By Thomas Ultican 11/18/2025
A propaganda rag, ‘The 74’, reported, “[R]esearch from Stanford estimates learning loss over the past decade has cost our country over $90 trillion in future growth.” The article was written by Eric Hanushek, a Stanford University economist, and Christy Hovanetz, an education researcher from Jeb Bush’s pro-school choice and pro-education technology organization, ExcelinEd. Unsurprisingly, the article linked above is a paper Hanushek wrote. Crazy-pants Eric has a long history of using his own papers to support new research which is typically long on assertions and short on convincing analysis.
Last year, he claimed COVID-19 “learning-loss” could cost America $31 trillion in future economic development. Hanushek’s latest paper asserts, “The present value of future lost growth would be approximately three times current GDP (which is $30 trillion).” The justification for this new assertion is more than uncertain. The new paper refers to a model from his 2011 paper written with Ludger Woessmann for the Organization of Economic Co-operation and Development. That paper was long on Arithmetic with several assumptions short on common sense. In his latest paper and this article, he is claiming that a small drop in the National Assessment of Education Progress (NAEP) scores added to what he calls “Covid learning loss” will cost America $90 trillion in future growth.
Eric Hanushek first gained notoriety with his 1981 paper, claiming “there is no relationship between expenditures and the achievement of students and that such traditional remedies as reducing class sizes or hiring better trained teachers are unlikely to improve matters.” This claim attracted conservative billionaires but had little relationship with reality. When providing solutions in ‘The 74’, Hanushek and Hovenetz contradict his 1981 paper writing:
“First, states need to invest in effective personnel. They can do this by incentivizing strong teaching and by supporting strong teaching through professional development in evidence-based practices such as use of high-quality instructional materials and assessment data to inform instruction.”
A few decades ago, a friend gifted me the book “An Incomplete Education” by Judy Jones and William Wilson. After my 1987 edition, Jones and Wilson have updated the book and republished several times. I fondly remembered their description of economists:
“Economists are fond of saying, with Thomas Carlyle, that economics is ‘the dismal science.’ As with much that economists say, this statement is half true. It is dismal.”
“Where once rulers relied on oracles to predict the future, today they use economists. Virtually every elected official, every political candidate, has a favorite economist to forecast economic benefits pinned to that official or candidate’s views.” (Incomplete Page 120)
It just so-happens that Eric Hanushek is an MIT trained economist who is good at creating reports that conform to the beliefs of conservative billionaires. His work is scientific propaganda masquerading as academic excellence.
Notices from Hanushek and Hovanetz
“No single event over the postwar period has had an impact on our educational system that comes close to that of the pandemic.” This statement from their article is hyperbole not fact.
The impact of vouchers in Washington DC and Ohio were worse than COVID-19. In his book, “The Privateers”, Josh Cowen shared that the losses due to COVID-19 were around -0.25 standard deviations while losses in DC due to vouchers were around -0.40 standard deviations and in Ohio they were as high as -0.50 standard deviations. (Privateers Page 6) However, these were not nationwide results. Unfortunately, we don’t have much information about the mumps, measles, flu and polio epidemics of the 1950s but there is every reason to believe their impact was close to that of the COVID-19 epidemic.
The authors hold up Mississippi, Tennessee and Louisiana as examples to emulate stating:
“Some states made noteworthy progress on NAEP this year: Mississippi, Louisiana and Tennessee. Each has a track record of high expectations and strong accountability.
“These states use an A–F school rating system that puts reading and math achievement front and center. They measure what matters — proficiency and growth — and they report results in a way families and educators can understand. Transparency and rigor are fueling their progress.”
The A—F school ratings system is worthless because it only measures parent income. As Diane Ravitch wrote on her blog:
“The highest rated schools have students with the highest income. The lowest rated schools have students with the lowest income.”
I wondered what Hanushek and friend were alluding to, so I graphed the average NAEP data for 4th and 8th grade reading and math since 2015 to see if it provided a clue.
NAEP Data for Louisiana, Mississippi and Tennessee
I added a bar indicating the NAEP cut scores for the 2024 NAEP designations of basic, proficient and advanced as a reference. It is well known that NAEP’s proficiency level is set well above grade level which means 70% of students not being rated proficient is not a bad score. Furthermore, writing in Forbes, Peter Greene shared, “An NCES report back in 2007 showed that while NAEP considers “basic” students not college ready, 50% of those basic students had gone on to earn a degree.”
The data graphed here indicates to me that all three states have solid public education programs. It is noteworthy that in reading their 4th graders only averaged basic but by 8th grade their average results were above basic. However, all three states are among the bottom 15 scoring states in the US and as we would expect are in the bottom 15 states in family income. What I do not see is why they are held out by Hanushek and Hovanetz as exemplars.
Concluding Remarks
The first blogger to label Eric Hanushek crazy-pants was Peter Greene. I have found Peter’s creativity worth stealing.
Let’s be clear, there is no such thing as “learning loss” involving students. This is a complete misnomer. Students may not achieve some state provided targets but they are always learning; maybe not what we want them to learn but always learning.
Eric Hanushek is a huckster for conservative billionaires. This latest claim is beyond farcical. Claiming to have knowledge of economic development in 40 years is fantasy and then tying that fantasy to NAEP testing data is bizarre. His $90 trillion claim for loss of future economic development would embarrass a carnival barker.
Hanushek is the guy who thinks education can be measured in days of learning which means that students learn in a near linear fashion. They absolutely do not but even this bogus claim would seem prescient when compared to this latest assertion.
Standardized testing only correlates with family income. To evaluate education based on this fictitious instrument is fraudulent. By extension, this makes economist Eric Hanushek a con artist.
It’s hard to imagine that Diane Ravitch was once friends with Hanushek.
None so blind as those who will not see.
The most troubling problem here is the genetic fallacy. You spend a remarkable amount of the piece telling us who Eric Hanushek supposedly serves — “conservative billionaires” — and calling him a propagandist, huckster and finally a con artist. But none of that establishes that his economic model is wrong. It could be spectacularly wrong. The way to demonstrate that is to identify the assumptions, data or causal inference that fail and show us why. A man’s politics, associations or presumed patrons cannot do that work for you.
And once I noticed that, several other logical problems became difficult to overlook.
You say there is “no such thing as learning loss” because students “are always learning.” But that equivocates on the word learning. Nobody using “learning loss” as a term of art means children literally stopped acquiring any knowledge or experience. The claim is that measured academic achievement or progress fell relative to an expected trajectory. You can dispute whether that construct is useful, but redefining it into an obviously absurd claim and knocking that down is essentially a straw man.
Then you answer a claim about the nationwide educational impact of COVID by pointing to effect sizes from particular voucher programs in DC and Ohio — before acknowledging yourself that “these were not nationwide results.” That qualification identifies the comparability problem rather than solving it. A localized policy effect and a nationwide shock are not interchangeable merely because both can be expressed in standard deviations.
There is a similar leap at the end: “Standardized testing only correlates with family income,” therefore the instrument is “fictitious,” its use is “fraudulent,” and Hanushek is a “con artist.” Even granting a strong correlation between test scores and family income, the conclusion simply does not follow. A measure’s correlation with socioeconomic status does not demonstrate that it measures nothing else. To establish that, you would have to address validity, reliability and the evidence concerning what the assessment actually measures.
There are smaller problems as well. Hanushek’s decades-old finding about aggregate education expenditures is presented as somehow contradicted by his present support for investing in effective teachers and evidence-based instruction. But “more spending does not automatically produce higher achievement” and “some particular expenditures can improve instruction” are perfectly compatible propositions. And the observation that poorer states tend to score lower does not by itself answer whether Louisiana, Mississippi and Tennessee have made unusual gains over time; levels and changes are different questions.
What bothers me about all this is that you may have a perfectly legitimate argument against Hanushek’s $90 trillion estimate. Forty-year macroeconomic projections obviously depend on assumptions, and those assumptions deserve aggressive scrutiny. But that is precisely why the logical shortcuts are so unnecessary.
Attack the model. Show which assumptions fail. Show where causation has been inferred from correlation. Show why the extrapolation is unreasonable. That would be an interesting argument.
Calling the economist “crazy-pants,” associating him with billionaires, redefining his terminology and eventually declaring him a con artist isn’t a substitute for making it.