James Bessen on Learning by Doing
May 23 2016

Learning%20by%20Doing.jpg Are workers being left behind when the economy grows? Is technology making the human workforce obsolete? James Bessen, author of Learning by Doing, talks with EconTalk host Russ Roberts about the role of learning on the job in the past and in the present. Bessen argues that during times of technological innovation, it often takes years before workers see higher wages from productivity increases. Bessen stresses the importance of the standardization of education on the job as workers adapt to new technology.

David Autor on the Future of Work and Polanyi's Paradox
David Autor of the Massachusetts Institute of Technology talks with EconTalk host Russ Roberts about the future of work and the role that automation and smart machines might play in the workforce. Autor stresses the importance of Michael Polanyi's insight...
Andrew McAfee, Megan McArdle, and Lee Ohanian on the Future of Work
Andrew McAfee, Megan McArdle, and Lee Ohanian talk with EconTalk host Russ Roberts on the future of work. Recorded before a live audience at the 33rd Santa Barbara Economic Summit, the conversation begins with each participant making a brief set...
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


May 24 2016 at 10:06am


I was happy to see that you pushed back on the optimism of rapid and drastic changes in technology. I know that you tend to believe that technological changes will benefit most people in the long run. However, I do believe as you seem to suggest that the transitioning period will be accompanied by a lot of pain. People do not simply learn new skills overnight.

Do you believe this is a case where we should try to do something to ease this transition? Or do you believe we should leave the market to clear itself and let the pain be felt until if clears?

May 24 2016 at 10:42am


Another great episode.

Under Australian common law, a “no compete” clause in an employment contract (or a restraint of trade clause, as they are known there) is primae facie uneforceable by the employer. There is a long stream of judicial commentary which considers them contrary to public policy.

Even if an employee voluntarily signs a contract with, say, a 2 year exclusion provision, if the employer wishes to enforce it the onus is on it to bring proceedings and prove to the Court that the provision is reasonable – that will include whether it is necessary to protect business good will or trade secrets. If they cannot do so the employee is not restrained.

That makes it easier for employees to move around and makes employers tend towards less restrictive provisions. A healthy balance, I think.

Glenn Mercer
May 24 2016 at 12:12pm

Just one thought, re licensing:

Assume a ridiculously-simplified economy: workers can work either in huge factories (e.g. Ford assembly line) or in small fragmented service facilities (e.g. nail salon). If:

1. workers seek to protect their wages and employment
2. and the economy shifts from mostly huge factories to small service facilities
3. then do we see a decline in UNIONS (suited to protecting W&E in big factories) and a rise in LICENSING (suited to protecting W&E in small service facilities)?

Is licensing just a modern adaptation of unionization?

(This is probably either blindingly obvious to real economists, or incredibly stupid, but I figured I would give it a shot. If it is remotely right it explains decline in U and rise in L.)

May 24 2016 at 12:51pm

Another great home run episode discussing the need of workers to keep their skills up to date at a greater frequency that ever before.

To offer my story about adaptation:

* I started my career in film and television post production. I went to a 4 year school and got a degree. I worked for one year.

* Then I attended a one year program to focus mostly on one software program (with a few others less emphasized). It was a very focused and specialized program

* I spent so much time enhancing my technical skills and not enough time nurturing my creative skills that it turned out I was now better and more marketable at supporting others who had problems with the hardware and software package.

* The software creator is having some financial troubles and I feel that in another 10 years there will probably be something else to take it’s place. I want to learn a larger platform of technologies and make a career change into IT.

* Although it helped me to get good work, if I were to do it over again I probably would not have spent that year focused on learning one platform but would rather have gotten experience with a variety of platforms.

* Now I have access to a larger world of work.

Even in IT, people skills are really important. It’s also important to try to figure out how something works on your own, that way when you seek out help your questions will be more focused. You’ll also improve your ability to learn as you go.

May 24 2016 at 8:47pm

Wow, such a sad episode. Two experts on the subject and the best that can be suggested is that we need to do better on teaching people to be lifelong learners, and we need to somehow get businesses and schools together to train and educate people.

Just think of the optimism in America 70, 60, 50, even 40 years ago. The space age, unlimited nuclear energy, flying cars, the Jetsons, and a middle class that was the envy of the world and perhaps the greatest accomplishment in the history of mankind.

Now look at us. An article in Business Insider today states that more young people are living at home longer than ever before. Their prospects are depressing.

The lesser educated underclass of our country has to compete with a never ending flow of cheap labor, illegal and legal, while our factories relocate to south of the border. And now the well educated even have to compete with H1Bs from the four corners of the earth…..Disney anyone?

Why in the world would any big business mess with training American workers when they can import desperate souls to do everything from pluck chickens to program computers, by the hundreds of thousands?

It seems like the only thing our titans of Washington and industry think American youth are good for is getting blown apart in middle east sandpits in furtherance of ‘spreading democracy.’

It’s amazing. With fracking and the like we literally have everything we need to prosper and employ our populace right here at home, but we just don’t seem to really want to do it.

The best suggestions for a real fix I’ve read recently were set forth by Andy Grove in a piece he wrote for Bloomberg years ago about having an economy that creates jobs. A quick internet search will find it. And more recently Peter Morici wrote a great piece about the damage caused by our trade imbalances and how we ought to fix these bad deals.

Some real action like this might help. But hoping for imparting lifelong learning to the masses won’t do it.

Andrew Wagner
May 24 2016 at 9:32pm

I make two connections from this:
One) Ronald Coase’ “Nature of the Firm”: In the early days of the textile mills, the transaction costs of training a “mill girl” were high–the work was “internal to the firm” and people stayed in one place. Later, with standardization, transaction costs were lower, so mill girls became interchangeable parts of the firm. As such, the employees became able to move from firm to firm and operate more like independent concerns.
Two) Clayton Christensen’s paper “Skate to Where the Money Will Be”: Christensen discusses how technologies go from integral to modular–subsystems standardize and with standardization–break down into different firms. The money tends to go to those firms in the chain that own a system or subsystem for which the technology is not ready to be fully and interchangeably standardized. The analogy might be a stretch here, but your graphics designers all know the same software to interface with the firm, but their human touch, design sense, etc, are the “value add”. Those are things that cannot be standardized. That’s where the money is.

May 25 2016 at 12:28pm

The part about schools not working is really important and it would be great to hear you do more on this subject. I know you’ve looked into it some, but I’d love to hear more. There are several interesting experimental schools out there, some new, and some (like Sudbury Valley School) have been around for over 40 years. Are you familiar with the various democratic school models? You mention the problem of not knowing how to teach students to learn… I’ve come to believe that humans are born with the desire and passion to learn and that most schools extinguish that desire and passion. Schools like Sudbury Valley School & have found a way to keep that passion and desire alive in young people with some very interesting outcomes. To understand the model I had to take a big step back with so many of my assumptions but I am very impressed with their approach and I highly recommend you look into it if you’re not already familiar with it.

Dr. Duru
May 25 2016 at 3:34pm

In Bessem’s story about the change in wages for textile workers in the 1880s, he noted that pre Civil War, workers were paid low wages due to non-transferability of skills. But why couldn’t experienced workers command higher wages at their existing mills given their value? I would have expected a co-dependent relationship where a factory would pay dearly not to lose experience and a worker would be loathe to leave only to start over. I would have also expected the experienced worker to at least command a higher wage in a new job by virtue of proving the ability to learn a manufacturing process. Surely when firm-specific knowledge is at a premium, the ability to learn is not equally shared in the general population? It is when standardization takes root that the ability to learn is not valued as highly. Just seeking clarification.

Scott v
May 25 2016 at 9:15pm

I am struggling with the story the author tells. If the current workers had scarce skills, why wouldn’t their employer pay higher wages to keep them rather than face the opportunity cost involved in training a new worker? Why hire women who were more likely to leave to start a family? Lots of men were involved in making cloth including the Luddites who smashed power looms.

Isn’t it more likely that cloth supply expanded faster than demand? Equally standardization that allowed people to move between firms would also increase the pool of workers competing for the same position?


Dr. Duru
May 25 2016 at 9:48pm

In Bessem’s story about the change in wages for textile workers in the 1880s, he noted that pre Civil War, workers were paid low wages due to non-transferability of skills. But why couldn’t experienced workers command higher wages at their existing mills given their value? I would have expected a co-dependent relationship where a factory would pay dearly not to lose experience and a worker would be loathe to leave only to start over. I would have also expected the experienced worker to at least command a higher wage in a new job by virtue of proving the ability to learn a manufacturing process. Surely when firm-specific knowledge is at a premium, the ability to learn is not equally shared in the general population? It is when standardization takes root that the ability to learn is not valued as highly. Just seeking clarification.

May 26 2016 at 9:07am

@FredC: Regardless of where you stand on immigration and trade, the problems touched on in this podcast will not be resolved by adding restrictions on those. Sooner or later, the problem of education will come up.

Restrictions on using foreign workers will likely only accelerate the automation of domestic manufacturing. These questions will have to be confronted whatever the future trajectory of US trade and immigration.

Dr. Duru
May 26 2016 at 2:59pm

In Bessem’s story about the change in wages for textile workers in the 1880s, he noted that pre Civil War, workers were paid low wages due to non-transferability of skills. But why couldn’t experienced workers command higher wages at their existing mills given their value? I would have expected a co-dependent relationship where a factory would pay dearly not to lose experience and a worker would be loathe to leave only to start over. I would have also expected the experienced worker to at least command a higher wage in a new job by virtue of proving the ability to learn a manufacturing process. Surely when firm-specific knowledge is at a premium, the ability to learn is not equally shared in the general population? It is when standardization takes root that the ability to learn is not valued as highly. Just seeking clarification.

May 26 2016 at 7:20pm


I loved this episode. I especially liked the discussion about certificates vs. licenses because it’s extremely relevant to everyday life in the developed world. As was pointed out in the episode, having transferable skills as a worker (i.e. the textile mill worker in the age of standardized machinery) drives up the cost of his/her labor.

*Certificates* are a great move to increase worker productivity and decrease training time by providing a workable standard. They probably increase wages by increasing the transferability of skills.

*Licenses* are primarily a way to restrict the labor supply. They are basically mandatory certificates. They increase wages only for the workers that have them by the same effect as certificates as well as by decreasing supply.

I would love to live in a world where we had only certificates for the vast majority of things (like hairdressing) that require licenses today.

Dr. Duru
May 27 2016 at 3:00am

In Bessem’s story about the change in wages for textile workers in the 1880s, he noted that pre Civil War, workers were paid low wages due to non-transferability of skills. But why couldn’t experienced workers command higher wages at their existing mills given their value? I would have expected a co-dependent relationship where a factory would pay dearly not to lose experience and a worker would be loathe to leave only to start over. I would have also expected the experienced worker to at least command a higher wage in a new job by virtue of proving the ability to learn a manufacturing process. Surely when firm-specific knowledge is at a premium, the ability to learn is not equally shared in the general population? It is when standardization takes root that the ability to learn is not valued as highly. Just seeking clarification.

May 27 2016 at 9:13am

A pretty good podcast this week, some comments:

  • I agree that non-competes had some effect on Silicon Valley pulling away from Rt 128, but the overwhelming reason was the geographic proximity to far east engineering and manufacturing resources and markets.
  • Math IS hard, which is why employers pay dearly for math skilled workers. Bessen too quickly gives up on the average worker, they are capable of learning much more math than they do in our schools or via on the job training. Advanced calculus might be too unrealistic, but they don’t need that, they need solid algebra, logic and some statistics, which our school systems are failing to provide.

    (I tell any college bound person to make sure to take a logic course and a statistics course. If they understand these concepts, it is much easier to tell when someone is lying to you…)

  • I strongly agree with Bessen that although you may learn things in college, the primary role of college is to learn how to learn (despite recent anti-learning college initiatives like “speech codes” and “safe zones”). Lifelong learning is not only required in this day and age, but it is fun as well (hence Econtalk!)
  • Lastly, I agree that technology is almost always a good thing, and that workers will be displaced and that there is a significant lag while the disruption is digested. I do worry about the tremendous reduction in the labor participation rate, but I cannot blame it entirely on technology.
May 27 2016 at 1:56pm

On the non-compete clause issue, I remember having discussions with people, including a couple of lawyers on this back when I was under one. The corporate lawyer at one company said that those things were largely unenforceable. He said that a company cannot prevent you from earning a living at something near your previous standard of living unless they are willing to pay you not to work. If there is comparable work in another industry that pays similarly that helps the company, but if you have very specific skills, they have to let you work or pay you not to work. At least that is what his understanding of the law was at the time. He said that our company at the time only really tried to enforce the non-recruitment provision of the non-compete stuff. And he said that they were very aggressive about doing so.

They way this was described by another attorney was that if you had skills that let you earn $100K per year in a particular industry. If you took a job in another industry, you would only make $60K per year since your skills were not as directly applicable. If your former employer wanted to keep you from working for a firm in their industry, they would need to compensate you the difference or a court would likely rule that you had the right to earn your proper income.

On the training side, it seemed like, particularly with large training costs, companies started having employees sign agreements to repay training costs if they left within a certain period of time. At a telecommunications firm where I worked, we would spend $30K-$50K training some new employees within the first six months. At that point, they could go somewhere else and make substantially more money than they were currently being paid because of the training and certifications. So the company instituted a program where if you left within the first year, you had to repay 100% of the training costs and 50% if you left during the second year. Usually, they could figure out who were the best new employees with the first year and compensate them well enough to keep them around.

Dr. Duru
May 28 2016 at 11:33am

Sorry about the multiple posts! I am not sure how that happened! I left this page open in my mobile browser. I guess whenever I came back to the browser, it would refresh and repost. Again, my apologies!

[No problem. I’ve removed the duplicate. You are exactly right: sometimes leaving things open in multiple browsers can cause duplication.–Econlib Ed.]

May 28 2016 at 8:08pm


As long as the US persists with unfettered immigration, legal and illegal, and bad trade deals, business will have NO incentive to team with educational institutions to train American workers. Why should an American corporation spend one second apprenticing someone when they can import an H1B from india who has the pinpoint skill they need and when the project ends, ship him back in the blink of an eye? And then rinse and repeat.

The only way that businesses will sign on to this cooperation model is if they HAVE to due to a tight labor market. As far as automation, well that’s going to happen anyway. Why exacerbate our problems with unfettered immigration too. Moreover, the vast vast majority of people illegally in the US today are exactly the type of people who will lose their jobs en masse due to automation. Are you ready for the social fall out?

Lots of immigration and putting up with bad trade deals has become so ingrained in the mantra of modern economists that it’s almost religion.

On his dumbest day, Andy Grove was smarter that a room full of your average economists. He saw that America was selling out its youth and future and had the guts to go against the orthodoxy in the piece I referenced. Thankfully, even many economists now are recognizing that the way the US does trade deals needs a rethink. It’s about time.

Trade and immigration can be great. But the way we’ve been managing both for the last 50 years stinks.

May 28 2016 at 9:18pm


Are you referring to Andy Grove, founder of Intel and, ummm, (legal) Hungarian immigrant?


The theory is that when you accept the job, the salary that you accept has a premium already built into it that compensates you for the non-compete agreement. I’m not agreeing with the concept, just stating the legal justification. I’ve signed a couple in my career and am embarrassed to say that no one bothered to enforce them when I left…

May 29 2016 at 10:45am


Yes, that Andy Grove. He entered the US in 1957, I believe, after previously fleeing the nazis. Interesting that he was able to enter the United States prior to the massive change in policy that expanded immigration in the mid-1960s. I guess even under the old rules, and before the H1B, the US could occasionally score a world class brain or two. Even before that we lucked out with that other wild-haired guy who was good at math. Thankfully, both managed to enter without displacing a hundred Disney tech workers.

Also interesting that Grove, considering his background, came to the conclusion that we give away the store with our recent trade deals, and a little “protectionism” might result in more Americans having a job and being able to support their families.

Trade and immigration can be great, just not the way we’ve been doing it for long time.

David Zetland
May 30 2016 at 3:29pm

Very interesting. I was kinda surprised that the two of you struggled to come up with an easy example of standardization. How about language? English is super popular (I teach in it in the Netherlands) as a standard of communication. The language standard is also why many migrants settled into “ghettos” and refugees have trouble finding jobs in new countries (although, as pointed out this week, Canada’s program of supporting “migrants” via community organizations points out how bridges between standards can be erected).

Jun 2 2016 at 5:59pm

I disagree on the software argument. People really overstate how hard it is for a reasonable intelligent person to pick up new software that does a very similar task. With the internet and being able to post question on forums and plenty of knowledge people volunteering their free time to answer them along with you-tube videos. I would say there is only a few month learning curve on average. The reason being it the underlying process and logic of how the software does the activities are very similar in most cases. A lot of the learning is finding out were and what steps you need to due to due the same or almost same thing that you knew how to due in another software package. The exception being is if a piece of software is designed by a culture that has a different logic and thinking process. For example most ERP systems coded in the U.S. have a logic and thought process that if you understand how one ERP system works in a given module, you can pretty easily figure out with a bit of help how the other works and it is not a couple year learning process a year at most maybe. Were as take SAP is based on German coder logic. It is actually a lot harder to figure out how SAP process are doing task being an American because German coders are using German logic which is different than American coder logic.

And that is the huge underlying problem with a lot of modern formal education. They teach software not the process of how and why the software does what it does. You need to teach people how and why of doing a specific task with a given software package. Once they understand how and why picking up another package yes requires some training but no more training and probably less than what one would have to learn when you go to work for a new company.

If you remember software is not the solution on its own, it is about having good processes and software is a critical part of that process. Since most software follows best practices, it is more important as a professional to understand the process and decision making and risk factors in a given process than how a given software package changes or slightly modifies that process.

Jun 18 2016 at 12:27am

Having worked in Accounting/Finance for 10 years, I’ve seen an overwhelming problem with how companies hire people. They are simply obsessed with what software job candidates have used, with “on the job” experience being the only thing they count.

For example, if a job requires “strong SAP knowledge”, the company will hold out for an applicant with multiple years history using SAP, and question them diligently in interviews.

It is silly to suggest that taking a MOOC, for example, would matter on a job application. I work at a Fortune 500 company and I doubt most managers could identify a MOOC. Even if they could, they would have no faith in the meaning of a certificate from one.

As an employee, you are captive to the system your company uses, and you know very little recognition of “learning ability” will occur. These days, you need to have the right keyword matches on your resume, in order to have it reviewed.

Comments are closed.


EconTalk Extra, conversation starters for this podcast episode:

This week's guest:

This week's focus:

Additional ideas and people mentioned in this podcast episode:

A few more readings and background resources:

A few more EconTalk podcast episodes:



Podcast Episode Highlights
0:33Intro. [Recording date: May 9, 2016.] Russ: Now, one of the central ideas of your book is what happens with the average worker as technology and innovation grows in the economy: Do they share in the gains? A lot of people worry the average worker isn't sharing in the growth of the economy over the last decade, decades. What can history teach us? Guest: Well, history lessons are always tricky to apply to the present. What history can do is make sure at least asking the right questions. In fact, the Industrial Revolution saw problems and patterns that are not really, or seem very similar to what's happening today. I won't say they are identical; they are not. They faced different challenges than we face. But I think people often forget or neglect the history or don't know it in the first place. And I think our understanding today would be a lot richer if we did know that. Russ: Well, I was very surprised and intrigued by the historical data that you bring to bear on the question, which is, in particular over long stretches of the Industrial Revolution, pay was stagnant, or appears to be stagnant. And that led some thinkers and writers to be very pessimistic about innovation, about capitalism. But things changed. In my mind, it's just a long, steady increase in human wellbeing since about 1750. And that's not true. So, talk about what the data show, how it was misinterpreted or at least assumed, some trends, were assumed to be permanent, and what we can learn from that. Guest: People understood very quickly that machines were taking over tasks of workers. And this prompted people to get very concerned, not unsurprisingly and very similar to today, that we're not going to have jobs. You combine that, then, with a long period of stagnant wages--so, this was true both in the Industrial Revolution in Britain and in the United States. I use data for the United States but we can see that the textile mill, which was really the forefront of the Industrial Revolution both in the United States and in Britain, productivity was going up very rapidly and wages were not. Wages didn't go up for textile workers until after the Civil War. So, you had a long period--the first textile mills were the 1814, 1817--and you had this very long period of tremendous productivity growth in output per worker. And completely stagnant wages. And so people--Marx included--looked at this divergence and said, 'Capitalism is this new system that is extracting all the profits to the capitalists.' And that time, it seemed like that was right. It seemed like that was happening. And people were very afraid that there would be massive unemployment, what we call technological unemployment today. Didn't happen.
4:21Russ: So, the stagnation in wages over that period--it's a very long period of time: it's roughly half a century-- Guest: Yep. Russ: What's your explanation for why that stagnation occurred and then why things changed? Guest: Yeah. Yeah. So, the--I think the--this is a complicated explanation, a little bit. The first thing to realize is: These were not unskilled workers. It's a common misconception. It's very prevalent. We still tend to think today of factory workers being unskilled, especially at the Industrial Revolution. And it's certainly true that they did not have many skills when they walked in the door of the mills. I look in detail, for instance, at weavers. And these weavers were typically teenage girls. Some of them had woven on handlooms at home. But they had never been in an industrial-type environment with sophisticated machinery, anything like was what was in the mills. And they had no prior experience, for the most part, when they walked into the mills. The mills hired people who had never worked in the mills before and who had no industry experience. Over the course of a year or so, they would learn a tremendous amount, and their productivity would go up 5-fold or 6-fold. Many of the new entrants didn't succeed. They couldn't survive in that environment. They couldn't manage to learn, or they found it too distasteful. But the ones who did, had, acquired very valuable skills. We can actually measure those skills. And we know that those skills were important and critical to actually the mill achieving productivity. Mills that failed to get the workers up to speed failed economically. Russ: The reason is--I want to interrupt for a sec. Which is, one of the challenges, I think, about thinking about these issues today, especially, as well as in history, is that we think about the phrase, 'Machines make workers more productive.' Which is undoubtedly true in the transformation and industry you are talking about--weaving--we'll probably get into it in some detail. It's rather extraordinary how much more workers could do than with the machines than without. With the automation and without. And yet, in many ways it's the machines that are productive. And so the fundamental question is, one of the fundamental questions is: What are the workers bringing to the experience that works with the machine, as well as that makes them valuable in and of themselves relative to their alternative employment? And I think that's something people often forget. Guest: Yeah. And so, I think the general answer to that question is: Technology is complex. And it's never 100% automated. When you have something that's even 98% automated, there are still lots of things for workers to do. And, it turns out that their proficiency as doing those few remaining tasks is critical for the through-put of the machine. So, what--I mean, literally, with the textile machines, if something went wrong, the machine would stop. And it would be down. And it was the ability of a weaver to recognize the problem was developing; if possible, fix it before the machine stopped. If not, fix it quickly so the machine was down for a short period of time. And fix it reliably so that it wouldn't go down again. So, the productivity of the machines was just deeply intertwined with these very specific technological skills that the workers had to learn on the job. And this was a particular challenge then; and I'll argue it's a very similar challenge, today. From a social point of view, what was difficult about these skills was that they had to be learned on the job. At least at first. They were skills that couldn't be taught in the classroom. And many of them were very unstandardized. You would go to one mill and would do things one way and would have one set of equipment--very often the looms were custom-built for the mills. It was very different from what was in other mills. So that skills learned in one mill weren't necessarily portable to another. You didn't have a labor market developing where workers who were trained at one mill could work at another mill proficiently. What happened after the Civil War was that standardization started taking place: That the mills coordinated their machinery; they coordinated the way that they hired workforce and used the machinery. There were training school developed for mill managers. So there was much more uniformity. And what you see happening is that robust labor market developed. That, in the 1830s, very few of the new hires had any prior experience. By the 1880s, almost all of the new hires had previous experience. What that labor market did, was it meant that mills would bid up the wages. You are talking a competitive market, workers with skills that could be applied across mills. And so, the labor market meant that if I was not earning enough as a weaver at one mill, somebody else is going to pay me more. And so, that was the essential thing to start wages on their upward trajectory.
10:30Russ: It's hard to understand why it would take so long. And one argument--I once a labor economist. I still play one on TV. But, you know, what sets your wages is your next best alternative, combined with your productivity. And if there are a lot of people who can work a loom, a power loom--and it doesn't take very long to learn how to do that--you're not going to get paid a lot of money. No matter how more productive the power loom is or the person is than an artist or craftsperson, loom weaving fabric by hand. That's my perspective. And I think it's yours as well, in the book. What changed in that story? Normally, you'd say, 'Well, in the early days, it was really hard--it was really easy to do. But it got more sophisticated.' I don't think that's your story, though. Right? Explain what, in a sort of standard labor economics way, why a worker in 1870 got paid a lot, whereas the worker in 1845 didn't. Guest: So, the work in 1870 was basically paid what they were worth. Because, if they didn't get it, somebody else would pay them--somebody else would hire them. In 1830, you just didn't have the prospect. If you left your job, you were not going to get a job at another mill, in all likelihood. I found something that only 18% of new hires had previous experience. That's a rather low figure. So there's no guarantee that I'm going to be able to take my work elsewhere. Now, why wasn't, why couldn't I find those alternative opportunities? Because basically if I went to another mill, things were so different between one mill to the next, that the new employer had no guarantee that my experience was relevant. They would more or less have to treat me as a new hire. And we actually see some of this: that even people who had some experience are not paid well at first. Then their productivity goes up over time, and then they make the earnings. Understand, most workers in these mills were paid piece rates, so their earnings were very closely tied to their productivity. Russ: So this is--a technical term for this is firm-specific human capital. So, I have some knowledge that I've learned on the job, but it doesn't apply elsewhere, so it's not valuable to new employers. I'm not much different, you are saying, from a new hire. So, what changed in the industry, then, that made me more productive outside the mill I started in? Guest: So, it was partly things became more competitive in the industry, and so employers--the mills started forming trade associations so that they exchanged all sorts of knowledge about the technology and about work practices--you see the Cotton Textile Manufacturer's Association forming in, I believe, 1865. You see the growth of a residential labor force. So, again, understand, in the early years they were hiring from all over New England. Where they situated the mills--these mills were driven by water power, so they had to be at locations that had good water power, which weren't typically in urban centers: there weren't many workers around. Lowell, Massachusetts was just a farming village. So, they hired girls off of farms from all over New England to come live in town. They were put up in boarding houses. And only very slowly did the residential labor develop. So, when somebody left the mills, they went back to Northern Vermont or whatever, weren't around to be rehired by anybody else. So, part of this development was the development of an urban labor force. And that's why it took especially long, I think, in the 19th century. And of course that's not a problem we experience so much today. Russ: Well, it sounds like China, actually. Guest: Right. Russ: Young women leaving the countryside, moving to cities, boarding-- Guest: Living in dormitories. Exactly. Russ: Not a very pleasant life at first, for sure. Guest: And you see this same pattern with Japan's industrialization. And again, beginning with the textile industries. It's striking how similar this pattern has been repeated. In Japan, they actually studied the Lowell pattern and attempted to follow it. In China, I don't think so much. But it's nevertheless you are facing some of the same problems, the same issues, and you see some of the same sort of developments.
15:44Russ: Let's move to the present. One of the areas I think people are worried about is manufacturing. And listeners to this program, I think, are pretty knowledgeable in the sense that they know that manufacturing is an incredibly successful sector of the economy in terms of output. And a very declining sector in terms of employment--as you point out as well: as a proportion of employment, it's fallen steadily since the 1940s, mid-'40s, at the end of WWII. That speed of decreases accelerated a little bit in the last 10 or 15 years, perhaps, due to, perhaps, China; perhaps the Recession of 2008. But a lot of people are worried that--the main thing driving that, of course, is productivity. It's that machines are being made that allow output per worker to be sufficiently high, that the demand for workers is relatively low. And you point out two very interesting things. The first is that that doesn't always decrease--the fact that machines make workers more productive doesn't always--or eliminates them, even--doesn't always decrease employment in an industry. So, let's start with that. Talk about tellers and ATMs (Automatic Teller Machines) as a sort of--to me a very powerful example. Guest: Well, actually, let me go back to the 19th century first, because you are talking manufacturing. So, what happened with textiles? During the 19th century, if you look at the tasks of weavers, 98% of the work was automated; yet the number of weavers continued to grow. The reason was the greater degree of automation meant that the price of cotton cloth went down, and people started using more cloth. The demand was elastic. We typically--when we think about this machinery question, we typically forget about the effect on demand. At the beginning of the 19th century, people had very few clothes. Clothing was dear-- Russ: It's hard for people to understand that. You had a couple of sets, maybe, and you washed them as much as you could bear doing. Because that wasn't fun. Guest: Right. You typically had one pair of clothes. I think that was probably typical. Most of the clothing was made at home. It was a very, very tedious process--I mean, just making the yarn took hours and hours. So, you had this tremendous drop in the cost of cloth. And people found more and more things to do with it. They got more clothes; they had the development of a fashion industry. We had cotton cloth used for draperies, carpets, rugs--all sorts of applications of textiles. And people found more and more uses. Each time the price dropped, the demand would kick in and it would increase more than the--enough to offset the labor-saving effect of the new technology. Now, eventually we get to the 1920s, 1930s, and we see that all of a sudden demand doesn't increase so much any more. The demand for textiles gets saturated. And we see this beginning of a long, but very slow decline in relative employment and then ultimately around the 1970s, 1980s, decline in absolute employment in textiles. Driven almost all by technology at that point. Russ: And of course the same thing is happening in agriculture--agricultural productivity is going through the roof. And after a while we don't need as many farmers, and it just continues to fall and fall and fall. Which is fabulous, unless you have a lifetime dream of being a farmer with your children's children's children. In which case, you'll be disappointed. But for the rest of us, who like to eat--maybe it's too good. We have an obesity problem instead of a poverty problem, for much of America. Guest: Right. Russ: But I found the tellers and ATM example fascinating. Guest: Right. Right. Russ: A lot of people assume that ATMs have destroyed the number of tellers in America. Guest: Right. And it's the same logic at work: that there's a demand effect. So, what happened when automatic tellers came in? Basically starting in the mid-1990s, ATM machines came in in big numbers. We have, now, something like 400,000-some installed in the United States. And everybody assumed--including some of the bank managers, at first--that this was going to eliminate the teller job. And it didn't. In fact, since 2000, not only have teller jobs increased, but they've been growing a bit faster than the labor force as a whole. That may eventually change. But the impact of the ATM machine was not to destroy tellers [?], actually it was to increase it. What happened? Well, the average bank branch in an urban area required about 21 tellers. That was cut because of the ATM machine to about 13 tellers. But that meant it was cheaper to operate a branch. Well, banks wanted, in part because of deregulation but just for deregulation but just for basic marketing reasons, wanted to increase the number of branch offices. And when it became cheaper to do so, demand for branch offices increased. And as a result, demand for bank tellers increased. And it increased enough to offset the labor-saving losses of jobs that would have otherwise occurred. So, again, it was one of these more dynamic things where the labor-saving technology actually created more jobs. This is in fact a much more general problem--a much more general pattern, I mean. We see a whole number of occupations where you might think that technology is going to destroy jobs because it's taking over tasks; and the reverse happens. So, if you look, for instance, when they put in scanning technology into cash registers, the number of cashiers actually increased. When legal offices started using, beginning in the late 1990s, electronic discovery software for doing discovery of documents in lawsuits, the number of paralegals increased rather than decreased. Russ: The other part of it that I found so fascinating, I think that often typically gets forgotten is that often--not always, but often--in these industries, what the people actually do isn't the same any more. It's not just that there are fewer or more tellers. It's that the tellers that are still there now are doing something a little bit more that's just different. And I think that's a really important part of this transformation. Guest: That's exactly right. And so, what's happened is that cash-handling has obviously become less important for tellers. But their ability to market and their interpersonal skills in terms of dealing with bank clients has become more important. So the transition--what the ATM machine did was effectively change the job of the bank teller into one where they are more of a marketing person. They are part of what banks call the 'customer relationship team.' But it's a different sort of skill. Maybe it's a higher skill. There is some evidence that their wages have gone up. They are hiring more college graduates as bank tellers. And in a whole variety of ways we are seeing changes of this sort where the nature of occupations is getting up-skilled in some fashion. Often very specific skills related to the particular technology, the particular job. This is happening across the board. And that's part of the challenge that technology is posing for us: How do we develop all of these new skills? Much of that was the challenge with the weavers in the 19th century.
24:29Russ: So, one of the things people worry about, we've talked about a lot on the program, is this fear that artificial intelligence or smart machines won't just make it cheaper to expand output and change the numbers in complicated ways through demand response, but just eliminate them totally. So, we'll go to online banking, where there will be no people. We'll go to a smart factory that just--the only thing the person does is make sure that the dog that's in front of the place doesn't get out of control--I think that's a joke. I'm not doing it correctly. But, you know--that's the concern. And just to take one dramatic example: A world of driverless cars and driverless trucks will eliminate thousands of jobs currently held by cab drivers who are drivers, truck drivers; and it's not like there will still be a few left to make sure that certain cars are driven well or to oversee the steering wheel process within the car; they'll just be more chatty and friendly or something. They just want to have a job. And their skills, those specific skills will now be worthless, that they've accumulated driving a truck. And so their next best alternative is going to be a lot worse-paying. Does that concern you? Guest: I think it does. The change we are facing now is different in a couple of dramatic ways. So, one of them is that we may be seeing technologies that are capable of replacing the entire set of tasks of an occupation. I don't think we are there yet. I think experts in artificial intelligence will tell you it's going to be quite a while till we're there yet. Except for maybe some very specific occupations. The other sort of reason things are very different now is that computer technology, information technology, is affecting far more workers. It's affecting a much larger share of the workforce than anything else we've seen before in terms of major technology changes. So, we can think of mechanization in the 19th century or electrification during the early 20th century as big technology changes. But these only affected a relatively small part of the workforce. With computers, we are talking about the majority of workers, already. I guess one of the--so, there is concern that we will see complete replacement of workers in some areas. There are a small number of lights-out factories. But that's not a very significant aspect today. I think, however, that the main thing that's happening and the main turbulence to jobs is coming when computer technology is being used by some workers to replace other workers. So, we can definitely point to occupations that have been diminished. Typesetters, for instance, have--the number of typesetters has dropped by 80% or something since 1980, with the advent of computerized publishing. On the other hand, the same technology is associated with an increase in graphic designers--a much larger increase in the number of graphic designers. So work has shifted from one occupation to another. And that's a typical thing we are seeing today. And I think it's a typical thing we are going to be seeing over the horizon in the next 10 or 15 years. It's not so much the specter of machines completely replacing workers. It's much more about jobs changing either internally, as with the bank tellers, where the nature of the job changes; or work shifting from one occupation to another. And all of this requiring new skills. The specter of wholesale job replacement and major impacts on unemployment, if it occurs--it seems to me to be multiple decades out. Russ: The graphic design is a nice example. Because if you are a typesetter, you are probably not so--let's say it differently: Being good at typesetting doesn't help you with graphic design. As you point out. It's just a totally different set of skills. Now, you can go acquire those skills, perhaps. It would depend on the person, their age, and their creativity and their brain and all that. But, I think that's one of the worries: that the transition time, as people leave an industry and join new industries, is maybe going to be a lot more difficult. I think one of the things--we had David Autor on the program talking about trade with China. And, if you lose your manufacturing job or your construction job--say, the construction sector was overheated due to a variety of factors--it's not obvious what your next best alternative is once you confront the fact that that job you lost may not be coming back. There's a human tendency to weight it will come back. And if it doesn't, you eventually have to confront the fact that you need a different set of skills or you are not going to be paid very much. I think one of the challenges is--your book, Learning by Doing, some of those skills are--the mechanisms by which a person can acquire those out of the blue is not so easy, when they are in the middle of their lives. So, do have any thoughts on that? Guest: Yeah. So, I think this is the big challenge. And it's a big challenge for two reasons. One is the thing I just mentioned, that we are talking about technology affecting more jobs now than ever. So, I think the scale of the transitions that people are going to have to go through is much larger than we've seen before--at least if it's resulting from technology. The other is the problem that these skills are often very difficult to acquire, for a number of reasons. So, one is that until things are standardized and well understood, they can't be taught in school. Or, it's difficult to be taught in school. As long as they are, continue to change very rapidly, it's difficult to be taught in school. And it's difficult to teach yourself. So we're seeing in numbers of areas great challenges in acquiring these skills. It's not something that we can expect our schools to do very easily. And we don't have much in the way of other institutions to help people with that. The graphic designer is an example where the skills are continually changing. And it poses a big challenge. So, what--initially you had this publishing coming in, in the 1980s. And so designers, who had been primarily trained as print designers now had to learn a new technology. And then the Internet came along; and so they had to learn web design. And then smartphones came along, and they had to learn mobile design. And throughout this process, standards and technologies keep changing. So, a few years ago, Flash was a requirement, more or less, for a lot of jobs. Today, Flash is seen as obsolete; and people are learning HTML 5. If you are not able to teach yourself the latest technology, such as Flash or HTML 5, either on your own or by taking an online course, or whatever, it's very difficult to--you can't get the very best jobs. And so what you are seeing is a divergence in pay between the average graphic designer, who still largely has print-design skills, and the top designers, who have these other much more sophisticated skills and is able to keep up with the pace of change. So, you project that sort of challenge across many, many occupations, which is what, I think, we are seeing, and we've got a social problem. A difficult problem.
33:03Russ: What strikes me about graphic design, as an interesting example, is that it gets standardized. But then the standard changes. Guest: Yeh. Russ: So, at first, you have to know Flash. Or you have to know--I'm thinking of just the artistic part of it--you have to know Adobe Illustrator; you have to know Photoshop. And so often software gets more and more complicated. They get more and more bells and whistles. It gets harder and harder to be a master of it. And all of a sudden, that software is out. People don't want to use that any more. There's a new program or a new package that's designed to make your life easier. But you don't know it. You haven't learned that piece of software. So you've got to now, in a sense, retool. And it's almost a cliche. But clearly the best skill to have is not Adobe Illustrator or Photoshop or Lightroom or whatever is the design tool you are using. It's to know how to learn to use a piece of software. And that's something you can't teach, except through, perhaps, teaching lots of different kinds and how people see they are connected. And I suppose it's easier to learn a new software package once you've learned a few. Guest: Yeah. I think that's right. It's this, what people call lifelong learning. And we're seeing a lot of people putting effort into different ways to do this. So, the rise of these online courses. We're seeing boot camps. We're seeing traditional schools trying to adapt their curricula. It's hard to know what works and what doesn't. Me, as a person needing new skills, it's often hard to know where to go. Russ: Yeah. And I think that's true at every level of education. I encourage my children to stay in touch with new things that are being developed and try to get them to think about going into them, if they are interested in them. They are in the middle of their teens or slightly past their teens. So it's a very exciting time. But, not quite sure what to work on, or what to invest in. But what seems to be true, for my kids, which is a blessing, is that there are a lot of things they could learn on their own, if they want to. They can get on Coursera. They can get on the web and find--they can almost get themselves a college major without the piece of paper. But they can have the skillset. Now, obviously this is a question of being able to prove that knowledge. Let's put that to the side. Yet, at the same time, there are many, many kids who are not prepared to do that level of self-learning. And I just want to come back to this point. It would seem to me that if you can teach people how to teach themselves, you really give them the greatest gift. It's almost the modern variation of 'Give a person a fish, they are fed for a day. Teach them how to fish, they are fed for a lifetime.' If you can teach somebody how to learn, you've really given them the greatest gift. Guest: Yeah. I think that's right. And, you know, one of the problems is I don't think we know how to do that, large scale.
36:19Russ: Well, let's shift gears and talk about this training issue a little more explicitly. There's a big--along the lines I'm talking about--it's a parallel track. People think it's the same track, but I'm not sure they cross. But, [?] to say, 'Well, we've just seen more college graduates.' And you are very critical of that claim. Why? Guest: So, college graduates might--college diplomas might very well be good things for most people. But it might not be what they need right now. This kind of skill--talk about the graphic designers: You have a lot of graphic designers who have college diplomas. You have a lot who don't. The college courses in graphic design, are, I have to say, most of them at this point are not able to teach the latest graphic design skills. And I'm not sure that they teach--and maybe the best ones do--but I'm not sure that they teach you how to learn. At least, it appears that a lot of graphic designers who do have college diplomas aren't necessarily faring so well. So, it's a story that a lot of these skills have to be learned in part through experience with new technology. Some of this stuff has to be learned on the job. Or, at least, in conjunction with on the job experience and, you know, a classroom. So, you know, one of the things that's encouraging is you see a lot of community colleges developing work-study programs with local employers, where people are able to learn both in the classroom and on the job. And I think that's an important aspect of solving this problem. We're seeing some other things developing that I think are helpful. So, because actual experience is so important, with many of the new technologies, we're starting to see more trade associations developing certifications. So that, somebody gets certified as having a set of skills, even if they've learned it on the job; but not in the classroom. And this becomes an important signal for employers, that this person is somebody who they can hire who will be able to get the job done. And that helps build the labor market and overcome these problems of firm-specific human capital that we talked about earlier. Which was a problem for the weavers and is a problem for many people today as well. Russ: The other part of this that you talk about so well in the book is the role of licensing. Once you have certification there's a temptation that a person, to work in this field, needs a certificate to get the license that makes them qualify for the job supposedly. Talk about the growth in licensing; why that's worrisome rather than encouraging. Guest: Yeah. So, it seems like certification and licensing are more or less the same thing. And they are not. There's really a very important difference between them. So, to the extent that certification presents and employer with some information about your characteristics, that's a good thing. To the extent that you can't get a job without a license, that's not such a good thing. Because it prevents you from getting the experience that you need in the first place. What we've seen is a tremendous growth in licensing, from about 5% of all workers in the 1950s to close to 30% today require some sort of licensing in order to get their jobs. Now, licensing may be very important where there are critical issues of, you know, health or safety. But, what seems to happen repeatedly--and there's very good evidence about this--is that licensing puts professionals within the occupation in a position to restrict entry into the occupation, raise prices, reduce service quality; and it may help the people in the occupation but it limits entry of new workers; it limits the adoption of new technologies; and it certainly hurts consumers. So we've seen a huge growth in licensing. And to my mind, and I think to the minds of many other people, excessive growth, where there needs to be some much better economic rationale for some of these licensing restrictions. Russ: As you point out, the political rationale is very clear. It protects the existing workers and gives them a little more security and lets them get paid more, but it's bad for the rest of us, including the people who would like to be in that industry. What are the chances that that political rationale is going to change any time soon? Other than people like you and me, and others which I've seen lately, who are saying, 'Hey, wait a minute. This is crazy.' Guest: Yeah. I don't think it's going to change very soon. It's rather discouraging how much political activity has been oriented around these things. A lot of this occurs at the state level, so we're talking--it may not even make it into the newspapers. People are generally unaware of what's happened, of what's happening. So these are things that tend to affect, you know, relatively small communities in the legislature but their impact on the broader society is much greater. And, I can't say I see any short term change in that pattern. Russ: A small glimmer of light would be the Institute for Justice, which if you don't know about it out there, check it out. They try to fight unnecessary licensing through the court system as a way of promoting economic freedom, and I'm a big fan of a lot of what they do. They are fighting this battle right now, city by city, state by state.
43:13Russ: So, let's talk about an example I knew nothing about--it's an interesting claim--which is the role of government procurement in encouraging innovation and technology. The last third or so of your book is about government policy; and a lot of people want government to champion certain kinds of innovations. As you point out, it's a very mixed bag; it hasn't been so successful. But you point out a role for procurement--government purchases of stuff--that I'd never heard or seen before. And it starts very long ago. So, start with that example of rifles and firearms, which I'd never heard. It was very interesting. Guest: Yeah. So, I think it was even before the--well, shortly after the Revolutionary War, Thomas Jefferson among others became aware that they needed a new way to build firearms that could be standardized so that replacement--each of the parts could be cannibalized from one rifle and put into another one, so that they could be repaired on the field. So, this was the ideal, the manufacturing ideal. Government played a key role in fostering this technology. Now, it did it for these very specific reasons of making firearms this new country could rely on, especially as it didn't have a large number of gunsmiths. So, they developed these techniques. And what happened was--it was done in a very open way, so that there were two armories--Harpers Ferry, West Virginia, and Springfield, Massachusetts--that were centers[?] of technology development. They shared the information widely. They hired numbers of private firms to produce the guns, and these firms would share their knowledge and their latest developments with other people in the process. And as a result what grew up was a whole community of people who were skilled in this new approach to machining. Well, it turned out that that community was able to then take those skills and apply them to a whole range of other mechanical technologies, and this really led to the emergence of the United States as the leading nation in terms of mechanical technology. It led to the development of the assembly line, but much more generally than that, it led to this real supremacy in terms of having workers who acquired these very unique skills to use technology with replaceable parts, which became known as the American system of manufactures. By providing the procurement program under the right set of rules where knowledge was widely shared, the government had helped spur this entire process. So, you hear people talking--and I think correctly--about government playing a key role in modern technology, such as the semiconductor or the computer, the Internet. And in many of these cases, it was government procurement playing a very similar role to what was played, to the role it played in the 19th century. And much of it related to either NASA (National Aeronautics and Space Administration) or defense procurement. But it was done--the reason these efforts ended up in having a much wider and more important civilian application was that they were done in ways where the key knowledge was widely shared; skills were developed among a broad base of people; and this provided the springboard to greater civilian application. And a key part in recent productivity growth. Recent technological growth. Unfortunately, it's not always the case that government procurement is done in such a beneficial way. Russ: Some of that's luck, right? It wasn't intended. The government, when it made those firearms, demands and purchases, wasn't thinking, 'Well, this will lead to all that great stuff.' Guest: Exactly. Russ: I think a lot of the Internet is the same way. People say the government created the Internet. Well, they created some infrastructure that allowed other people to create the Internet. So, obviously there is some public role that is relevant, but it's not necessarily their best thing, the government's best thing, to start from some idea and try to take it through fruition. But there are these unintended positive consequences sometimes.
48:41Russ: But I want to bring you back to standards. Because I don't think we got at the issue that I think is really at the heart of the question we started with, which is: When do workers share in the gains and when do they not share so much. Which is partly a role of standardization. So, standards can emerge from private activity, through either associations of manufacturers or industries. Sometimes they can be imposed from the top down--they can decree a standard of some kind, sometimes correctly, sometimes not so correctly and you can go down the wrong path. Obviously we can make mistakes in all kinds of settings. But I want you to try to bring back why standardization is important for worker productivity and compensation and their ability to learn on the job, because I think we didn't give that enough attention in the first part of the conversation. So, why is that relevant in terms of allowing more learning by doing, more specific human capital that allows workers to be compensated and productive? Guest: Right. So, this is the--this was something with the example of the rifle, with the Internet, with all these things: is that a key aspect was open standards that were widely shared. So, why was this important? And people who write about standards, typically write strictly from the point of view of the importance of standards in allowing inter-cooperation, inter-operation of different parts. My argument is that standards play a perhaps more important role, or at least as important role, in terms of their ability to make skills portable. To allow--and having portable skills allows robust labor markets to develop. So, what happened when we had standards for making machinery with interchangeable parts were that there were a whole set of skills, and not only standard tools, but standards of measurement that a worker could learn. And if they had learned these standards, their skills were then widely applicable to other employers. And so that meant that they had a wider labor market to appeal to, when they were looking for work. Russ: So, I didn't really appreciate this until I read your book. And I think I appreciate it even more from this conversation--which is, thinking back to the graphic design idea. So, if you are in a company that wants to have, say, a corporate newsletter, or some kind of ad that you want to produce for some kind of publication; and you've got a brilliant genius in your IT (Information Technology) Department, and she figures out a way to help the art people draw that ad or create that newsletter. And then that firm dies or you need to move or you get fired or whatever it is, and you go want to go to a new company; and they have their own genius who has created a totally different system. It's a fabulous system for creating newsletters or for drawing or for an ad. But you've never used it before. So you've got to start from scratch. And a software package--one of the things that--we tend to think about it as, 'Well, it's great.' Because then you can do the stuff easily without being a great artist. But the other thing it does, which I hadn't thought about till I read your book or had the conversation is that what it means is that somebody can teach you how to use that package somewhere. You can take a course in it. You can take--you can go to night school. You can learn it online. You can buy the package yourself and practice it. And once you've learned it--most firms are using that package--you are very valuable to all of them. And so your wages are going to be a lot higher than they would be. Unless you are that rare person who can develop the package in-house for that customized one-time application. And so I thought the part I really enjoyed, hadn't appreciated, was that, from your book, was this idea that the ability to acquire a skill, once standardization sets in, gets a lot easier. You can teach it in a class. You can write a book about how to learn it. Without that, you've got nothin'. And so that allows the average worker to invest in a skill that has an application beyond the firm that they work in. And that's huge. Guest: Yeah. And very aptly put[?]. I've been struggling to say it as nicely. And there's something else, though, too. We were just talking about how do these governmental technologies--government plays a role in the early stages, and it fosters--it flowers from there. And when you have something where you can take that skill and use it and multiply it across employers, that's how new technologies take hold. That's how they get widely adopted, and then develop the productivity gains that we all want so much from technology. So, technology struggles to get adopted when it's very firm-specific and when it's not standardized. And it's inevitable that in the early stages because it's changing rapidly and it's new, when there are multiple variants, things are different from firm to firm and things don't get widely adopted. And so we have new technology, but we don't have necessarily the productivity growth we would hope from that new technology. It's when things become standardized that the technology itself becomes more productivity. Because we have the skills, the skilled people, who can take it and spread it widely.
54:45Russ: The other thing I just want to make sure we mention is the synergy between people across firms that can occur in geographic areas. And in particular, famously in the United States, Silicon Valley. And you contrast Silicon Valley with Route 128, which is a fantastic example I'd never heard before. I just don't want to miss that conversation point. It's just fantastic. At least the claim. I don't know if it's true. But talk about the non-compete role, and the difference in the two places. Because it's an issue that's in the news lately, the role of non-competes--being the contractual thing that when you are hired somewhere that if you change jobs or quit, you can't necessarily work wherever you want. Which is a bizarre thing that you would agree to such a thing. But a lot of firms have non-compete agreements. So, talk about that and how it makes a difference. Guest: So, the use of non-compete agreements, particularly for technical employees, has been growing over the last couple of decades, as well as--it's not just licensing. Non-compete is an issue. Non-compete is starting to be an issue in lower-wage jobs as well. But I think it's particularly important for technical employees. So, from the point of view of the employer, there's a rationale for a non-compete, which is: If I'm going to, as an employer, train you on some specific skill, I want you to stay employed at my firm rather than take that skill to somebody else. So, having a non-compete might make the employer's incentive to provide training greater. Now, as it turns out, employers, don't, in the United States, provide much training to employees. But there is a flip side to that coin, which is: If I as an employee have a non-compete agreement, I have less incentive to invest my own time in learning a new skill. Russ: Yeah. Guest: And so that's one of the drawbacks of non-compete. There are two others I think. One is that in a place like Silicon Valley you see a lot of knowledge being transferred from one firm to another as employees go from one firm to another. And this is often critical with newly emerging technologies. So, nobody typically has a monopoly on all of the good ideas. So, there's Firm A and there's Firm B; and Firm A has figured some things out, and Firm B has figured some things out. And if, through employee exchange, which is what happens typically in Silicon Valley, both firms can benefit from what they have jointly developed. The third thing is non-competes make it difficult for new firms to start up. So, we are seeing cases in Massachusetts where a firm wanted to--a Silicon Valley firm, actually--wanted to, for instance, wanted to open up a Massachusetts office. But everybody they wanted to hire had been employed some time in the last two years in the industry for a competitor, and so they could not hire anybody. And they eventually gave up. From the employees' point of view, these things are sometimes very unfair. So, there are stories where somebody has been working, got a Ph.D. in the technology, who has been working in the field, that they take a job; things don't work out, they've got a clash with the boss; and they leave. And then for 2 years their entire--everything they've learned over a long period of time is put on iceballs, and they've got to take a detour. So, economists have done some very good work showing that there are some detrimental effects on startups, on innovation, patenting rates, with non-compete agreements. And because we are seeing some state-to-state variation, they are able to measure this quite nicely. So, there is good empirical support for this explanation, of the difference between Silicon Valley and Route 128. Russ: Well, so the claim is that Massachusetts is much more tolerant[?] of non-competes than California? But my impression now is that California seems to have a growing issue with this. Guest: No. Yes. In California, non-competes have not been enforced. In Massachusetts, they have been. So, Massachusetts, there are some efforts, including by the Governor, to change the law in Massachusetts. But this has worked in favor of Silicon Valley as opposed to Route 128. So, if you go back--this is, Annalee Saxenian first made this argument a while back--if you go back to 1980, both areas seemed very comparable in terms of tech firms. They had tech firms with large universities. They looked very similar. But California was able to adapt to challenges much more effectively and Silicon Valley grew much more rapidly, where the firms around Boston did not. Now, there's something of a resurgence in New England, particularly with biotech. But there seems to be some good evidence that noncompetes played a role in the poor performance of Massachusetts.
1:00:46Russ: So, we're almost out of time. We had an episode with Adam Davidson a while back on manufacturing, and of the lessons from that was that the manufacturing jobs that are left are actually quite high-skilled. Which is why many of them do pay very well. They require some knowledge of advanced mathematics. And that's one way you can imagine that an ever-shrinking industry might still produce valuable employment for some people. Similarly, there are going to be jobs for people who create the smart machines. So, one of the--I think the temptation is to say, 'Well, we just have to teach everybody how to code and teach some mathematics.' And I think that's the wrong lesson to learn, because I don't think everybody can code well, and mathematics is hard. So, what are your thoughts for what our school system should be doing, either K-12 (Kindergarten through 12th grade) or universities or community colleges? Tell people who are not going to be coders and mathematicians to be productive going forward in a world with a lot more technology in it. Guest: So, this gets back to the kind of transitions we were talking about earlier. You look at the bank tellers or the graphic designers--they don't need to be coders. Which, some graphic designers are html coders, but I think for the most part, not. But, on the other hand they need to be comfortable with computers; they need to be able to think quantitatively. And--I think this the most important thing--they need to be able to learn new systems as the systems are continually changing. So, if our schools can do a better job of that--and I'm not sure how to tell them how to do a better job--I think that's the challenge for our schools, is to figure out how to produce people who can be lifelong learners. Russ: Yeah. I'm not sure we know how to do that. But certainly the other part, which you allude to I think briefly in the book is that there are some nontechnical skills that are going to be valuable in a world of computers and smart machines, and those are human skills. Empathy and so on. Guest: Right, yeah, yeah. Russ: So, I don't know how to teach that, either. But that's another route I think that will still allow people to be compensated well. Guest: Right. Increasingly the jobs are going to be jobs that involve human aspects--interpersonal skills, abilities to listen, to develop trust, those sorts of things. And at the same time to be able to relate to computer systems. That's the combination. You can even think of the bank teller: works with a computer, needs to understand the computer systems--not know how to code them necessarily but needs to be able to work with them. And yet at the same time be able to tell the small businessman who just walked in about the latest loan offerings that the bank has and might be useful.