft flight engineers have pretty much vanished, and air traffic controllers can do much more with the same number of people, or the same amount of work with far less people, now they can just click a mouse to send a message to an aircraft rather than having to call them over the radio. Many of the old satellite communication stations have been closed because one station can now do what required dozens in the past, and the stations that still operate may no longer have full-time staff as advances in electronics have made them more reliable and advances in communications mean we can do most things remotely.
Trent, would you say there’s just as many farmers as there were 200 years ago? Or just as many manufacturing employees as 20 year ago? Obviously neither is true. Sometimes job categories shrink.
There’s more farmers than ever before. It’s simply a myth than automation killed farm jobs. It actually made farming more profitable which drove more people to farming. The statistics you hear about about “independent farmers”, a term that basically means whatever the statistician wants it to mean.
Your statement about farmers is incorrect, at least in the US. At the beginning of the 20th century, over half of all Americans lived and worked on farms. Today, the percentage is less than 2%. Even with today’s population being much greater, the total number of farmers and farm workers is much lower today due to the incredible increases in productivity due to mechanization, fertilizers, and perticides.
Many types of jobs have almost disappeared in the last 50 years. Corporations and governments used to have large bookkeeping departments but today, that task requires fewer people. Unskilled laborers on construction sites were once a common sight but not today. It’s the low skilled people who are becoming not only unemployed but unemployable as even mundane jobs are requiring skills. These people are likely left to a life of permanent welfare status.
Point of order – it’s not “unskilled” labor that’s being automated. Anything that can be reduced to math or algorithms is in danger. Like accountants – or even Doctors. IBM’s Watson is already better at diagnosing cancer than an oncologist.
Also consider how programs like Autocad could get much smarter. An end-user, maybe with the help of an interior designer, could design their own house in the software and the computer would make sure it was compliant with building codes and the limits of engineering. One-click to have the parts assembled, packed an shipped from a warehouse to be assembled on site – which automates away a lot of middle-skill workers in logistics.
Speaking of logistics, those people are pretty well trained and paid too. But a Traveling Salesman program on a massively-parallel computer could be better.
Also, BitCoin is a real danger to banking, finance and law. The current coins aren’t a good replacement for money, but then Compuserve wasn’t a good replacement for mainstream news or local bookstores. First iteration. But if finance goes open source, lots and lots of complicated finance that generates big fees gets reduced to code. Contracts like loans or letter of credit get reduced to code.
This is great for consumers of financial services. Cheaper and more abundant anything is great. But it’s going to be disruptive – and it’s not going to be just the unskilled.
To be fair, doctoring has largely been a ‘database in the head’ job for decades, and many doctors’ jobs could have been eliminated if not for governments treating it as a magical occupation that must be protected. What most doctors do is not a particularly skilled job, and, when I was at school, it was the third-raters who went into medicine, not the top of the class.
The last few times we’ve had to visit a doctor, we already knew what the problem was because Google told us, and we just had to get the doctor’s signature so we’d be allowed to buy the drugs Google told us we needed. That’s a huge amount of money wasted for trivial work that could otherwise be replaced by a search engine, and automated diagnosis seems likely to replace much of the non-trivial work as the technology becomes cheaper and more sophisticated.
Improving K-12 ed. now will be almost useless, because automation is going to improve faster than education, no matter how many reforms we make.
I also don’t buy that the dislocations that have led to the current underemployment are going to resolve and millions of new jobs will be produced in new market niches. Those new niches are going to automate from the git-go. Furthermore, at some point, demand saturates; human beings only have so many hours in a day that they can use to consume stuff. If we’re not there now, we will be some time in the next 20 years.
Bottom line: productivity will continue to increase faster than GDP does, removing more and more jobs from the workforce. We’re looking at large numbers of unemployable people. We’re going to have to support them with something like a guaranteed annual income or negative income tax.
I don’t know how you solve the moral hazard of distinguishing deadbeats who could be employable from those whose labor is simply not worth enough to be hired. My guess is that you don’t. This is the welfare state writ large, and I’m not happy about that, but I just don’t see a solution. It may be that the best we can do is to make it a nice, simple welfare state that doesn’t intrude into its clients’ lives very much. Replace TANF, SNAP, EITC, section 8, Medicaid, Medicare, Social Security, and anything else I’ve forgotten with a guaranteed annual income and decide what percentage of GDP you’re willing to spend on the whole thing.
The good news: everything will be really cheap, so a small mincome goes a long way.
Replace TANF, SNAP, EITC, section 8, Medicaid, Medicare, Social Security, and anything else I’ve forgotten with a guaranteed annual income and decide what percentage of GDP you’re willing to spend on the whole thing.
That’s the current leftyist solution. But it can’t possibly work, because it just encourages people to sit at home breeding more kids until the entire productive economy is working solely to provide a guaranteed income to the unproductive. Then it all collapses.
The real solution is 3D printing and other technologies that allow people to live on very little income. There will always be things that people are willing to pay other people to do, at least until AI makes us obsolete.
…it just encourages people to sit at home breeding more kids until the entire productive economy is working solely to provide a guaranteed income to the unproductive.
Your assertion here rests on two assumptions:
1) That people who could be productive won’t be productive because they’re lazy. There are always going to be some people like this, but I assert that we’re not preventing that today with things like SNAP or SSI disability, so mincome won’t be much worse. You can also mitigate this problem by setting mincome at a level that’s only high enough to avoid abject misery and ensuring that supplementing one’s income with productive work has an advantage. For example, for every n dollars you earn on your own (where 1 < n < 5ish), you only lose one dollar of mincome. (NB: That are zillions of implementations of this, and I'd be happy to debate the various merits of them, but this comment is already getting long.)
2) That the number of people entering the system is unbounded and can swamp the rest of the economy. But you yourself said that things like 3D printing and other tech will allow people to live on very little income. So if you can support 75% of the population on 15% of GDP, I'd contend that that's not going to lead to economic collapse.
Ultimately, all systems like this depend on the patience of the people footing the bill, i.e. the productive part of your economy. In the US, we've been remarkably consistent about what we're willing to pay to support the federal government: It's about 20% of GDP. That number is almost totally immune to any changes in tax policy or enforcement. Because of that, I've always advocated that the amount we spend in entitlements be set by working backward from what we're willing to pay. Say we decide that that number is 15% (it's about 14% right now from the feds). That tells you how much money's available that year. Then you figure out how many people are eligible, and how much they make on their own (you can do this year-by-year quite accurately), and that yields what this year's mincome level is. That way, you've got negative feedback: the more people enter the program, the lower the payout per person.
You also wind up having an honest debate about how much to spend. There's no hiding things in fifteen different programs. There's no outright denial that the funding levels of things like Medicare are inadequate to the task. And if you decide to raise or lower the amount you spend, you know exactly what it does to the level of taxation or to the discretionary parts of the budget, and the constituents for that stuff will know exactly how much they need to fight.
I've been a laissez-faire, small-government, libertarianish sort my whole adult life, and I'm not happy about this. But over the past 5 years or so it's become pretty clear that we're in the grip of historical, not political, forces. The problem is that, unlike previous labor dislocations, automation isn't a point innovation that does its damage and then gives the labor market time to adjust. Automation just keeps getting better and better, and it does so at a rate that's faster than any possible adjustment the labor market can make. Lots and lots of otherwise hard-working people are going to be permanently unemployed. We have to deal with that. I prefer we deal with that in some way that doesn't involve them and their families starving to death. I'd also prefer that we deal with it honestly, not by setting up warring constituencies fighting for dollars for their own pet program at the expense of the other fifty pet programs. And, finally, I'd prefer that we deal with it realistically, rather than doing things like paying unemployment benefits to the permanently unemployed in an open-ended fashion that can't be controlled. Mincome looks like it fills the bill.
You sound like the British welfare statists of the 40s and 50s. Obviously giving people money for nothing won’t result in an underclass who do nothing productive, go for generations without working, and breed faster than the workers who have to work long hours to pay for the welfare state.
Oops.
People are lazy. If they’re paid enough that they don’t have to work, many, if not most, won’t. If they aren’t paid enough that they don’t have to work, the whole thing will be a dismal failure. Unless you believe you can change human nature and create The New Soviet Man, of course.
You might be right. Women on welfare have 2-3x the fertility rates of women not on welfare, although there’s considerable debate about whether they can’t/won’t control their fertility because they’re on welfare or whether they’re on welfare because they can’t/won’t control their fertility. The answer to that question is pretty important to this discussion.
I have a hypothetical for you: Suppose that the behavior of those on the dole is about as bad as you say, but also suppose that we discover that we’ve got 25% of the 18-65 workforce being long term unemployed. (Note that this is something not well captured even in U6, because the long-term unemployed may be discouraged workers for longer than 6 months.) That would be indicative of a vastly different labor situation than what we’ve seen historically. What do you do then?