Lighting Fires and Filling Buckets
Policy Ideas for the Age of Automation, Part One: Education
Everyone is starting to become concerned that the machines are about to take away all of our jobs—at least, all of the jobs that we do now. Everything is on the chopping block to be automated, from flipping burgers to driving trucks to filing legal papers to making a medical diagnosis. Even writing computer code is supposedly going to be automated. With the rise of artificial intelligence, this is all supposed to happen extremely quickly, leaving a vast population of superfluous workers.
A lot of this is overhyped and exaggerated, of course. The process of transformation is likely to take a lot longer than predicted, and there is good reason to think that while many tasks will be automated, fewer actual jobs will be automated, so in most cases machines will end up augmenting human workers instead of replacing them.
Those of us who are technological optimists can point to the history of the Industrial Revolution as evidence that a big economic and technological transition might seem terrifying at first, and it might eliminate or reduce some of the old jobs that we're familiar with, but it will produce many new jobs that had never been conceived before, along with enormous amounts of new wealth for everybody.
Yet the example of the Industrial Revolution is not quite that reassuring. It was a wrenching disruption, not everybody adapted to it successfully, and the result was an age of political and cultural disruption that provided the roots for populist, nationalist, and totalitarian movements. It's not something we necessarily want to repeat.
That raises the question: can we do better this time? Are there things we can do to make the transition to the new era of automation faster and smoother? What government policies can make the problem better or worse?
I look at this question from the perspective of a skeptical free-marketer who doubts there's much government can do to predict what the future will look like or to help people adjust to it.
For example, the obvious answer to the disruption caused by automation would seem to be government-sponsored job retraining. But that's been tried before as a cure for unemployment and has never made much difference. Not everyone who needs a new job has the initiative to sign up for the retraining. Worse, the government doesn't always know what the skills are that people need to be trained for. Especially with the fast pace of technological change we're expecting, we run the risk of training people for new jobs because their old ones have been automated, only to find that the new jobs get automated, too.
Then there are the unintended consequences. You can throw billions of dollars in funding and subsidies and loans into higher education and push people to go to college, on the theory that jobs that require a college degree are less likely to be automated. But you may also find that a lot of that money ends up being absorbed by bloated college administrations, and when you push young people to go to college, they might end up taking on a lot of debt for degrees they don't complete.
And don't get me started on the downside of the Basic Income. (Too late. I'll have more to say on it in a later installment of this series.)
Perhaps the better focus for government policy would be on figuring out how to remove the roadblocks that make it harder for people to make their own adjustments to the new era of automation. In that spirit, I will propose five broad policy ideas for the new era of automation, beginning in this installment with the first and probably most urgent: education reform.
In an era when old skills are being rendered obsolete and new skills will be required, education is the most important way for everyone to adapt to the new era.
But just throwing more money at the problem has been a disaster, resulting in (among other things) soaring costs and debt as colleges just increase their tuition to soak up the extra federal money. The same can be said of the attempt to make another two years of education beyond high school mandatory, which is a confession that existing high schools are failing to turn out students with actual job skills.
To help people adapt to the new era, we desperately need to reduce the expense of education, but more important, we need to change how we approach it. Education is currently treated as single event that consumes a lot of time and money when you are very young, up to age 18 or 22, and after that most people never return to any formal, organized, serious kind of education. Yet that is exactly what they're going to need to do, and education needs to be instilled as a lifelong passion.
Here are some specific policy proposals to help us get there.
1. More choice and competition in primary and secondary schools.
In addition to teaching students the specific knowledge and skills they need today, the primary and secondary schools—everything young people do before college—need to instill a love of learning and the skill of knowing how to learn. They have to prepare students for education as an ongoing process.
Unfortunately, public schools tend to be bureaucratic and sclerotic and have not done a good job of this. The past two administrations have tried to reform public schools by imposing standardized tests that are now universally hated by teachers and students alike. William Butler Yeats said that "Education is not the filling of a pail, but the lighting of a fire." Standardized testing has the schools busy franticly filling pails that they hope will remain full long enough for the next round of tests.
I suspect the only way to reform the whole system is to remove it from the government bureaucracy, through charter schools that are given more freedom from union rules and bureaucratic restrictions, and through school choice—some sort of vouchers or tax credits that will make private schools affordable for more people.
In a competitive marketplace, products and services tend to get better, not worse, and when customers aren't getting what they need, they have the freedom to take their business elsewhere. It's a powerful incentive that is missing in today's public schools. It has certainly been my own experience, with my children, that a private school is far more responsive to the parents and much less responsive to any internal bureaucracy.
It is also very expensive, a cost that some of us are fortunate enough to be able to bear (more or less). School choice would help make that possible for many more people.
2. More focus on blue-collar skills and apprenticeships.
Public schools tend to produce two kinds of students: those who go on to college, and those who emerge with no marketable skills. Over the decades, they have de-emphasized the middle: students who need advanced blue-collar skills. The result is that employers are going begging for skilled mechanics and pipefitters and welders, well-paying jobs that are less likely to be automated. These skilled professions are too complex to be done entirely by robots, at least not any time soon, and are instead more likely experience technological advances as increases in productivity rather than a threat to employment.
Again, the important thing isn't just learning a specific skill, it's learning how to learn a skill. It's a lot easier to retrain yourself from, say, diesel mechanic to welder, when you already have a base of mechanical skills, or to retrain as the operator of some fancy new semi-automated equipment, than it is to go from flipping burgers to a highly skilled job.
3. Apprenticeships for white-collar jobs.
Lower-skilled white-collar jobs including some forms of computer programming and graphic design need to be treated less like work that requires a four-year degree and more like blue-collar skills. There is a good argument that coding should be the new coal mining—a new well-paying semi-skilled job for non-college-educated workers.
One of the most exploitative parts of our current system of education is the way students are charged six-figure four-year tuition for skills that don't require four years of full-time study to learn and are unlikely to produce enough income to pay back the loans. Before they were absorbed into the universities, semi-skilled white-collar jobs and jobs in the arts were taught through apprenticeships, just like the work of skilled craftsmen. Having a fine arts degree might carry a lot of prestige, but when it also carries a giant student loan, maybe the prestige isn't worth it.
4. Rein in the cost of higher education.
Higher education is the most universal skill of all. Work that truly requires a college-level education—in law, finance, engineering, scientific research—tends to involve the kind of higher-level abstract thinking that cannot be automated, and those who have been trained for this kind of thinking are the most flexible workers, who are most likely to view a changing economy not as a threat but as an opportunity to rush into newly discovered fields of work.
If this kind of high-level knowledge work is going to be more important in the era of automation, then we need to make it possible for more people to acquire an advanced education without having to go into ruinous debt. A few years ago, by some calculations, we passed the point at which it is no longer mathematically possible for a bright and dedicated young person to work his or her way through college. There aren't enough hours in the day to work a regular job while also paying for astronomical tuition fees. You can see how this is a formidable barrier for students from lower-income families, who can't pursue higher education without taking on crippling debt.
The counterintuitive solution is fewer subsidies, with less emphasis on encouraging students to take on debt and more emphasis on making regular tuition affordable without subsidies and loans. That's going to mean stripping down a college education to its practical essence, treating it less like four years of spa and party time, and making massive cuts in the higher-education bureaucracy.
Enterprising state governors who want to help their constituents forge a viable path to the future of work would do well to follow the example of trying to craft a $10,000 four-year college degree at state schools. Even if $10,000 is unrealistically low, it would be a huge step forward just to get close to this goal.
5. Make existing higher education obsolete.
For better or worse, the Internet has conditioned everyone to get all sorts of information, from breaking news to encyclopedias, for free. So how come the most important information of all—higher education—is more expensive than ever?
So-called "EdTech" has been experimenting with Massively Open Online Courses, or MOOCs, which make the best teaching from the greatest minds available to anyone. In the age of the Internet, the problem of how to transmit information quickly, inexpensively, and on demand has been thoroughly solved. So why hasn't it been solved for education? The reason is that when people pay for college, they're not paying for the actual knowledge. They're paying for a certification—a piece of paper that marks them as employable.
The real challenge for the disruption of higher education is to find alternative ways to provide this kind of certification, at less ruinous expense for students. To the extent that professional regulations help entrench traditional universities as the only acceptable source of educational certification—particularly in fields like law, architecture, and medicine—they help prevent new technology from offering competition. But hobody has quite solved this yet, and it's going to require a change in mindset among major employers.
Remember what the challenge is in the age of automation. If robotics and artificial intelligence make whole categories of existing work obsolete, workers are going to have to adapt quickly and acquire new skills—and possibly do so again another ten or fifteen years in the future to adapt to a further wave of automation. More to the point, the concern is that automation will target every job that involves tasks that are routine and repetitive and therefore easily imitated by machines. To compete, humans will have to offer increasingly abstract, higher-level thinking skills that computers cannot replicate. Both of these trends, to the extent they are occurring, make education and the development of higher-level intellectual skills all that more vital. They make education reform into a crisis-level priority.
The ideas I have suggested above are the kinds of things we would do if we took that crisis seriously and were really concerned about solving the problem—instead of being concerned about appeasing the entrenched constituencies the current system serves.
This is a theme that will continue in the next installment of this article, when we turn to another area with a lot of entrenched constituencies: urban policy.