Reports of drops in student achievement due to the pandemic are now treated as old news. Amid abstract reporting of test results, a sense of inevitability and complacency has developed. After all, could the fact that students’ math scores fell by “nine points” truly be important?
The reality is that the cohort of students in school in March 2020 has been seriously harmed—implicitly facing a lifetime tax on earnings of 6 percent. And the harm is not going away.
A simple way to assess learning loss from the pandemic is to compare the performance of students tested in 2023 to students taking the same tests in 2020. The most recent data come from the National Assessment of Educational Progress for 13-year-olds. Often called the Nation’s Report Card, NAEP provides regular assessments of American students’ math and reading skills at different ages. Comparing 2023 results with those for students tested just prior to the pandemic reveals that losses averaged nine points in math and four points in reading. This drop erased all the gains in students’ math scores since 1990 and moved reading scores back to where they were in 1975! Low-achieving students lost more than high achievers, poor students lost more than nonpoor students, and both Black and Hispanic students lost more than white students.
But NAEP, like most tests, uses an arbitrary scale to report scores that makes the size of changes hard to interpret. The implications of lost learning are better seen by translating these sterile numbers into economic losses. Past research confirms that people who know more, as measured by their performance on tests like NAEP, earn more. The research considers how individuals’ earnings throughout their working lives differ according to the skills measured by scores on standardized math and reading tests. Importantly, the U.S. labor market rewards these cognitive skills more than almost all developed countries—which in turn implies that the U.S. punishes the lack of these skills more than almost all developed countries.
Historical earnings patterns make it is possible to estimate what the learning losses documented by NAEP will cost the average student in the Covid-cohort: 6 percent lower lifetime earnings than those not in this cohort. In other words, the pandemic learning losses for this cohort are equivalent on average to a 6 percent income tax surcharge throughout the students’ working lives. This rises to 8 percent for the average Black student, who suffered greater learning losses according to NAEP.
The economic costs do not end there. The economies of nations with more skilled populations grow faster in the long run, and the pandemic learning losses imply that the U.S. population will be less skilled in the future than it would have been. Using historical growth patterns, it is again possible to project the aggregate losses to the U.S. economy of having this lower-skilled cohort move through the labor force. The economic loss from the lower-skilled workforce amounts in present value terms to $28 trillion.
Costs in trillions of dollars are perhaps no easier to understand than drops in test scores. To put this figure in perspective, consider that the projected loss of $28 trillion amounts to more than one year’s Gross Domestic Product. Or that the aggregate losses due to unemployment, business closures, and related economic fallout from the pandemic totaled about $2 trillion. The losses from the “Great Recession” in 2008 totaled about $5 trillion. In short, the impact on the economy we should expect from pandemic-era learning loss dwarfs the impacts that have so captured public and policymakers’ attention in recent years.
We are struggling as a nation even to get our schools back to where they were in terms of supporting student learning, but these costs will be permanent if we just return schools to the status quo in March 2020. Our schools must improve if we are going to eliminate the burden of lost learning. Evidence from a variety of experiences in other nations shows that the losses students experienced will persist if schools simply return to business as usual. For example, several German states had short school years in the 1960s when policymakers sought to standardize school calendars nationwide. The earnings of students educated during that period stand out throughout their careers from those of students educated before and after the adjustment, and not in a good way. Other examples of extended school disruptions—for example, due to prolonged teacher strikes—show similarly persistent impacts.
What has been done so far to address learning loss? The federal government provided almost $190 billion in Covid relief aid to schools under three separate appropriations. Only a small portion, however, was required to be spent on ameliorating learning loss, and most schools have yet to spend much of these funds even though they disappear in a year.
States and districts have adopted a variety of strategies that most frequently include added instructional time or intensive tutoring. Unfortunately, the results of these efforts to date have not been good. Even if we optimistically project that the best available programs will be implemented with fidelity, the losses will not be erased. The scale of current recovery efforts is simply insufficient to overcome the deficits. Moreover, when recovery programs are voluntary, as is typically the case, higher-achieving students are more likely to participate, leading to a widening of achievement gaps.
At the same time, the pandemic strengthened a number of harmful policy trends that may cause school quality to decline. For one, it reinforced a general drift away from test-based accountability policies. Additionally, teacher unions saw the pandemic as an opportunity to push a variety of their preferred policies—including policies well beyond pay, benefits, or anything related to learning. For example, the Oakland Teachers Association, after agreeing to a substantial pay and benefits hike, nonetheless went on an eight-day strike in May 2023 over “common good” clauses, including reparations for Black students and “environmental justice.”
There is a clear roadmap to success, albeit one that leads to political tension. The one policy that is known from research to be effective is ensuring that all students have an effective teacher. Recruiting and retaining more effective teachers has, of course, been the goal of many policy initiatives, but a variant of this emphasis can be the solution to the learning loss problem: Simply provide incentives for the most effective current teachers to teach more students. The highly-effective teachers could teach larger classes or added sections of courses with both monetary incentives and additional support for this work. Unused federal funds could immediately support this tactic. Indeed, one could go further and use part of the funds to buy out the contracts of the least effective teachers. These steps could instantly improve the average effectiveness of instruction, both making up for pandemic-era learning losses and improving schools going forward.
Such policies have been shown to work in a few large districts, including Washington, D.C., and Dallas. Deploying them now at scale could save the Covid-cohort from a 6 percent lifetime tax. The alternative, saying change is “too hard,” amounts to accepting the lifetime injury to current students along with a $28 trillion national loss.
Eric Hanushek is a senior fellow at the Hoover Institution of Stanford University and a prior member of the National Assessment Governing Board.