Q&A: Stanford economist Eric Hanushek on COVID’s trillion-dollar impact on students
Kevin Mahnken | October 12, 2023
Your donation will help us produce journalism like this. Please give today.
Experts have spent years trying to quantify the pandemic’s toll on a generation of K–12 students. Some have focused on the months of incomplete or nonexistent learning opportunities while instruction was being delivered remotely in 2020 and 2021. Others were most disturbed by the deferred development of social-emotional skills for the youngest students, or the damage dealt to the mental health of adolescents.
All significant harms. But then there’s the bottom-line figure that appeared last winter: $28 trillion.
That’s the projected cost, in lifetime earnings, to the children whose academic abilities were set back during the COVID-era, totalling about $70,000 per person over the course of their careers. The figure was reached by Stanford economist Eric Hanushek, based on the cratering eighth-grade math performance measured by last year’s National Assessment of Educational Progress. And it could be permanent if schools don’t do something about those diminished skills.
One of the most influential and widely cited academics studying American education, Hanushek has spent over 40 years studying how schools lift kids’ achievement and whether test scores translate into later-life success. A longtime fellow of the conservative Hoover Institution, he won the prestigious Yidan Prize for Education Research in 2021, which brought a $3.9 million award for new projects. He is also deeply involved in some of the biggest ongoing debates in education policy, including whether the achievement gap between rich and poor students is growing or slowly closing.
Above all other issues, Hanushek is tied to the question of whether spending more on schools will consistently result in better student outcomes. He has been both resolutely skeptical of the proposition and influential in the statehouses and courtrooms deciding whether to increase education funding. Over the last decade, a raft of studies have offered new evidence that increasing expenditures on schooling does, in fact, lift achievement — a developing literature that Hanushek acknowledged in a lengthy review released this spring, while still insisting that the amount of dollars expended matter far less than the quality of the interventions they underwrite.
Money is also at the center of what the veteran researcher calls his “simple and complete” fix for COVID-related learning setbacks, detailed in a 2022 policy brief. But rather than paying to lengthen school days or fund tutoring programs, Hanushek advises that districts spend remaining pandemic relief aid on incentives for the best teachers to take on extra students; he also proposes buying out the contracts of their least effective colleagues.
In a conversation with The 74’s Kevin Mahnken, Hanushek talked about how the most advanced economies reward skill (and punish the lack of it), why he believes voluntary learning initiatives tend to increase achievement gaps and what the United States could do with $28 trillion.
“The cohort that suffered these learning losses isn’t going to be around for much longer,” he said. “We’re graduating 3.5 million of them each year, and they’re going away without any real chance of recovery.”
This conversation was edited for length and clarity.
Can you contextualize what $28 trillion of potentially lost wages actually means? It’s a sum so large as to be almost mystical.
Sure. You know that we had a recession because of the pandemic, when businesses closed. The total cost of that is about 1/15th the same amount, maybe $2 trillion. The best estimates I’ve seen of the effects of the 2008 recession are something like $5 trillion. And if you want a third option, $28 trillion is a little more than one year of America’s GDP. So it’s as if you closed down the economy for a whole year.
Am I right in thinking about this as essentially a huge number of deferred opportunities for growth, like businesses that don’t expand or innovations that are slower to arrive?
It’s everything about the economy. But a better way to think about it for most people is that everybody who was in school during the pandemic will experience 5–6 percent lower lifetime earnings. It’s almost like a 5 or 6 percent added tax, with this cohort earning less than the cohorts immediately ahead of it and immediately behind it, because they’re just less skilled. Unless we do something about it.
But how confident are we that these test results are really measuring skills that are important in later life? Do we have proof that there’s an actual relationship between student scores on NAEP or state exams scores and their future economic activity?
Historically, people haven’t looked so much at the effect of test scores or cognitive skills on earnings because we just haven’t had much data. We’ve recently gotten data from a bunch of countries in the OECD [Organisation for Economic Co-Operation and Development], and it shows that the U.S. rewards economic skills more than almost any other country on earth.
But consider that in reverse. It also means that the U.S. punishes the lack of skills more than almost any other country on earth. What this is essentially showing is that employers really do pay attention to the skills people have. Those skills determine how well people can adapt to new jobs, and they help determine how capable people are in actually doing their jobs. Nobody objects to the idea that people who are smarter generally earn more. And while there is a lot of variation around that — from NBA players, who often don’t finish their degrees, to college graduates who find themselves driving taxis — that’s the kind of relationship we would expect to see.
Is that connection between demonstrated academic achievement and economic success just a result of the highly knowledge-based economies that prevail in rich, Western countries like the U.S.?
Precisely. One country that rewards skills slightly more is Singapore, and Singapore basically doesn’t do anything except for knowledge-based industries.
If it’s so clear that increased educational attainment is linked to better economic prospects, does it worry you that we’ve seen a decline in the number of students applying to college the last few years?
We don’t know if that’s going to last. But in the short run, it’s a potential concern that people aren’t going off to get more skills. Part of it is that in the current economy, it looks like you can get a huge reward from going straight into the labor market after high school. But it just won’t compare with the reward from going to college.
There are variations in the returns [to college]. You can go into some college programs that ultimately don’t pay off, at least if you pay the full tuition. But on average, college graduates are earning something like 75 percent more than high school graduates, and that differential is large enough that people who get fixated on the cost of tuition and debt cost are really making a huge mistake if they don’t go to college. The differential more than compensates for the cost of getting that college degree.
Could you please explain what you’ve dubbed the “simple and complete” way of addressing learning loss?
Every time I mention my preferred strategy, people sort of turn away. [Laughs] People are much happier saying, “Let’s do exactly what we’re doing already, but just a little bit more of it.” And it boils down to lengthening the school day, lengthening the school year, trying some tutoring here and there. But even if you could implement that — which nobody knows how to do — the data suggests that it wouldn’t make up for the kinds of losses that kids have suffered. It’s just not a strong enough treatment.
Something everyone mostly agrees with is that teachers are really important, and that effective teachers are very, very valuable. People start to disagree when you ask, “What do you make of that?” My answer for dealing with the current learning losses of the COVID cohort is to try to get our most effective teachers to just work more intensively by taking on a few more kids. You can, in fact, provide incentives from the existing ESSER money — which will be around for at least another year — and provide them with help in grading, essentially lighten the load and make it possible for them to teach more kids.
The place where everybody stops listening is here: We know that our most ineffective teachers are just harming kids. If we could take some of the ESSER money and buy out the contracts of our least effective teachers, the combination of those things could be enough to make up for the learning loss we’ve seen.
We’re almost completely back to normal schooling. But the cohort that suffered these learning losses isn’t going to be around for much longer; we’re graduating 3.5 million of them each year, and they’re going away without any real chance of recovery. So we have to do something that’s quick — we can’t wait to hire and train a new group of teachers, or wait to figure out which kinds of instruction might help them the most. We just don’t have the time. But my point is that our current workforce is good enough to make up for this! We have, on average, a really good teaching force, though we don’t use the most effective teachers enough, and we overuse the least effective teachers.
The issue of urgency is obviously critical, but I wonder if it really supports your point. Given the logistical and political difficulties of that kind of approach, it seems much simpler to increase the quantity of instruction students receive — by providing tutoring and lengthening the school year, for instance — than by improving the quality through the strategy you describe.
The problem is that nobody is willing to actually make it mandatory that people stay longer in school. Los Angeles had a voluntary program to try to add four days to the school year, and essentially nobody showed up. But even if they did, what would it mean? Say a quarter of your kids actually show up for two extra days: What do you teach them, and how do you integrate that with their normal classroom instruction? You could, at best, improve their soccer skills.
Voluntary added days generally expand variation in academic performance because the kids who need [extra schooling] the least are the ones who tend to show up for voluntary learning days. It’s people at the top of the academic scale who take advantage of voluntary activities, and the people at the bottom don’t. That’s why I said before that we don’t know how to implement things like tutoring and supplemental instruction at scale. We have a few examples of it working, but we don’t know how to put it into 100,000 schools.
Does it surprise you that, in the five years since the Janus ruling, the organizing strength of teachers’ unions seems to have risen in much of the country? Teacher pay became something of a national issue during the summer of 2018, and now you see Republican governors excited to announce pay raises.
It’s a huge obstacle that you can’t get schools and teachers’ unions to agree to anything. We have a strike going on in Oakland Public Schools, where the union is trying to wring the last bit of money out of the pandemic; they’ve been offered over 20 percent pay increases over three years, but they don’t think that’s enough. [After nearly two weeks of school closures, the Oakland teachers’ strike ended on May 16.]
One of my favorite books on my bookshelf is called Teacher Shortages and Salary Schedules. It has a 1962 copyright. The thesis of the book is basically that if you pay all teachers the same amount, you’re likely to get either very underpaid math teachers or very overpaid PE teachers; you tend to see the former, and that’s where some teacher shortages come from. It’s a problem that certainly exists, and in my opinion, the biggest shortage is a shortage of highly effective teachers that we would like to keep in.
Do you think there’s something positive in the settlement we’ve reached in the U.S., wherein we hire lots of teachers but have largely deferred salary increases? I’m reminded of a paper written by your frequent collaborator, Ludger Woessmann, which found that students earn better grades and stay in school longer when they have an adult mentor in school — that seems like something that small class sizes would foster, no?
It can be very important to build close relationships [between students and teachers]. The problem is that everything we measure suggests that reducing class size is not the most important thing, and that having an effective teacher in a classroom is so much more important than having a slightly smaller class.
People get excited if they can see anything positive in the data about smaller classes. But the impact is so small relative to the impact that great teachers have. As you mentioned, teachers are now concerned about low salaries. But part of this Oakland strike has to do with hiring more people too; it’s just hiring more people and paying them more. There are demands for more ancillary people in schools, like counselors and nurses, and they want to put caps on class size as well. So no one’s given up on quantity, which is the thing that unions like. Unions make their money by having more people, so they’re always happy to push for smaller class sizes.
If I could change the subject slightly: We’ve recently finished some work on Dallas, which radically changed the way it evaluates and pays teachers. It’s much more related to effectiveness in the classroom now than under the traditional salary schedule. The other thing they did was introduce an incentive scheme to get the best teachers, measured by prior performance, to work in the lowest-performing schools in Dallas.
We found two things from this experiment. One is that the most effective teachers were willing to do that for relatively modest average salaries. The second was that once they moved to the lowest-performing schools, the worst schools in Dallas began approaching the citywide average within two years. This is a policy that has been shown to work, and work at scale where you can turn around entire schools.
But the unfortunate thing they did was say, “Oops! Now that these schools are performing well, they aren’t eligible for the program anymore.” They took the incentives away, and the best teachers left because there are lots of easier places to work. And they will if they don’t get paid.
You’ve got another paper on Dallas’s pay-for-performance reform. It reminded me of the debate over Washington, D.C.’s IMPACT teacher evaluation system, which people have said is really hard to replicate, in part because the political blowback in the District was so fierce.
There was a bit of blowback in Dallas as well. Mike Miles was the superintendent who devised the entire Dallas system, and he worked very hard to convince a somewhat grumpy school board to put it in place. Then, just as it was starting to come together, he left. So there’s no doubt, this is politically difficult.
The question is, are you willing to give up $28 trillion and settle for having the 31st-ranked school system in the world because it’s politically difficult? I’m not somebody who needs to make my living by politics. But it’s clear that the rewards of adopting these policies are so, so large, even if there are also political costs. Someday I hope the teachers’ unions also see that it’s in their interest to make some marginal changes to the old “no differentiation” policy.
I realize that’s just your estimate, and it’s not as though it would be dropped as a lump sum in the national checking account; but if the value of restoring lost learning is anything approaching $28 trillion, it could fund all our K–12 schools many times over.
The cost of the current U.S. education system is about $750 billion, so you’re talking about potentially funding that system for decades with the money that you could get out of this. Frankly, we have to get some people who are willing to lead in this time of crisis.
Are there any useful lessons from the catastrophe of the pandemic — or perhaps technology, including advances in virtual and asynchronous learning — that we might use to improve schools going forward?
The thing I learned was sort of the opposite.
It became very clear that having a teacher in the classroom is much better than hybrid instruction that only sometimes has a live, in-person teacher. And hybrid instruction, in turn, is better than fully remote instruction. There’s noy doubt that we gained a greater appreciation of the importance of the classroom teacher, even in these kinds of chaotic circumstances. We had other ways of trying to teach, but we couldn’t do away with effective classroom teachers.
We did mobilize the tech industry to try to improve virtual instruction in a variety of ways, and maybe we’ll come out a bit ahead from that. But I worry that we won’t try to learn from when and where virtual instruction works best, and when and where it doesn’t work.
Do you think the public sector, including federal entities like the Institute of Education Sciences, needs to take a more active role as a catalyst for breakthrough learning platforms and technology? Given the mixed record of ed-tech generally, I wonder if we actually have reason to be hopeful.
Sure, I like what IES has been doing. The amount of money we’re talking about there is pathetic, really, compared with the size of the problem. We have systematically underplayed any role of research and evaluation in education compared with other industries. But education is the base industry for most of this other stuff that we’re pouring money into.
The most recent NAEP release shows decades’ worth of progress — this time in civics and U.S. history — essentially erased since the beginning of the pandemic. I’m wondering whether you think the gains of the last 25 years or so, often characterized as the “education reform era,” were all that big to begin with.
Reform has been less reformist than many of us would have liked. It hasn’t accomplished what we wanted.
On the latest NAEP release, the thing that I thought was most important is that we really expanded the gaps in scoring. The top end of the scoring distribution did okay in social studies, but the lower-performing students fell further. What we’re seeing just reinforced these gaps, the spread-out distribution. At the bottom end, it seems that some eighth graders don’t even know what the separate branches of government do; they barely know the difference between the legislature and the executive. That’s very worrisome to me because the U.S. has thrived by having a common society. We might be losing some of that, and the polarization we see in our politics can only be heightened by these kinds of results.
The expanding achievement gap, which we’ve seen on other NAEP tests, is really concerning. Forty percent of eighth-graders scored below the NAEP Basic level in U.S. history, for example.
Yeah, and NAEP Basic is really basic. On subject-matter tests like these, kids scoring below that level really don’t know much at all; on math and reading tests, they’re kind of hopeless because they are not likely to be able to go on past eighth-grade knowledge. In the math test, it’s also roughly 40 percent of eighth-graders in the country scoring below Basic. That’s not a position you want to be in.
There’s been an enormous amount of newer research from scholars like Jesse Rothstein and Kirabo Jackson on the effects of education funding on school outcomes, with much of it suggesting that money really does matter. As perhaps the single figure most associated with the opposing view, are you getting ready to wave the white flag?
[Laughs] I thought the debate was settled, but it’s come back. I spent the last year and a half reviewing, in great detail, all the studies that people cite for this claim. But those people largely choose the studies they like based on their answer to this question. It turns out that there is huge heterogeneity in what these studies find about the importance of spending money on education.
You get very different effects across these studies, and it’s not well understood when money has a big effect and when it doesn’t. There’s no reason to infer from some of this recent work that we can be assured of a great achievement gain, or a lifetime earnings gain, by simply putting more money into the existing system.
At a certain level, this is largely a political debate because Kirabo Jackson and Jesse Rothstein recognize that how you spend money is also really important. My view is that how you spend money is considerably more important than how much you spend. We’re currently spending something like $17,000 per kid on education, and an extra $300 is not going to make much difference unless you spend that money well.
I’m reminded of a paper from a few years back that argued that the positive effects of school finance reforms are mostly driven by states that also adopted test-based accountability.
Yeah, it was very simple and straightforward that if you put more money into systems with good accountability for outcomes, you get much larger results. Part of the problem is that the current federal accountability law, ESSA, is loosening accountability all over the place and particularly for teachers. What states have done since 2016 is backing off holding teachers responsible for student outcomes. This has negative effects on learning.
The effects of spending have been argued endlessly over the last few decades. What are the biggest questions you’d like to see investigated going forward?
One of the developments coming out of the pandemic is that schooling is going to start becoming very different. It’s clear to me that just doing what’s called homeschooling is not going to be the answer, but parents have often looked for something different than what they got during the pandemic and even before the pandemic. You have lots of political fighting over education savings accounts and choice.
I don’t think any of these things is going to be the answer, but understanding the possible choices is important because choices are going to expand for lots of people. We don’t understand where they’re going to go and what the effects of those choices are going to be.
If states continue to expand school choice, does it become harder to improve policy through the districts? It seems like the lever of traditional public schooling will just move fewer students in the future.
Well, if we had Milton Friedman-style vouchers everywhere, we’d still see 80 percent of kids enrolled in traditional public schools. I like having more choice, and I think it’s important. It’s important to improve the charter sector as well. But the traditional public schools are still going to be there for the rest of my lifetime, and probably yours too. So we still have to be concerned about how we operate traditional public schools because they’re the backbone of our economy.