Intro. [Recording date: July 30, 2019.]
Russ Roberts: My guest is Daron Acemoglu of MIT [Massachusetts Institute of Technology].... This is his fifth appearance on EconTalk. He was last here in November of 2014 talking about inequality, institutions, and Piketty. Today we're talking about a different monograph he did for a group called Economics for Inclusive Prosperity (EconFip). His piece is entitled, "It's good jobs, stupid."... Now, I want to start with a more basic idea--we'll get to good jobs in a minute, but I want to talk about what the idea of shared prosperity means to you, which is a phrase you use in a number of places in the essay.
Daron Acemoglu: Yeah; I think it has two components, really. I don't think that there is a level of inequality that you can say it's broadly agreed that you shouldn't go above or anything like that. But if you have a period of, you know, almost 4 decades in which some groups are left behind while, you know, GDP [Gross Domestic Product] per capita and other measures of prosperity are forging ahead, that is clearly not shared prosperity. And then the second element that will, you know, perhaps take us to the heart of the discussion: I think you really need some amount of wage growth and employment growth to go with prosperity, because that's the only way of sort of creating the basis of shared prosperity.
Russ Roberts: What do you mean by that? What do you mean by wage growth? Don't, by definition, you need growth? Or are you just saying that we could imagine a world where everyone is relatively comfortable?
Daron Acemoglu: In the United States, you know, we've had growth for the last 3 decades, but average wages have been more or less stagnant; and employment has grown not much more than population growth, or actually slightly less. So, the ability of the economy to sort of generate more jobs, and especially higher wages, is not guaranteed. And the reason why I am actually emphasizing that is because, you know, one view is that perhaps we can create shared prosperity by redistributing. And that's not what I think is generally long-term sustainable--that there are a number of political, social, and economic reasons why we have to push the market economy to generate shared prosperity. Not, say, let the market economy rip and then we can use taxes and transfers in order to achieve whatever level of distribution of consumption and resources we wish.
Russ Roberts: Yeah; that latter vision is one that gets invoked, particularly when people are worried about extreme increases in, say, the ability of artificial intelligence to automate work--
Daron Acemoglu: Exactly. And that's why those topics are really central to the conversation we'll have.
Russ Roberts: Yeah--I think if we really have--I don't think that's going to be a reality, but if we really had a world where only a handful of people were able to work, creating the machines that made us immensely wealthy, that somehow, that we would then extract resources from that handful of people to share with the rest of society. It strikes me as a pretty dystopian future.
Daron Acemoglu: Absolutely. It is dystopian; and for more than one reason. First of all, I think people would be really unhappy, alienated--meaning missing the meaning of life--because I think staying at home and living off government transfers is not going to be satisfactory for the vast majority of people. Not for everybody: for some people it might be fine. But for the majority of the people it won't be. But second, it won't be feasible to maintain that, because at some point those who command all the resources of the economy are going to find ways of getting out making those transfers. And, finally, I think the democratic, political equilibrium that undergirds what we have right now wouldn't be maintained, wouldn't be sustainable, if we had an economy in which a handful of people generate and capture all the value and the rest just live off transfers.
Russ Roberts: Yeah; I think we'd be in uncharted territory. Just on the political side, I think the idea that somehow we'll peacefully rearrange things so that we're all doing fine seems to me to be a bit of a fantasy. It's a fantasy--
Daron Acemoglu: Absolutely. Absolutely--
Russ Roberts: Whether it's a realizable fantasy is--
Daron Acemoglu: No, I don't think it is realizable. I mean, it's extreme; but it is useful to contemplate it. Because, if you look into the details of many of the ideas that come from the Progressive Wing of U.S. politics and some other countries' politics, they really amount to a world in which something like that is going to be an important pillar of it. So, in some sense what I'm--the essay that you mention is really trying to articulate an alternative Progressive vision.
Russ Roberts: Let's talk for a minute about this issue of meaning and a satisfying life. Sometimes we call it, people call it 'flourishing'--a sense of mattering. And, certainly through human history that we know of, work is central to that--
Daron Acemoglu: Absolutely--
Russ Roberts: There's a lot of horrifying work that people did just to stay alive. But, through, I would say, the last 200 years or so, to a large extent, people, a lot of people derived a sense of meaning from either their work itself or the opportunity to take care of their family, to support their family. Is this just a cultural phenomenon that would change in a world of extreme automation? Is it possible that meaning is over-rated and that people would enjoy--well, you know, my dystopian vision is people watching YouTube all day. Which I feel like many of us already do to some extent. But, what do you think about this question of meaning and work?
Daron Acemoglu: I think--it's out of sample, so we have to extrapolate; so it's hard to know. But, I think most people who do not use their jobs, their main profession as an important part of their identity find it very hard to, sort of, sustain a meaningful, motivated existence. And in some sense there is a good reason for that. You know, evolution really--we are programmed to work. Otherwise, it would have been just much harder for people to be motivated to go out there, hunt and gather or go and do, engage in productive activities. It's not a coincidence that everybody, starting from childhood, really takes pride in doing something well and identifying with something. You know, if you watch children, you see that it's completely ingrained in them. Now, again, you can say, 'Well, perhaps instead of a job that could be a video game.' But, I think there is a sense in which that's not going to work as well. And, in history, you know, sociological accounts find how people identify with very, very unpleasant and bad jobs because it gives them a community; it gives them solidarity; and it gives them a meaning of life. There is a work by, you know, sociologist Runciman [W. G. Runciman], for example, you know, describing, you know, coal miners, which sort of gives you a very good sense of this.
Russ Roberts: Yeah. I worry that my experience--which, my work is deeply meaningful to me--I have other things that are deeply meaningful, too--but, so it's not just my work. But my work is deeply meaningful to me. And I worry that--I suspect you like your job, too--
Daron Acemoglu: Absolutely--
Russ Roberts: I wonder if, you know, we are wildly unrepresentative, or--
Daron Acemoglu: Yeah. We might be. I mean, that's the problem. You know. We are--it's very difficult to sort of leave your own identity when commenting on these things. I often warn my students when they work on labor issues: Don't think that the world over the labor market looks like the U.S. academic labor market. And, you know, if you look at the, sort of the totality of human history, a lot of work was entangled with coercion. That's obviously not everything that people did. They liked and enjoyed, and that's why their employers and the elites chose to coerce them, sometimes viciously, in order to get them to do the job. But, but, finding meaning in jobs and what you do is extremely common in history.
Russ Roberts: Yeah, and I don't think--an interesting question; I'm sure sociologists have looked at it. I don't know how many economists, if any, that have looked at retired people. And you hear people bemoaning--I mean, I can't imagine retiring. Literally.
Daron Acemoglu: Exactly. Yeah.
Russ Roberts: Maybe I'd do more [?] and maybe I'd play golf more than once a decade if I were retired, maybe I'd take a lot of photographs--I love photography. But--
Daron Acemoglu: Lots of movies I want to watch.
Russ Roberts: Yeah. I just don't see that work[working?]--and, you hear that from people. You know, 'The first three months are fun, and'--
Daron Acemoglu: I think the evidence[?] that I have seen is that, you know, there's some fraction of people, I don't know exactly what, but perhaps 15%, 10-15%, who really enjoy retirement: They travel, they find some completely new hobbies and new ways of doing things. But the overwhelming majority, they find it depressing. And go through various crises because of retirement.
Russ Roberts: So, I want to challenge one of the presumptions that underlies part of your empirical story. And it's a very common presumption. I'm out of the mainstream. But, this view that, you know, wages have been stagnant for 40 years, say. And, I was alive in 1979. I was 25 years old; I remember it pretty well. It was not a particularly prosperous time for the average person, compared to today. And this raises the possibility that the data that we have to try to measure prosperity are not very accurate--that, we don't count compensation; we often just count money wages. We don't measure inflation correctly. Do you really believe that the average person--
Daron Acemoglu: Well, yeah, so--
Russ Roberts: has been left behind for 40 years?
Daron Acemoglu: Absolutely. Absolutely. Yes, I do. But let me take it on. And, I don't think the skepticism that you have expressed is as uncommon as you make it out to be. I think, uh--
Russ Roberts: Yeah. Maybe.
Daron Acemoglu: I think, you know, uh--there are a couple of facts that challenge people. One of them--and that's why many people end up where you are, at some level, is that productivity growth has been very disappointing, and if you put this to, you know, the titans of industry, especially the tech industry, or people involved in the tech industry, or tech enthusiasts, then they are cornered into saying, 'Well, we must be mis-measuring it. So the quality of products we have is much better.' Which, amounts to saying CPI [Consumer Price Index] is exaggerated and therefore, you know, real wage growth is much more like nominal wage growth, or in fact even perhaps greater than nominal wage growth. Now, that is a possibility. But, you know, careful studies, especially by the BLS [Bureau of Labor Statistics], have not found that the mismeasurement of quality is greater over the last few decades than their estimates of mismeasurement of quality [?--Econlib Ed.]. Because, you know, don't forget: there were many, many new products, and quite transformative new products, in the post-War era that people did not have access to, that really spread in the population. And there's a lot of sort of quality mismeasurements, and also, the other thing to bear in mind is that, you know, if there is quality mismeasurement, it would have some implications. For example, it would show that in places where you expect more mismeasured quality, you should have more nominal spending growth, for example. Because they are--especially if the elasticity of demand is greater than 1. And, generally, we don't find those things. For instance, one paper that I wrote with David Autor, David Dorn, Gordon Hanson, and Brandon Pierce [?Brendan Price?--Econlib Ed.] has looked at the manufacturing industry. And, in places where you have more computers--so, you might think, those are the sort of technologies that are contributing to more mismeasurement because there are increasing quality of products or introducing new types of products. Actually, both nominal and real spending track each other very strongly, so there is no decoupling that would indicate some sort of price mismeasurement. Now, not all this is dispositive. Not all that proves that CPI is not mismeasured. But, I think there isn't, uh, great evidence that it is. But, for great reason it is useful to look at things that are not affected by CPI mismeasurement. And, uh, one that is not affected by CPI mismeasurement is if we look at the share of national income or GDP that goes to wage earners versus capital owners. And there, this is completely immune to, uh, sort of overall CPI mismeasurement. And the picture is very striking. So, if you look at the, uh, the sort of the share of national income going to labor, it starts dropping very sharply, especially in some sectors such as manufacturing, mining. Uh, and it doesn't recover at all. And it sort of creates a very striking pattern of some groups being left behind, whatever is going on with GDP or productivity. And let me add one other thing. One additional argument that is suggested in regards to some of these facts is about fringe benefits, health care--
Russ Roberts: yeah--
Daron Acemoglu: and other benefits that are not included in some of the standard weight measures. But, there is a very good study of that. It's not updated, but it sort of covers the early 2000s, by Brooks/Pierce, and we have a lot of the patterns that happened in terms of inequality, increasing[?] median wages being stagnant already in the 1990s and early 2000s. And what he finds that, using the confidential census data and sort of Bureau of Labor Statistics data, what he finds that if you add those fringe benefits, inequality is even higher. So, a lot of those go disproportionately to high earners. So, the sort of 50%, 60% of the population being left behind, that's not going to change much by adding to fringe benefits.
Russ Roberts: Well, I don't agree with that. But, that's probably a longer story. The reason I don't agree with it is, I just mention in passing is that--I don't care about inequality. I do. But for this conversation, I care about the bottom half, say, is doing. It just--to suggest that--and it may be that including fringe benefits boosts the upper more than it boosts the bottom--I find it hard to believe it doesn't boost--well, I know it boosts all groups; the question is how much? How it boosts, say, measured progress relative to the past would be interesting. But I think--
Daron Acemoglu: I mean, I don't disagree with that, by the way. Let me say that, I don't think it's healthy to just focus on inequality. But I think I emphasize inequality because it gives us a very good sense of shared prosperity. So in fact if, you know, the U.S. economy, say, generated 2% a year improvements in the welfare of the bottom 10% and the top 10% increased at 5%, I wouldn't find it as troublesome. The problem is, again, all of these measures really show that the bottom 10%, and even bottom 40%, 30%, and even the median, is not doing anything like that.
Russ Roberts: Yes. I think there are numbers that suggest that the bottom has done well; and the median especially if you look at following people over time and you look at the panel data rather than cross-sectional data. Cross-sectional data, of course, is often confused by the increases, say, in immigration, changes in the age structure of the workforce. People play all kinds of games. And I'm not suggesting that all the people play games. Some do. Many don't.
Daron Acemoglu: Yeh.
Russ Roberts: But, let me ask it a different way. And, you know, of course there are explanations for the labor-share point that have--look at how government has changed, how it treats, say, self-employment. Which, of course, self-employed workers are sometimes called labor, sometimes called capital, if you are running your own business--that makes that challenging. But, the bigger point I have, and I'm going to tell this in a different way. I was telling about this story when we were discussing this narrative with a prominent journalist, who I will not name. So, I don't think these numbers are representative of what's actually happening. And he said, 'I think they are.' And he said, 'I think most people would agree with me.' And I said, 'Yeah, you're probably right.' And I said, 'The problem I have, on top of all this, even if you accept the measurement issue, is: What's the underlying cause?' And he actually said to me--this is 10 years ago--he said, 'We haven't figured that out yet.' Meaning: Well, we know there has to be one, because the numbers don't lie. My view is I think the numbers are challenging. And complicated. But let's say we accept your view. Let's say that starting--and most people would say in the 1970s--that prosperity was no longer being shared. You'd want to see a massive, structural change in the U.S. economy, that, to me, that would explain that. And, when you press people--and it's funny--when I pressed people on this, 15 years ago, 10 years ago they were more prone to say things like: 'Well, we haven't found that out yet.' Now they claim to know, and they same things like: Market power on the part of firms. You know--and you allude to a number of these arguments in passing.
Daron Acemoglu: Yeah.
Russ Roberts: So, try to make the case for me, to me, of what's going on--
Daron Acemoglu: Yeah. I mean, I think I would say--
Russ Roberts: Daron, I want to make it clear, one of the reasons it's important and not just a fun thing to argue about--one of the reasons it's important is to understand why good jobs aren't being created. And, of course, you talk about that in the essay--
Daron Acemoglu: Absolutely. Absolutely. Look, I think, unfortunately for social scientists, this particular thorny issue is multicausal. I like unicausal stories because it's clear and it focuses the mind.
Russ Roberts: Well said.
Daron Acemoglu: But in this case I think that there are three sets of powerful explanations. And I have worked, myself, and I trust the work of others on two of them. But that's not reason for me to be able to dismiss the third. So, the first one is globalization. So, I think trade with low-wage countries has helped firms but at the same time destroyed or prevented the creation of jobs that would have been created during a different era. Second, which is very related, is automation. Which is that: we have been changing the production technology in a particular way that emphasizes the substitution of machines for labor, which, again, improves productivity but does not help create jobs--in fact, that replacement is at its root a way of reducing the labor share, reducing payments to labor, both as an incentive and an outcome. It is very much so. And I'm more than happy to get into the details of both the theoretical, conceptual ideas here, and the empirical evidence. And I think both of those are important, because, in my opinion, economists have underestimated the extent to which globalization and automation really change the structure and the nature of labor demand, so it's important to have a conceptual discussion.
Russ Roberts: Yeah, I agree.
Daron Acemoglu: And then the third set of reasons are institutional. And here, I include market power, both in the labor market and the product market. And I also include the changing priorities and norms in the corporate world that have made it much easier, sometimes under the guise of shareholders' value, sometimes under the guise of other sort of ideas, to take away from workers and benefit a small group of shareholders and managers. And, unfortunately, I cannot give you as clear evidence on some of these. And, I share your view that some of the evidence bandied around on some of these things is not completely convincing. But there are bits and pieces of evidence that is indicative even if it's not comprehensive.
Russ Roberts: Let's talk about globalization just for a minute, and automation. To me, they are--to me, the same thing. They are just ways of producing things more cheaply, sometimes through technology and sometimes through finding a way to produce it overseas rather than producing it domestically.
Daron Acemoglu: Absolutely. And, in fact, that's actually a very important observation because they both amount to replacing tasks that were previously performed by labor, by either foreign labor or machine.
Russ Roberts: Correct.
Daron Acemoglu: That's why you have very similar implications.
Russ Roberts: And they've been going on, those two changes--the mix of trade versus automation has shifted in the last 50 years. They've both gotten so much larger, is part of it, also. But, certainly, this has been going on in America and the world for a long time. And it didn't lead to the kind of issues that you and I are both worried about.
Daron Acemoglu: Great question. That's where my research in this area started.
Russ Roberts: So, talk about that.
Daron Acemoglu: Because, you know, you are 100% right. They have been going on for thousands of years. But most prominently, you know, the Industrial Revolution was nothing but a concerted effort in automating, mechanizing production in several industries more or less at the same time. You know: spinning and weaving innovations at the early stages of the British Industrial Revolution were, are just quintessential examples of automation. So, then, the question that you ask is obviously a very important one: How come things worked out so well in the past? And I have two answers for that. First of all, they haven't worked out so well at every turn.
Russ Roberts: Exactly.
Daron Acemoglu: But more importantly, the way that they have turned out well has not been an effortless, automatic implication of automation. It's been because other adjustments--technology and social--have been made at the same time. And, let me elaborate this point, starting from the British Industrial Revolution and then bringing it up all the way to the early--to the recent economic history. So, one of the important surprising facts that economic historians of the Industrial Revolution have struggled with is what they used to call the living standards paradox, or what more recently, Bob Allen, economic historian at Oxford called Engels' Pause, after Friedrich Engels. Which is that, for about 80 years--data are scant, so, you know, this is a rough estimate--but, for a long period of time, exceeding half a century, there were all these amazing innovations--flood of machines that made people dub this period the Industrial Revolution; but, at the same time, wages were stagnant. So, wages start increasing in the British economy only in the, towards the middle of the 19th century. And, when do they--why do they start increasing? Well, it's over-determined, but we know several things happen around that time. First of all, technology starts changing, and not just putting the emphasis on automation but a variety of other things that are starting to create new tasks, new occupations, new activities in emerging new industries. And this has been undergirded by a complete institutional change--you know, Britain starts becoming democratic, with the First Reform Act of 1832. Then, you know, the vestiges of some sort of a social safety net are administered by the state starts cropping up. Investment in education increases. And, quite crucially, trade unions start organizing and protecting workers against excessive power of firms, especially because early stages of the Industrial Revolution, you know, was replacing the labor of high-paid, relatively high-paid by the time, by the standards of the time, artisans, by machines and labor of women and children. So, all of these things sort of start changing at the social, political, and economic level. And, it is this combination that finally brings wages back to a state of growth. And if you look at the U.S. history over the last--since, say, since WWII, something I have done, for instance, in my work with Pascual Restrepo, what you find is that there is rapid automation. But, this automation is being coupled with a lot of other technologies that are at the same time creating new tasks, new jobs. And it's like a race between these two things. On the one hand, you have automation reducing the labor's share and pulling down labor demand below the growth of productivity; and at the same time you have these new sectors, new occupations, new tasks cropping up, and those are pulling it, pushing it up. And it's sort of a remarkable picture, that these two forces are going and some of them are faster in some industries and others are faster in other industries. And then you aggregate it up, they balance each other out. But there is nothing inherent in this balancing out. It is a function of institutional choices, and it is a function of technological choices that we make. And then that balance completely disappears from 1980 onwards, or sometimes the 1970s onwards, where automation starts growing much faster, and the other, what we call reinstating technologies slow down sharply.
Russ Roberts: What we're talking about here often goes under the name creative destruction, where innovation is portrayed destroy often certainly factories, if not industries. And new industries and new factories come along. But, of course, they don't come along at some kind of even pace. And they don't come along at the same rate for all workers. I'm very skeptical of the role of trade unions, say, in that British story--although I'd be interested in--maybe I don't know enough about it. But I think in America, the irony is when you look at the post-war period, the post-war period shows pretty much a steady, secular decline in unionization. The first half of that post-war period, prosperity is widely shared: jobs are strong; wage growth is strong across the board, across different levels of skill--even though unionization is declining. So, that trend--it's hard to invoke that as a cause. It could be related. There could be a tipping point where below a certain level it becomes important--
Daron Acemoglu: Yeah; I think the point isn't that you necessarily need unionization to grow; and also I think unions are very complex organizations--I mean, how could they not be? And they are sometimes captured--I would not want to touch a trade union by Hoffa with a barge pole. But, if you look at history of industrialization, unions play a critical role. They are the biggest organized block fighting for democracy. The Chartist movement, which was really very important for British democracy, because the first Reform Act of 1832 just extended the franchise to the middle classes. And, the Chartist movement was the one that pushed for the broader extension. You know, they are very much coming from the heart of the organized labor; and they also play a very important role in protecting workers against excessive employers' powers--you know, getting children out of the factories and very difficult working conditions. So, they are complex organizations. But the important thing is that as that particular episode illustrates, they play a balancing role against the power of organized business. And I think it's a fantasy to think that business is not going to organize. Business is going to organize. You see that in political parties; you see that in employer associations; you see that in other different forms. And I think the genius of the market economy or liberal economy, democratic economy, whatever you want to call it, is that you find multipolar structure that balances the organized power of different groups. So, if trade unions were extremely powerful and organized and businesses were unorganized, that would be a very unbalanced structure, too. But, in the United States I think we've gone in a direction of, over time, as you've rightly pointed out, trade unions became weaker and weaker, and businesses became better and better organized.
Russ Roberts: I don't see that at all. Although, in recent years, in the tech sector, they seem to have more market power. But, I'm not even convinced of that. I don't see the--I don't see the organization of unions as an important counterbalance. In fact, unions generally raise wages by restricting access to those jobs. It's such a mixed bag. It's so not clear. It's similar to the minimum wage, which you also invoke, right? The minimum wage certainly helps the workers who benefit from it, who keep their jobs; we worry, I think correctly, about the ones--
Daron Acemoglu: Well, actually, the minimum wage--I'm a bigger fan of the minimum wage than the unions. I'm more than happy to talk about that. Because--and in fact, some of the efficiency-inducing effects of unions work very similar to the minimum wage, but in a more complicated way and with other distortions. That's why I like the minimum wage better, as long as it's not at excessive levels; and in the United States, clearly it's not at excessive levels. It's essentially very low. And I'm more than happy to talk about that. But, I want to make one other point before--you can go wherever you want to go, but--
Russ Roberts: Go ahead.
Daron Acemoglu: But before we go there, I want to make one other point. I completely agree with you: Creative destruction is key. You know, I will always be a big admirer of Schumpeter and his ideas; and we make creative destruction an important part of our narrative in Why Nations Fail, James Robinson and I. But, creative destruction can happen in many different ways. And I guess one way of articulating what I'm saying is that there is no guarantee that, left to its own devices with no government intervention and no other political intervention, social intervention, the market economy will automatically generate the right type of creative destruction. And again, let me go back to automation versus new tasks. So, it may well be that there will be a lot of creative destruction: new firms coming with more and more automated technologies and making more profitable and cheaper products. But that won't created shared prosperity unless some of that creative destruction energy is directed towards not just automating but creating more labor-intensive new tasks that contribute to productivity and at the same time contribute to the creation of good jobs.
Russ Roberts: So, we're going to find an area of agreement in about 5 minutes. I just want to let listeners know that that's going to happen. And I'm letting you know, Daron.
Daron Acemoglu: Hah, hah, hah, hah!
Russ Roberts: But I want to disagree with one piece of that. It's a bit of a straw man, right, to say there's no guarantee that the market without any government--of course, without any government? Sure. Without any government, maybe there'd be cartels--
Daron Acemoglu: Sorry; I should have been clearer. "Without any government"--I didn't mean, like, complete anarchy. I meant, like, the night watchman state of Nozick or the simplest principles of economics course, we let the market rip but don't interfere with it. I think the role of the government in regulating and redirecting entrepreneurship and innovative activity. So, yes: it's actually by, I mean, a relatively present government as being important.
Russ Roberts: So, then I would respond as follows--and soon we'll get to the agreement. Here's the question, then. There's no guarantee that government will do that, either. So, why is it that you are willing to have an assumed role for government that is precise, surgical, designed by economists, etc., that would steer this complicated machine that has all these multi-causal?
Daron Acemoglu: Oh, I wouldn't trust the economists further than I could throw them. No. I'm trusting democratic process.
Russ Roberts: Whoa. Really? The democratic process?
Daron Acemoglu: Right. Oh, yeah. Of course. Right, right. I'm a big optimist about democracy. I think the problem we have is we don't have enough of it.
Russ Roberts: What do you mean by that?
Daron Acemoglu: I think part of this problem, now we're going to the next layer; we haven't reached our point of agreement--
Russ Roberts: It's coming--
Daron Acemoglu: But part of it is because our democracy is broken. I think all of these issues are related to the fact that democratic monitoring mechanisms on the government have weakened.
Russ Roberts: For example?
Daron Acemoglu: And of course, there isn't--let me just finish this thought and then I'll come to that. And of course, there isn't--it's a fantasy to think that the government or its technocrats could perfectly design anything or, you know, can find just the right balance. But the argument is that government regulation often driven by a complex of vote-getting incentives but also well-meaning democratic efficiency-related incentives, puts constraints and creates guardrails, safety rails, that helps the process. And I think the best example of that, actually, is the Nordic model. There are lots of things wrong with the Nordic model. Anybody who goes and studies it will see there were lots of things that didn't work, and there were lots of inefficiencies. But it's also remarkable how well many aspects of the system worked. And, the whole thing was undergirded by democratic politics. It was a democratic government introducing it; and it was introducing it as a vote-getting strategy because people were really depressed about, you know, the Great Depression and things not working, and looking for, of course, something new: it was the Swedish Workers' Party after 1932. And it did so, you know, not as an imposition on businesses, but trying to create this sort of balance of power that I was mentioning, that was the essence of the corporatist model. And, not through some sort of very heavy taxation--'We're going redistribute everything from businesses'--but 'We're going to create incentives for businesses to do the kind of investment and innovation that would create jobs and high wage growth.' And that's sort of worked remarkably well for, you know, essentially about 5 decades in Sweden, and then subsequently in Norway and Denmark, and later on in Finland. And Germany. So, it's that sort of democratically-girded government as sort of one of the multitude of parties that are part of the bargain that make sure the system has some stability.
Russ Roberts: I think that's--so, I don't find that compelling. I think the--I find part of it compelling. Let's make it clear. The Swedish story I think is quite complicated. My understanding is that there have been quite a few changes--
Daron Acemoglu: Of course--
Russ Roberts: through the democratic process of that role of government--
Daron Acemoglu: Of course. Absolutely.
Russ Roberts: which was very interventionist initially; less so later. More capitalist now--
Daron Acemoglu: Well, you know, I think those interventions are exactly the essence of the democratic process. So--
Russ Roberts: So, I agree with you. I agree with you. That's a great point. So, here's the challenge--
Daron Acemoglu: So, let's go into the details. If you want.
Russ Roberts: Go ahead.
Daron Acemoglu: I mean, the details are that, you know, the Workers' Party comes to power, in coalition with the Agrarian Party--that's why they call it the Cow Trade, Red/Green sort of coalition. And their first priority is to sort of increase wages. And create labor demand. And the way that they do that is actually by bringing businesses into the coalition. So, they say, 'Well, let's form a sort of coalition between the government, so run by the Social Democrats--workers, trade unions that are very important and powerful, the agrarian interests and the corporate interests.' And, and, at the heart of it is, 'We're going to have relatively high wages, but those high wages are not going to become excessively taxing on the investment.' And the way that they achieved that is by setting these industry-level wage bargains. And one of the geniuses of this industry-level wage bargains is that, if your wages are set at the industry level, and if you invest more and you become very productive--you introduce new technologies--you are not, you are not paying higher wages. So, you are at the margin the residual claimant. And that's what really drew the businesses into the model, into the system. And especially the productive businesses really thrived. The export-led businesses really thrived under this system. Because, you know, what happens in the United States with or without unions, in the United Kingdom with their multiple[?] of unions, is that, if you invest and you increase your productivity you start paying higher wages. Not so much in Sweden. But then, exactly like what you said: The system is not perfect. It's a complex system. So, after about 30 years, trade unions started pushing for, uh, these Wage Funds, saying, 'It's not fair that some of the companies are making a lot of money, so we should get that money away from them.' And that would have completely been a repudiation of the system, because what made the system work as a market mechanism was this incentive for firms to introduce new investments and technologies. And they do that. The convince the Social Democratic government. But then the voters see this and they hand a huge defeat to the Social Democratic government. And they claw that back. So, this is a place in which you see the democratic process when it's open, transparent, and open to political competition. You know, it's got the ability to actually self-correct these types of excesses, whether they come from businesses or they come from workers.
Russ Roberts: And it seems to me that that system is pretty well--has a much better self-correcting aspect to it. Which is, I think, where we agree on this. I think the challenge is whether the United States is anything remotely like the Nordic countries in its political process, or whether it ever could be--
Daron Acemoglu: Of course. Of course. We're not, at the moment.
Russ Roberts: They're small--
Daron Acemoglu: Yeah, yeah. Very heterogeneous: We're very, very big. It's [?].
Russ Roberts: Yeah. So, it doesn't seem to be so responsive. And effective. Just like: 'Our schools aren't as good as theirs.' There's a lot of things that are different.
Daron Acemoglu: Well, our schools used to be better.
Russ Roberts: Yes.
Daron Acemoglu: Our schools used to be much better than theirs. So we were a leader in schools. And, incidentally, it's a little anecdote: You know, what FDR [Franklin Delano Roosevelt] originally tried to do was very similar to what the Workers' Party did in Sweden. So, FDR's industrial and agrarian policies were very similar. But they got blocked. And, whether they would have worked in the United States, that's an open question. But the solutions again, through a political mechanism--that a democratic government in the United States in the early 1930s came up with--was actually not that different from the Swedish experience.
Russ Roberts: My only point is that I just don't think the American political system has the nimble responsiveness of those Nordic systems. Either because theirs are dramatically smaller, or their society is more homogeneous. What will be an interesting test--is, they have become--
Daron Acemoglu: Yeah--
Russ Roberts: less homogeneous--to see how they sustain that.
Daron Acemoglu: Yeah. I don't disagree. I don't disagree that they face, uh, less extreme political challenges. But, you know, the American political system--I think on this one you have to agree with me: I don't think we've found the point of happy agreement yet.
Russ Roberts: It's coming. It's coming.
Daron Acemoglu: But, it's not--it would be that the American system has become much more dysfunctional.
Russ Roberts: Yeah. I don't think we understand why. But I think it's, I think part of it is--the standard answer is that it's, you know, the role of money. My view is that we have changed--just as we have changed the norms of what's acceptable, corporate level, which you talk about--and I'm somewhat sympathetic to that--I see that in, say, the pharmaceutical industry where pricing, I think, would have been unacceptable 25, 50, 60 years ago, is now considered like, 'Well, why not?' You know, 'It's okay. Let's raise it 20%. We don't have an excuse, but it doesn't matter: We can.' And I think similarly there are some norms in how politicians behave which have changed.
Daron Acemoglu: Absolutely. 100%. I completely agree. And, and--but there are multiple aspects of this. You know, the data, I think, again multiple interpretations are possible. But the data are striking. So, the sort of, the type of analysis that Nolan McCarty, Poole, and Rosenthal do, which is to look at polarization in the legislature. It's striking. So, if you look at the voting patterns of Democrats and Republicans in the Senate or the House in the 1950s and 1960s, you know, many Democrats voted like Republicans on many bills; and many Republicans voted like Democrats on many bills. Today, there is no--
Russ Roberts: [?]
Daron Acemoglu: there is not even--not even the most moderate Republican votes way more Right-wing than the most moderate Democrat. And I think this cannot be completely unrelated to money and politics, because one of the things that happens--and we saw that during the Obamacare: If a Republican were to go and join the Democrats--and I think this was really tempting for some of the moderate Republican Senators like Olympia Snow--they would immediately face an extremely well-funded, thanks to Koch brothers or you know, other energy sector, or other sort of stalwart supporters of extreme Right-wing causes--they would force an extremely well-funded challenge in the primaries. And many of them that weren't actually so moderate, like Cantor or Paul Ryan have lost, because of that sort of, or came close to losing because of those sort of challenges. So I think that is one part of the equation. But, you are absolutely right: In the past, even if money spoke so loudly, norms would have also gotten in the way. And I think organizations--the organization of the Republican Party, you know--the Republican Party, and the Democrats, also, to a great extent--have become much better at controlling dissent within themselves. Which also forces more group voting. And that is an institutional element adding to the role of money and norms.
Russ Roberts: The only thing I would say to that, besides the fact that I would just defend the Kochs as being--they are not Right Wing--on many issues they are not Right Wing. I think it's important to say that. They get vilified as if they are responsible for everything negative about our society--
Daron Acemoglu: Oh--I used to have much more sympathy for them in the past when they, you know, supported Libertarian causes like the [?]--
Russ Roberts: They're pro-immigration; they're still pro-immigration.
Daron Acemoglu: Right. Oh, well, they don't speak much about it because they don't want to anger Trump, so I don't know.
Russ Roberts: Let's put that to the side. I think where we agree, and it's a troubling reality, is that the two-party political system is not very competitive. It's less competitive than it used to be. There is less--their ability to control their members is stronger; and I think that's a bad thing. I don't think we understand exactly why that is, but it's true.
Daron Acemoglu: I agree. I think there's much more to learn there.
Russ Roberts: Now let's talk about where we really agree. So, where we really agree--
Daron Acemoglu: Minimum wages. I didn't know you really agreed with me on minimum wages, Russ.
Russ Roberts: No, it's not minimum wages. Where we really agree is the people at the bottom. So, when we think about creative destruction and the opportunities, it's interesting you did not mention the benefits of globalization to consumers. You mentioned producers, but I think it benefits, obviously, consumers--
Daron Acemoglu: Oh, yeah, obviously.
Russ Roberts: as does automation, through competition, through forcing businesses to share the cost, improvements with consumers. And normally, that reduction in resources required to produce stuff creates opportunities elsewhere in the economy: it draws workers into those areas. The workers who are left behind then become opportunities for businesses to create work for them. And that isn't happening the way it used to, it seems to me. And in particular--this is where we agree--it's not happening for the lowest-skilled workers. Workers who, say, graduate high school or do not even graduate high school used to be able to have a decent standard of living in manufacturing, say, or in some parts of the service sector and are not finding those opportunities now. So, the question is--
Daron Acemoglu: I would go further. I would actually say, or college graduates. So, college graduates without advanced degrees--Masters, MBAs [Masters of Business Administration], Ph.Ds. [Doctors of Philosophy], specialized degrees--they are in the same boat. They are finding it harder and harder to get those good jobs.
Russ Roberts: It's what they study. If you are in a STEM [science, technology, engineering, and mathematics] field, I don't think that's true. I think even if you--most business majors do well. I think if you major in the liberal arts--and those ranks are diminished greatly, which makes me sad because I'm a fan of liberal arts; I'm a fan of the humanities; I'm a fan of English and Philosophy--and there's a lot of debate about why they are less attractive and effective. But it's certainly the case that the people who major in those things struggle to find--can struggle to find work. Including social sciences, I would add--not just the--
Daron Acemoglu: Well, I'm generalizing. But, again, subject to the CPI issue--if you look at the real wages of college graduate men without post-graduate education, they have been fairly stagnant for the last 20 years.
Russ Roberts: Yeah; of course some of that is because--
Daron Acemoglu: Post-graduate wages are growing very rapidly--
Russ Roberts: One of the challenges--
Daron Acemoglu: but not--
Russ Roberts: of interpreting that, which is often forgotten, is that the proportion of the population going into college has grown dramatically, especially in the last 20 years. So, it's not the same people. It's different levels of ability. You would expect that to slow--at least the rate of growth. If it's stagnant, then that's a little bit alarming; I agree. But let's talk about the people who we know are struggling, which people who, say, don't finish high school, who have an inability to be part of the so-called knowledge economy. And they end up, say, driving Uber. But maybe Uber is going to disappear soon, and so--because not Uber [?], but say autonomous cars show up: I'm starting to think they're not coming. But, the worries that people have about low-skilled workers being able to participate in the modern economy I think is a real concern, and I think more importantly--
Daron Acemoglu: Absolutely. Agreed.
Russ Roberts: and more importantly--
Daron Acemoglu: One more point of agreement.
Russ Roberts: And I've got one more for you, too: Even though the focus is on the material side of the standard of living, I think that echoing what we talked about earlier, I think the dignity part is more important. Equal if not more important, in that there are too many people in America today who don't feel like they matter, and who are struggling to feel self-respect. And those are the folks that I'm most worried about. So, what could we do for them? You have some suggestions in your essay. We're not going to [?] on those probably. But, lay them out.
Daron Acemoglu: Yeah. No, 100%. And I think actually it was--so let me reiterate the agreement. I think people call it middle class or lower middle class falling behind is a real problem. But, also, the fact that they feel neglected and ignored is even worse; and I think it's one of the failures of the--[?] the intellectual elites, including, you know, the sort of the leaders of the Democratic Party--you know, they did not understand all of that resentment in the heartland of America. And the same thing in Britain, where, you know, people were not getting good jobs; they were not getting the wage raises; they were not finding that they were sharing the gains. And many of them were actually losing their jobs because of automation and globalization. And their voices were unheard. And the longer they became, they were unheard, the more resentment they started, rightly, to embody. So I think it is critical to deal with this problem. And my sort of perspective on this is that, you know, of course, the standard levers that economists emphasize are important. Social safety net--the better the social safety net, the more the sort of the protection we can provide to help them get into new jobs, better education for the next generation. But, I think, critical, also, is a more concerted effort to create good jobs, and good jobs not just for tech entrepreneurs, but good jobs for the middle class, for people without college degrees, without post-graduate degrees. And to do that, I think we need both institutional and broader measures. So, going back to my earlier discussion, if indeed there is no guarantee that the market economy will channel its creative destruction energies, technological dynamism, into technologies that will support, underpin shared prosperity, then the government does have a role in--as a provider, the most important provider of funding for new innovations and technologies, to actually take that into account. The example I give there is climate change. Look, we've made such amazing progress against climate change over the last 30 years, because 30 years ago green technologies--solar, wind, geothermal--they were no where to be close to competitive with fossil fuel energy. Today, they are pretty close to being competitive. If it wasn't for the shale gas discoveries--you know, today, green energy would be head to head with most of the coal powered energy. So, that didn't happen because we slapped big carbon taxes--I mean, carbon taxes play an important role and we should have them. But it happened because the government actually supported green technology. In other words, the government played a role in the direction of technological change. It's the same when it comes to new technologies, and one of the things I think the government has to take into account is: Are we investing enough in the blue-sky technologies that are going to be the job generators of tomorrow? Are we putting too much emphasis on automation, automation, automation at the expense of creating new jobs, new tasks, new labor-complementary technologies? I think those are important questions we have to ask. And if you look at the track record of the government, it's far from perfect, but it also has many successes of, you know, being at the early stages of completely transformative technologies. I would also add to that list other institutional factors. Minimum wages. You know, this is the conceptual point that I wanted to make, and let me make it. This goes back to research I did in the 2000s, late 1990s. So, if you look at an economy where labor markets are not competitive--in particular, for instance, you have search and matching and some sort of wage bargaining--you can easily show, it's a first-order prediction, that the market economy will have an incentive to create too few good jobs and too many bad jobs. Where good jobs and bad jobs are distinguished by how much up-front investment you make. And it's for a very simple reason. If I make an up-front investment, I produce more. And I pay higher wages. But those higher wages also mean that I'm not the full residual claimant. So, I under-invest and I tend to opt for lower-wage jobs. And you can show that, if that's the way the economy looks like--and there is some evidence suggesting that's not a very bad approximation to many sectors--then minimum wages that put a floor on how low wages can get is part of the optimal policy. And that's the sense in which I think minimum wages should be justified not just on sort of monopsony, but on this good jobs wage bad jobs [?], and that sort of brings it back to, really the priority of good jobs, and that's why we need the institutional foundations. Such as minimum wages and to some degree unemployment benefits and unions to make sure that there isn't an excessive--too excessive--incentive for bad jobs at the expense of good jobs.
Russ Roberts: So, I don't agree with that, but that's neither here nor there--
Daron Acemoglu: Oh, I thought you were going to agree with that 100%.
Russ Roberts: No. No, I'm not--
Daron Acemoglu: Hah, hah, hah, hah--
Russ Roberts: No, but I'm not going to push on it. I think--we'll put a link up to the paper. You'll send me a link to it, and we'll let readers think about whether that's convincing on their own. But, I want to try a different approach--
Daron Acemoglu: Please.
Russ Roberts: I want to find a different area of agreement. Regardless of whether these structural issues in the labor market are important or not--and I may underestimate them, and you may overestimate them--it would certainly help, regardless of what your view on that is, it would certainly help if workers had more skill--
Daron Acemoglu: Absolutely.
Russ Roberts: And I find it remarkable, and tragic, that we don't spend--'we' meaning our profession--don't spend more time on how to make our education system better. We have very disturbing evidence on the relationship between spending and outcomes. We can debate--people debate it all the time. But the default position that people have to explain is that there doesn't seem to be much of a relationship between per-student spending and various measures of outcomes. You could argue that outcomes aren't well-measured, that they are flawed. But the idea of "spending more money" has not proven to be very helpful--
Daron Acemoglu: Well, let me not get into that, because that's complex topic. But let me agree with you on the education point. I think I 100% education is key, and we are doing a lousy job of asking what type of education. You know, it's actually really ironic that we're witnessing in economics a complete resurgence of education research. There's so many papers written on education, because data are available, you know, on school assignment, on various aspects of what goes [?]--
Russ Roberts: Charter schools--
Daron Acemoglu: Charter schools. Etc., etc. But, none of that, in my opinion, is asking the important question, which is: How should we design the curriculum--
Russ Roberts: Agreed--
Daron Acemoglu: What are the skills that the students need? And what are the skills that students from different backgrounds need? I mean, go and visit an elite school, especially where children come from upper-middle-class backgrounds, and one where, you know, they come from lower-middle-class backgrounds. You'll see they are just completely different environments. And those students need very different sorts of instruction. The type of instruction that they need is different. The skills that they need to get, they are not getting them, because our curricula are completely outdated. Our students don't have, our teachers don't have necessarily the right skills. I think those are really critical issues that we are not dealing with. But I would also add to that training--you know, one of the things, again, going back to the work that I did in 1990s, the sort of the same things that create this imbalance between good jobs and bad jobs also creates two little training, too little investment in the skills of workers by their employers. And again, some of those arrangements such as unions and minimum wages have the added benefit that they tend to trigger, if embedded in the right institutional setup, they tend to trigger more training investments.
Russ Roberts: Yeah--I think the--one way to summarize our expanding sphere of agreement, Daron, is that the current emphasis on K-12 [Kindergarten through 12th grade] as the entry port into a college experience is a mistake. Not everyone can go to college. Not everyone should go to college. And, we need some alternative ways to enrich the skillset of folks who aren't going.
Daron Acemoglu: Absolutely. 100%. But it's a very difficult question. Because, one alternative is the German system--and I've studied the German system a little bit, and I have respect for it. But it is too scripted, and it does not give second chances which I think are very important given the fabric of U.S. society. So, the challenge for us is: How do you design a system that really doesn't treat every course as if it's preparing you for, you know, junior calculus in one of the state universities, while at the same time gives second chances and ability for social mobility to people who come from all kinds of different walks of life. So I think that's the American challenge.
Russ Roberts: Yeah, I agree with you. And I think--I just wish we would spend a lot more time on it--a lot more research. It just seems to me it's the central problem that we could make--again, obviously you and I would agree on this--if we could improve that, there would be more good jobs no matter what the state of the labor market, creative destruction, globalization, etc. is.
Daron Acemoglu: Yeah. Let me--I'm sure we're running out of time. But let me end with one observation: There are many things that are different, but the evidence is that when U.S. manufacturing firms adopt robots, they cut employment; especially they cut employment of their production workers. When German companies do so, they cut less; they find other ways of using the skills of their line workers. And I think that is not unrelated to the fact that those are trained workers that have more embedded health care in them. They have expertise and the employment don't want to lose them.
Russ Roberts: Excellent point.