ḤeshbonCenter on Data and Jewish Life
Jewish EducationCost & Scale
Jewish Education ← Library
Original dataset · ~497 PK–8 schools · multi-year panel

Growth stops saving most schools money at about 150 students. Does it stop saving these?

Independent schools get cheaper per student as they grow, up to about 150 students, when the savings run out. So we took a hand-built panel of roughly 497 PK–8 schools and drew the cost curve for the Jewish day schools next to the curve for ordinary private schools of the same size in the same regions.

Two lines below, on the same pair of axes. Left to right is enrollment, from a small school to a large one. Up and down is cost per student, so lower on the chart is cheaper to run. The slate line is the comparison schools. The rust line is the Jewish day schools. Follow each line from left to right and watch where it changes direction.

0schools compared
0where others stop saving
0where these keep saving
0analyses that agree
Scroll to follow the curve
How this dataset was built

An original dataset assembled by hand from public IRS Form 990 filings, historical and archived filings, Internet Archive / Wayback captures, and school financial records and public web sources, reconstructed across multiple years. Because the record follows the same schools year over year, it can watch a budget move as enrollment changes, which is what the year-over-year design in this story rests on.

IRS Form 990 filings Historical & archived filings Internet Archive / Wayback School financial records
~497schools in the panel
4independent analyses that agree
PK–8grade range
Multi-yearlongitudinal panel

Every finding on this page rests on the assembled record above.

The short version

Start with why a small school is expensive.

Why is a small school expensive in the first place?

The reason is easiest to see one classroom at a time.

You have to staff third grade whether it holds eight children or twenty. The teacher, the classroom, and the schedule cost about the same either way, so every additional student spreads that fixed cost across more people and the price per student falls.

The saving runs only so far. Once a classroom fills, the next student needs a second section and a second teacher, and the gains stop. Across independent schools generally, the turn arrives somewhere around 150 to 200 students, the oldest and most replicated finding in the economics of school size. Past that point, cost per student levels off and then begins climbing again.

A single-curriculum school

One track per grade to fill.

One set of classrooms and one faculty. The rooms fill quickly, and the savings from growth are used up early.

A dual-curriculum school

Two tracks per grade to fill.

General studies and Judaic studies run side by side in every grade. Twice as much fixed staffing waits to be spread, so there is far more room to keep spreading it.

01 · Where the extra money goes

Running two programs means hiring more people.

If a school runs two curricula, what does the second one cost?

The typical comparison school employs about 21 staff for every 100 students. The typical Jewish day school employs about 24, roughly five more employees in a school of 200. Comparing schools of the same size in the same part of the country, Jewish day schools spend about 17% more per pupil, and that gap is solidly established.

Who the comparison schools are

The comparison group is a set of ordinary independent, community, and parochial private schools, matched to the Jewish day schools on the things that actually drive cost: enrollment and size, operating budget, campus and facilities, grade range (PK–8), and region and local cost of living. These are everyday private day schools, not elite, high-endowment flagship prep academies. Matching on the cost drivers is what keeps the comparison like-for-like rather than Jewish schools measured against fancy ones.

Every comparison on this page holds school size and region constant, so the two groups differ in how they are built, not in scale or geography.

0
staff per 100 students at comparison schools, at the median
0
staff per 100 students at Jewish day schools, at the median
0
higher cost per pupil, $27,682 versus $32,497 at the median

Two bars below. Each one is the number of full-time-equivalent employees a school carries for every 100 students, at the median, running rightward from zero, so a longer bar means more adults on the payroll per child. The slate bar is the comparison schools. The rust bar is the Jewish day schools.

Look at where the two bars end, and at how little separates them.

Staff per 100 students, median school
Full-time-equivalent employees across all categories. Comparison group is comparable-size private schools in the same regions.
The slate bar stops at 21.3. The rust bar under it runs to 23.8, two and a half staff further along the same scale. That small a gap is doing all the work, because class sizes are essentially the same across both groups and so is faculty experience (class size differs by 0.7 students, t = −0.44; experience by about a year, t = 1.59). The extra money buys more sections rather than smaller ones, which is the whole reason the next part follows. The cost figures behind the premium are $27,682 versus $32,497 at the median, with p < 0.001 in the fitted model, meaning a gap this size would almost never turn up if the two kinds of school really cost the same to run.
A bigger fixed base costs more to carry. It also leaves far more room to keep diluting as the school grows.
02 · A curve with a different shape

More fixed cost to spread means more room to keep spreading it.

Does that bigger fixed base show up in the shape of the cost curve?

A dual-curriculum school starts out more expensive because it carries a bigger fixed staffing base. That same base is also more forgiving of growth, since there is more of it to dilute and it takes far longer to run out of room. The prediction is about shape: the cost curve should fall more steeply and keep falling at sizes where other schools have already flattened.

The chart below is the pair of axes from the hero drawn large. Enrollment runs across the bottom, from 100 students to 600, on a scale that stretches the small schools out. Cost per student runs up the side, with gridlines labelled $25k, $30k, $35k and $40k, so lower on the chart is cheaper per child. The five buttons build the argument a step at a time. Comparison schools draws the slate curve on its own, with a dot at the point labelled "savings run out". Both sectors lays the rust curve for Jewish day schools over it. At 400 students drops a dashed vertical line at that enrollment and puts a dot where each curve crosses it. Staffing channel and Year to year clear the curves away and replace them with two smaller panels, one of bars and one of a number line. Hovering anywhere on the curve reads out cost per student at that enrollment.

Watch what the slate curve does after the dot.

Cost per student as schools grow
Comparison schools (comparable-size private schools, same regions) in slate, Jewish day schools in rust.
Hover or tap the curve to read the cost per student at a given enrollment.
Start on Comparison schools and follow the slate curve down to the dot near 150 students, then up again for the rest of the chart: at 400 students, 10% more enrollment predicts about 2.7% higher cost per pupil. Press Both sectors and the rust curve comes in above the slate one on the left and crosses under it further right. Press At 400 students and read the two dots on the dashed line: +2.7% for the slate curve, −2.2% for the rust one, close to $700 a year on a $32,000 per-pupil budget. The rust slope also steepens as the school gets bigger, from −0.11 at 250 students to −0.22 at 400 and −0.32 at 600, so growth is running the wrong way for one group and the right way for the other.

The savings show up where the theory predicts. Press Staffing channel and the panel becomes two bars measuring what happens to a school's staffing as it grows from 150 students to 400. Jewish day schools shed 15.9% of their staff per 100 students, the long rust bar. Comparison schools shed only 4.2%, the short slate one. Pay per employee shows no difference between the groups, so the saving comes from filling two sets of classrooms rather than from paying anyone less.

Press Year to year and the panel becomes a single horizontal line with a mark at each end. The left mark is −1.0, labelled "budget stays fixed", where a school's spending does not move at all when enrollment changes. The right mark is 0, labelled "budget follows enrollment", where spending rises in exact step with the number of children. The two dots on that line are the two sectors, each school compared with itself twelve months earlier. The rust dot sits at −0.92, almost against the fixed end. The slate dot sits out at −0.53, about halfway. A nearly fixed budget is what heavy fixed staffing looks like, and it means each new student arrives close to pure margin.

0
Jewish day school budget elasticity, where −1 is a fully fixed budget
0
comparison school budget elasticity, same year-over-year design
0
Jewish day school cost slope at 600 students, still falling
03 · What it means at real sizes

The paths separate exactly where day schools live.

At the sizes real day schools actually are, how much does the difference come to?

Under about 150 students, both kinds of school are still filling classrooms, and the case for growth is obvious. Between 150 and 250, comparison schools cross over and growth starts costing them money, while Jewish day schools are only beginning to see their savings arrive. At 400 students and up, the gap turns strategic: a 400-student Jewish day school weighing whether to add a section faces genuinely different math than the independent school down the road, and the standard sector advice, built on schools like that one, points the wrong way.

Five pairs of columns below, one pair for each school size: 100, 150, 250, 400 and 600 students. The horizontal line across the middle is zero, marked "no change". A column standing above that line means 10% more enrollment leaves the school costlier per student. A column hanging below it means the same growth leaves the school cheaper. The slate column in each pair is the comparison schools and the rust column beside it is the Jewish day schools, with the figure printed at the far end of every column.

Follow the two colours left to right and watch them pull apart.

Effect of 10% enrollment growth on cost per student, by school size
Bars above the line mean growth raises cost per student; bars below mean it lowers cost.
At the 100-student pair the two columns are short and on opposite sides of the line, −1.0% for slate and +1.0% for rust. At 150 they are almost nothing, +0.1% and 0.0%. From there the pair opens out at every step: +1.4% against −1.1% at 250 students, +2.7% against −2.2% at 400, and +3.8% against −3.2% at 600, the slate columns climbing higher above the line and the rust ones hanging further below it.
What follows from this

If the curve holds, enrollment growth works as a financial strategy, not only a mission one.

What would a school do differently if this is true?

Day school leaders are usually told, correctly, that they cannot grow their way out of a budget problem. That advice comes from research on schools with one curriculum. The structure of a dual-curriculum school appears to change the arithmetic.

01

Recruitment spending reads as an investment

If each additional student lowers cost per pupil rather than raising it, admissions and marketing budgets stop being pure overhead and start carrying a measurable return.

02

Tuition-reduction programs may partly pay for themselves

Discounting to fill seats looks like lost revenue on a single-curriculum cost curve. On this one, added enrollment lowers unit cost, which changes how a middle-income program pencils out.

03

Merger and consolidation math shifts

Two 200-student schools that combine cut duplicate overhead, and they also land in the size range where the savings run steepest.

One caution belongs alongside all of this. The shape of the curve is supported by every analysis the paper runs, yet it rests on the couple dozen Jewish day schools that report financial data, and the author is careful to call it a well-supported hypothesis rather than a settled fact. Roughly forty more schools reporting would likely settle it for good.

The bottom line

Nothing in this data finds the floor.

So: growth stops saving most schools money at about 150 students. Does it stop saving these?

Not anywhere the record can see. The comparison schools reach their cheapest point near 150 students and get more expensive from there. The Jewish day schools are still getting cheaper at 600, and the slope is steeper there, at −0.32, than it was at 400 or at 250.

Here is the strongest version of the objection. The 17% level premium is established at conventional significance, but the shape of the curve is not. It rests on the couple dozen Jewish day schools that report financial data, and the author calls it a well-supported hypothesis rather than a settled fact.

What holds the hypothesis up is that four separate research designs point the same way: the fitted curve, the split at 400 students, the staffing channel, and the year-over-year check of each school against itself. Roughly forty more schools reporting would likely settle the question for good.

What this analysis cannot say yet

Read before quoting a number.

The Jewish sample is small

The curve rests on the couple dozen Jewish day schools that report financial data. Roughly forty more schools reporting would likely settle the question for good.

The shape is a hypothesis, not a verdict

The difference in curve shape is supported by point estimates in all four of the paper's research designs, and it is not statistically confirmed under the paper's preferred inference method. The author presents it as a well-supported hypothesis.

The curves are reconstructed

The cost curves are drawn from the elasticities the paper reports at five reference enrollments and anchored to its reported median expenditures. They reproduce the shape and crossing behavior described in the paper and are not a reproduction of its original figure.

The premium is the firm part

The 17% level premium is established at conventional significance. The difference in curve shape between the sectors is the part awaiting broader financial reporting.

Source & method

A visual reading of "Scale Economies and the Cost of the Dual Curriculum: Evidence from a National Sample of PK–8 Independent Schools" (Gavriel Brown, working draft, July 2026). The underlying dataset was assembled by hand from public IRS Form 990 filings, historical and archived filings, Internet Archive / Wayback captures, and school financial records and public web sources, reconstructed across multiple years, covering roughly 497 PK–8 schools. The Jewish day schools are cross-referenced against a peer set of comparable independent, community, and parochial private schools, matched on enrollment, operating budget, campus and facilities, grade range, and region rather than drawn from elite flagship prep academies. This is a well-supported hypothesis about the shape of the cost curve, not a settled fact.

How to read the numbers. "Slope" is the percent change in cost per student for a 1% change in enrollment; a negative slope means growing makes the school cheaper per student. All comparisons hold school size and region constant, so the comparison schools are ordinary private schools of similar size in the same regions. FTE counts full-time-equivalent employees across all categories.

Statistics shown are drawn from the paper's reported results. The cost curves are reconstructed from the elasticities the paper reports at five reference enrollments and anchored to its reported median expenditures; they reproduce the shape and crossing behavior described in the paper but are not a reproduction of its original figure. The level premium is established at conventional significance; the difference in curve shape is supported by point estimates in all four of the paper's research designs but is not statistically confirmed under the paper's preferred inference method, and the author presents it as a hypothesis awaiting broader financial reporting.

The paper behind this story

Scale Economies and the Cost of the Dual Curriculum

Working paper · 2026 · Ḥeshbon · Center on Data and Jewish Life

First page of the whitepaper Scale Economies and the Cost of the Dual Curriculum

Request the full whitepaper

Tell me who you are and where you work. I’ll send the complete paper.

Sent to the author directly. No list, no forwarding, no third parties.

Ḥeshbon — Center on Data and Jewish Life

A data project on how Jewish life is paid for. The work comes in seasons. This page is part of Season 1 — Jewish Education. It states its own sources, sample sizes and limits.

Ahead: Season 2 — Household Affordability · Season 3 — Federation & Charitable Giving

More in Season 1

 

Working prototype. Every figure comes from the underlying research. Where a chart simplifies a published result, the page says so. Nothing here audits an individual school.