Chris P. Dialynas Distinguished Service Professor of Economics
John Templeton Keller Professor of Behavioral Science and Neubauer Family Faculty Fellow
V. Duane Rath Professor of Economics, Deputy Director of the Becker Friedman Institute, and John E. Jeuck Faculty Fellow
Is there an inherent tension between efficiency and equality?
Bertrand: Definitely, and as economists, we’re really focused on efficiency. We have scarce resources, and we are always trying to think about how to allocate those resources to create the most value possible. The issue of distributing the value that’s created comes up typically as a second step.
Epley: Whenever you’ve got two goals you could be pursuing at the same time, it’s very hard, and that produces a general bias in behavior of pursuing one goal at a time, often at the expense of the other. In general, decision makers are myopic, and end up prioritizing one goal over the other.
Bertrand: Economists would argue that the right thing indeed is to focus on efficiency, to make the pie as big as possible, and then it’s for policy makers or voters to decide how they want to redistribute the pie. But economists also have some views as to the efficiency considerations that should go into that second stage.
Hurst: In our models, we say it’s good if countries move toward their comparative advantages—one country specializes in Good X, another one specializes in Good Y, and they trade with each other. But if you think about manufacturing in the United States, the people who used to produce tradable goods that are now being produced abroad got dislocated, and they paid some costs.
In a frictionless labor market, those costs are small. If you get fired on this side of the room, you just move to that side of the room. In real-life transitions, we know that the labor markets aren’t totally frictionless, but how quick the speed of adjustment is from Sector A to Sector B is important. The frictions come along three dimensions. One dimension is: people could move sectors within a given location for the same level of skill. The second is: they can move locations and get jobs in another place, again, for the same level of skill. And the third, which is what we see over long periods of time, is: they could accumulate more skill. The frictions are: How costly is it to move? How costly is it to transfer into another sector? And how costly is it to acquire the skills the labor market is demanding? Thinking about the costs versus the benefits is important. We tend to spend a lot of time thinking about the benefits. We’re not always as focused on costs, or how they might be changing across people or across time.
Bertrand: If we assume perfectly functioning labor markets, and solely focus on efficiency, we may make the wrong calculations. There might not be efficiency gains from a trade policy if there are permanent losses for some displaced workers. Once we look at the real labor markets and the losses that are created, we might find the policy does not increase the pie but makes it smaller.
Epley: There’s a deeper psychological friction here that has to do with individual identity. Think about people losing their jobs in the coal industry in West Virginia. They’re coal miners, but that’s not just a job; it’s their identity. It’s who they are, and what their parents and grandparents were. It’s deep in them. Even if they could go to Walmart and make the same amount of money, many will have a hard time doing that. Switching jobs into something that doesn’t seem as meaningful or as consistent with their identity creates another friction. Models tend to focus on things we can measure easily, which makes it hard to account for factors such as loss of identity. Where do you put that in the equation? What does that even look like?
Why is this issue particularly pertinent now?
Bertrand: We live in a moment where some of these groups that have been on the losing end of many of the shocks that were very good for society overall have been expressing themselves, probably much more than they were 10 or 15 years ago.
Hurst: It’s about technology and automation. We produce certain types of goods with technology more than labor now, particularly low-skilled goods, and that displaces workers with lower levels of skills. That has always happened, and so what’s different? Think about how the introduction of the tractor affected agricultural employment. The booming sector at the time was manufacturing. And the skills of agricultural workers and manufacturing workers were very substitutable. Migration was easier. Now, we’re basically moving lots of low-skilled jobs together.
Bertrand: Machines are certainly a bigger factor than trade, but it’s very clear that people have picked on trade because it’s a much better narrative for politicians who want to harness this energy. It’s hard to get people angry at machines. It’s easier to get them angry at the Chinese, or immigrants. There’s not as much of a backlash against trade in Europe as there is in the US. The reason may be because Europe has a stronger social-safety net, and so when jobs get destroyed, there’s more compensation from the state to partly undo some of the concentrated losses.
Hurst: The US lost a lot of manufacturing jobs in the 1980s, although not as many as in the ’00s, but it didn’t show up in employment rates the same way that it is showing up now. At the same time, we see declines in mobility across space. And it’s not only manufacturing jobs that are disappearing; lots of other low-skill ones are too. We’ve had lots of adjustments to past shocks, and the people who haven’t adjusted yet to those past shocks might be inherently different than the people who were exposed to previous shocks.
Think about education. Traditionally, very few people went to college. Then a shock came along, the return to skill went up, and we went and got educated. This scenario repeats itself. So who’s going to get the education? The people to whom it’s most beneficial. Who is left? Those for whom skill is less beneficial. And over time, the pool becomes skewed increasingly toward the latter group, so that for any given shock, they might be less responsive to these margins of adjustment than in the past. This means that for each successive skill-based shock, the pool that is left becomes less responsive. The unskilled who are left might be harder to move than in the past. In my own work, I’m trying to formalize this and test it. The propensity to get schooling, particularly for men, has flattened pronouncedly over the past 20 years relative to the prior 100 years, even though the return to skill actually grew over that time period.
How should policy makers think differently about equality and efficiency?
Epley: You could perhaps align efficiency and equality if you thought more about maximizing well-being, which social safety nets do. All this raises the question, what should we expect from people? As a society, when we form these policies, what ought we be trying to do? Economic models of efficiency assume that people ought to be maximally responsive to shocks. The working mother whose job is lost should be willing to up and move to another job. The coal miner who loses his job in West Virginia but wants to continue living in this community and be part of this system, should he just be expected to up and leave the community and go to California, where there are jobs?
We may decide that’s not the kind of country we want to live in. Instead, we could try to make people happy where they are. If we thought more about well-being, and incorporated some of these nonmonetary factors into well-being, it might lead to very different kinds of policy, because we wouldn’t assume that people would want to do things or should do things.
Bertrand: There are people, such as [University of Chicago’s] Eric Posner, who are thinking about how we could adapt cost-benefit calculations to account for the permanent impact of job losses. On top of just asking the standard questions—What’s going to happen to prices? What’s going to happen to cost?—you could also incorporate factors such as the number of jobs that would be destroyed and the total dollar value of these job losses, factoring in the loss of skills that people have invested in. This might lead to different choices on specific, concrete policy decisions. Should the Environmental Protection Agency pass this rule versus that rule? Should we open the door to trade? We could try to put a dollar figure on some of the factors Nick is talking about.
Note that some of this is already happening, particularly when it comes to accounting for possible concentrated losses. There’s such a thing as a feasibility analysis, where a government agency might be considering a policy that may pass the cost-benefit test, but it would mean shutting down 80 percent of the plants in a particular county, and that’s something we would deem to be infeasible. It’s about doing two things: first, looking at costs and benefits, and second, determining a threshold of how many job losses you are willing to have in the particular part of the country. Then you have to find a way to balance these considerations.
Hurst: The auto bailout that we had in the US during the Great Recession had the concentration consideration in there for a particular area. We don’t have a lot of policies that are designed to deal with these long-term adjustments. If you think about the big transfer policies, particularly for men in the US, there’s unemployment insurance, but that’s usually for short-term fluctuations—you go into a recession, you want to help somebody for a couple of months, or maybe up to a year or two, to ride out the storm. Then we have disability, which is basically a long-term insurance program, usually for health shocks, to help smooth out those types of events.
We are relatively underdeveloped in terms of policies in response to industry decline, which is not a health issue, and not a temporary measure such as unemployment insurance. We do have some trade-adjustment policies around retraining. But this doesn’t really get to any of the types of stories that Nick said might be around in the background. It doesn’t talk about some of the long-term dynamics that we see. It doesn’t talk about this long-term skill adjustment. When you think about these issues, how do you implement policies that also don’t have large efficiency effects on the background? That’s the trade-off. If you start transferring money to people in certain areas, that might slow down adjustments for the next generation, or the generation after that. So how do you get the right policy?
Now suppose we want to do efficiency-gain-type policies. There are some losers. We want to redistribute the pie. We really don’t have any mechanism for that redistribution for the types of shocks we’ve experienced in recent decades.
Epley: Let’s go back to the concrete example of the coal miners in West Virginia. There are analyses of what really drives well-being. Social connection turns out to be hugely important. It’s much bigger than the effects of income on day-to-day well-being and happiness. A policy could take into account the cost of relocation when a group picks up and moves, because now they’re not around friends and relatives whom they’ve been around
for a long time, which is likely to affect their health.
Coal miners are involved in the energy industry. That’s part of their identity—they’re making power for the planet. You could imagine a policy that tries to keep them there, and tries to subsidize some other social good, such as wind power. Instead of blowing the tops off the mountains, you could put windmills on top of them. Retrain these workers to stay close. It’s probably not the most efficient outcome. That’s probably not the best place for the windmill, and so you’d sacrifice some efficiency. But would there be an overall gain in utility, not just for those people but for society as a whole, of keeping them there, retraining them to do green energy in the state of West Virginia? Would that be better overall for well-being? If you were to add well-being and health into some of these metrics, not just money in terms of efficiency, some of these policies that wouldn’t perhaps make sense on a pure efficiency calculation might start to make more sense.