← Nerdy Dragon
Nerdy Dragon
Do the Math. Follow the Money.

Do the Math. Follow the Money. #1
America spends 7x more on school than in 1976. Math scores haven't moved.

America spends 7x more on school than in 1976. Math scores haven't moved. The receipts.

🐉 ₿ ∑

America spends 7x more on school than in 1976. Math scores haven't moved.

First in a series on the education system, its incentives, and its receipts. Every number below is sourced.

In 1975-76, America's public schools spent $1,504 per student (NCES Digest, Table 182). In 2023-24, they spent $17,360 (NCES Digest 2023, Table 236.15).

That is more than 7 times the money. Even after adjusting for inflation, real spending per student has roughly doubled in about 50 years.

Now the second number. In 1978, the average American 13-year-old scored 264 on the federal government's longest-running math test, the NAEP Long-Term Trend assessment. In 2023, that same test (same scale, same measurement) put the average 13-year-old at 271 (NAEP LTT 2023).

Seven points in forty-five years. On a 500-point scale, that is a rounding error.

Worse: the score peaked at 285 in 2012. A decade later we gave back 14 of the 21 points we had gained since the early 1970s. The 2023 class of 13-year-olds is doing math at a level their great-grandparents' kids basically matched in 1978. NCES's own headline: "Scores decline again for 13-year-old students in reading and mathematics." The White-Black score gap widened from 35 points to 42 in just three years.

So follow the money. Where did roughly $15,000 of added annual spending per student, several trillion cumulative dollars, actually go?

Not to results. The receipts above say so.

Partly to smaller classes. In 1970, American public schools had 22.3 students per teacher. By 2021 it was 15.4 (NCES Fast Facts). That is a 31% reduction: the single most expensive "improvement" in education, and it shows up nowhere in the score trend.

Partly to credentials. In 1971, 27% of teachers held a master's degree or higher. By 1991 it was 52.6%, more than doubled (NCES Digest, Table 68, the last survey that tracked it back that far). Today it is about half. Master's degrees come with pay bumps ("steps and lanes," in the trade's language). They come with tuition bills, too.

And the tuition bills are where the money system walks in. A teacher in 1976 earned $12,005 (Digest Table 68). In 2020-21 the average full-time public teacher earned $61,600 (Digest 2022, Table 211.20). Adjust 1976 dollars into 2021 dollars ($57,177) and teacher pay has grown 7.7% in 45 years. Half a decade of work for essentially nothing, while the credential required to keep the job climbed from "bachelor's" to "bachelor's, increasingly a master's, paid for with debt."

Look at what else that 1971 to 1991 teacher survey says. In 1971, 69.6% of teachers held only a bachelor's degree. In 1976, the median teacher was 33 years old and the median classroom had 25 kids. In 1981, asked if they would "certainly would" teach again, 21.8% said yes, down from 49.9% in 1961.

Here is the incentive chain, and you can check every link:

  1. Federal policy made student debt cheap and everywhere (Higher Education Act expansion, 1972 to 1992).
  2. Schools tied pay raises to credentials, not results ("lanes").
  3. Universities, facing customers whose bills were pre-financed by the federal government, raised tuition. The customer stopped being price-sensitive because the price was someone else's future problem.
  4. Teachers took on debt to get the credentials that unlock the pay bump. The salary that was supposed to pay it off is, in real terms, flat.
  5. The system now needs teachers to keep taking degrees, and teachers need the degrees to afford the degrees. That is not a conspiracy. It is a ratchet.

    6. Nobody in this chain has to be evil for it to grind exactly this way. Each participant responds rationally to the incentive in front of them. The system just happens to pay every party for doing more of the thing that does not show up in a 13-year-old's math score.

    The Department of Education opened its doors in May 1980. It has never taught a child. What it did do is standardize the flow of federal money into a system that already had all the incentives it needed. Enrollment grew about 10% since the mid-1970s. Spending, after inflation, doubled. Scores: +7 points, then a 14-point fall from the peak.

    The people hurt worst are the ones the system claims to serve. Lower-performing 13-year-olds fell 12-14 points in math between 2020 and 2023, more than twice the decline of top performers. The kids who needed the money most got the least out of it.

    The fix is not more money. We have run that experiment for fifty years at a cost of trillions, and the control group is sitting in a 1978 exam booklet.

    The fix is to change who the money answers to. When funding follows the child instead of the system; when teachers are paid for craft and results instead of credit hours; when a parent can take their kid's education budget somewhere that actually teaches math, the incentives flip, and the ratchet runs the other way.

    That is not a fantasy. That is exactly what we are building at Math Success: research-backed math instruction, printable materials and direct instruction that go straight to teachers and parents, no curriculum-industrial middlemen, with results demonstrated in real classrooms. We will show you how in upcoming posts.

    But first, the receipts. This series will follow the money through every layer of the system: the publisher consolidation, the assessment industry, the federal loan machine, the credentialing ratchet. Every claim numbered, every number sourced.

    Do the math. Follow the money.

    Sources

    Corrections welcomed, with sources. Every number here traces to a public dataset.

🐉 Do the math. Follow the money. · nerdydragon.com