International School M&A: Verified Transactions, Prices and EBITDA Multiples (2021 to 2026)

How every figure is graded

School transaction data is dominated by undisclosed prices and repeated estimates. This directory separates what is actually known from what is merely repeated.

Where a figure in circulation could not be re-verified against its claimed source, it was downgraded or removed. Two previously circulated multiples were withdrawn during verification; each entry says so.

Global platform transactions

The deals that set the top of the market, and the disclosure behind each headline number.

Most Southern European exposure sits inside these platforms rather than in separately priced local deals: Globeducate is Spain-headquartered, ISP holds Spanish and Portuguese schools, and Inspired holds assets in Spain, Italy, Portugal and Greece. Individual school prices in those markets surface only where a listed seller must disclose, as with the H-FARM sale to Nord Anglia set out under Continental Europe. Higher education deals such as Permira and Universidad Europea are excluded from this directory as adjacent rather than K-12.

GCC transactions

The most transparent regional market, because listed operators and REITs must disclose. Where privately held deals dominate, headline values stay private.

Three consecutive Amanat monetisations now sit in this table with published returns: the Taaleem stake at 2.2x and a 21 per cent IRR in 2021, the school property at 1.7x and 10 per cent in 2025, and the Almasar listing at 2.2x cash-on-cash later the same year. One investor, three exits, three disclosed outcomes. Nothing else in this directory offers that. The property return is the lowest of the three and should be read that way: a building let to a school earned roughly half the money multiple of the operating stakes over a comparable period, which is the arithmetic behind every argument for splitting property from operations rather than owning both.

UK and Ireland

Record deal volume driven by the VAT change, the sector’s clearest lesson in checking what a price actually includes, and a growing distressed channel where schools transfer through administration rather than a competitive sale.

Alpha Plus: the OpCo/PropCo case study

Inspired Education acquired 17 Alpha Plus schools, including Wetherby School and Pembridge Hall, at an enterprise value of over GBP 230m Published per Delancey’s own disclosure. The sale explicitly excluded the three pre-university Colleges and several freehold properties, which remain with Delancey client funds. Shortly before the sale, the group’s London freehold portfolio was valued at GBP 145m and the group was reported as loss-making at holding level Strong estimate.

An EBITDA figure of GBP 30.1m circulates from aggregator databases only and sits uneasily against the reported losses Reported. GSE does not publish a multiple for this transaction; an implied figure previously in circulation was withdrawn on verification.

Other UK and Ireland activity

Continental Europe

Prices surface where a listed seller is obliged to disclose. Where the seller is private, Continental European school deals close without a published number, which is why the platform transactions above carry most of the region’s valuation evidence.

Greek trade press estimates these school acquisitions at EUR 15m to EUR 30m each depending on enrolment, campus and results Reported. The range is included because its circulation makes omission misleading, and it should not be used as a comparable: no party has disclosed a price on any of the six Greek transactions above.

Asia, Americas and Africa

Including the regulatory event that repriced more K-12 equity value than any transaction in this directory.

Historical benchmark transactions

Pre-window deals with disclosed values, kept separate because their analytical value is the re-rating arc they reveal, not comparability with current pricing.

Education real estate yields

Where school property trades separately from the operating business, published yields exist. They anchor every OpCo/PropCo valuation split.

Occupancy cost is the figure operators actually negotiate against. KYS KL East is the only entry where rent, designed capacity and price are all published: MYR 8.2m of net rent across 1,300 designed places is MYR 6,308 per place per year, about USD 1,560, before a single lesson is taught. Escalation compounds harder than it reads. Ten per cent every three years is 3.23% a year, so rent in the final period of a thirty-year initial term reaches MYR 7.78 per sq ft per month, more than double the entry level, and the operator carries all of it.

The widely quoted benchmark of school rent at 8% to 12% of revenue is a market rule of thumb. No published source verifies it, and GSE labels it accordingly. The wider yield evidence is set out in the GSE comparison of cap rates and yield in education real estate.

The listed comparables

Nine listed vehicles give the sector a continuous public mark. They are the only school valuations that reprice daily, which makes them the natural sanity check on any private multiple.

Curro also supplies the sector’s clearest published affordability signal. Its expected credit loss provision rose to 42% of gross receivables in FY2024 from 32%, with total bad-debt costs at 4.3% of turnover against 3.3% the prior year, and enrolment fell 1.4% at the start of 2025. Fee-paying schools carry consumer credit risk that rarely appears in a transaction multiple, and Curro is one of the few operators anywhere that publishes it. Any underwriting model built on enrolment growth should be tested against these disclosures.

One name to watch is absent from this table for a reason. Grupo Salta Educação, Brazil’s largest K-12 operator, holds listed-company registration on the B3 and is bound by the same governance rules as a public company, but has not completed an IPO and its shares do not trade. It is a registered issuer awaiting a window, not a listed comparable, and it is recorded under Asia, Americas and Africa with its private transaction evidence instead.

How GSE treats multiples for these names. Market data providers disagree on EV/EBITDA and P/E for listed school operators depending on lease treatment, minority interests and the earnings period used, so GSE does not republish provider figures. Where a company publishes enough to derive a multiple, GSE derives it and shows the working. Curro is currently the only name in this table with all the required inputs published: an enterprise value of roughly ZAR 10.8bn to 11.1bn, being a market capitalisation of about ZAR 7.6bn plus net debt of about ZAR 3.2bn to 3.5bn, against FY2024 EBITDA of ZAR 1.2bn, gives roughly 9x GSE derived. Market capitalisation is a moving input and published figures for Curro range between about ZAR 5bn and ZAR 7.6bn depending on the date and source, so the derivation is an order of magnitude rather than a mark. The remaining names are derived once their results are published, on the same basis as the private transactions above.

Enterprise value per student

The metric practitioners reach for when EBITDA is undisclosed, which in this sector is most of the time. It is also the fastest way to see that a single global multiple range does not exist.

Three caveats govern this table. Student counts are as reported at announcement and are not audited, so they move with the enrolment cycle. Enterprise values include IFRS 16 lease liabilities where the party reported on that basis, which inflates both the EV and the per-student figure against a freehold-owning peer. The USD column applies indicative rates at announcement rather than a fixed date, so it is directional and should not be used for precise cross-deal arithmetic. The Curro row is a listed mark rather than a transaction: enterprise value is derived by GSE from market capitalisation of about ZAR 7.6bn plus net debt of about ZAR 3.5bn at the reference date, and published market capitalisation figures for Curro vary by source between roughly ZAR 5bn and ZAR 7.6bn depending on the date, so the derived figure should be read as an order of magnitude. GSE publishes this as a screening tool, not a valuation.

Operating benchmarks, fee levels and leverage

What sits underneath a multiple. Prices are the visible layer; the cost structure, the fee level and the debt capacity are what determine whether a price was sensible.

Revenue per student, and why it beats EBITDA for screening

Revenue is disclosed far more often than earnings, so revenue per student is available where a multiple is not. Paired with enterprise value per student it also yields an implied EV to revenue multiple, which is the most widely available comparison in this sector.

The global fee ladder

Fee level is the variable that explains most of the valuation spread in this directory, so it is worth setting out at its full range. These are published or reported annual fees, not derived figures, and they are not adjusted for discounting.

Cost structure

Few operators publish a cost breakdown. Curro is the useful exception, and its disclosures give the clearest published anchor for a school P&L. Staff costs run at roughly 70% of operating costs, described elsewhere in its reporting as about two thirds of operating expenses Published. Head office costs have been held at 6% to 7% of revenue over five years. Revenue is roughly 92% school fees and 8% ancillary, with ancillary income led by boarding at about a quarter, after-care at about a sixth and transport at about a seventh. Around 13.6% of learners receive a discount of some form, covering staff children, bursaries and affordability concessions.

Wendel’s Globeducate disclosure adds a second set of party-published operating benchmarks, this time at the premium end: roughly 35% of the school portfolio is owned freehold with the balance on long-term leases, maintenance capex runs near 2% of revenue, the FY25e EBITDA margin is about 27% including IFRS 16 and 22% excluding it, tuition is largely prepaid before each term giving negative working capital, and average student tenure is 8.2 years. The same documents record Providence’s hold from 2017 to 2024 taking the group from 21 schools and about 11,000 students in four countries to 65 schools and more than 40,000 students in eleven, through 21 bolt-on acquisitions, the most completely documented buy-and-build in this directory.

BDA’s July 2025 snapshot of listed operators, on Capital IQ data and a pre-IFRS 16 basis, adds a seven-name margin ladder: SISB in Thailand at 43.9% LTM EBITDA margin, NCLE at 35.6%, Taaleem at 26.4%, Overseas Education at 24.4%, Curro at 23.1%, Ataa at 22.8% and Al Faleh at 19.2%. SISB sits well above the 20 to 35% band usually quoted for well-run schools; a point-in-time LTM margin on a smaller listed name can flatter, and it is recorded as an outlier to interrogate rather than a benchmark to underwrite. The same snapshot marks the low end of the listed spectrum: Overseas Education at 5.3x LTM EBITDA with earnings in decline, and Al Faleh at 9.7x. A comparables set that keeps only the interesting names overstates the sector.

CIRA in Egypt reports on a different basis again, with an adjusted EBITDA margin of about 56% in an earlier reporting period and around 41% more recently Strong estimate, against Curro nearer the mid-twenties. Before treating Egyptian margins as evidence of superior economics, note that they are struck in a currency that has devalued heavily against the dollar over the same period, so local-currency margin expansion and hard-currency return can move in opposite directions. CIRA also publishes two operating metrics almost nobody else does: K-12 capacity utilisation, which stood at about 91% in the period reported, and a pupil-to-teacher ratio of about 11.3. Utilisation is the single most useful operating number in a school model, because it tells you whether growth requires capital or simply fills existing seats.

Two implications for underwriting. First, a school is a payroll business with a property attached, so a model that flexes revenue without flexing teacher cost at the learner-to-teacher ratio is not a model. Second, the discount rate matters as much as the headline fee: a published fee schedule overstates realised revenue per student wherever bursary and affordability concessions are material, and Curro is the only operator in this directory that quantifies it.

Leverage

The GEMS line requires care and is the reason GSE publishes the input rather than a leverage ratio. The USD 3.25bn facility refinanced existing debt and funded minority shareholder exits including CVC, so the quantum is not net debt struck against operating earnings in the way the Curro and Salta figures are. What it does show is the scale of debt a GCC school platform could raise against roughly USD 380m of estimated earnings, which is the more useful fact for anyone sizing a facility.

The multiples framework

Read together, the verified evidence supports the following working ranges. Scaled premium global platforms transact in the high teens to low twenties on EBITDA, with Globeducate’s published 16.7x to 19.4x pair as the cleanest anchor and Nord Anglia and ISP’s headline values consistent with the top of that band on estimated earnings. Individual schools and established groups have historically transacted at 8x to 14x, with premium single assets and regional platforms in supply-constrained markets now underwritten at 12x to 17x. Mid-market schools sit in the high single digits to low teens, nurseries at roughly 3x to 7x, and SEN commands a premium supported by structural demand. No Indian multiple is carried in these ranges. The 13x widely reported on the Pathways transaction is not verified here: the reported earnings imply an operating margin close to 70%, and the price appears to capitalise the Gurgaon site alongside the operating business. The frameworks behind these ranges are set out in the GSE guide to how to value an international school.

Two adjustments must precede any use of these ranges. Establish the IFRS 16 treatment of both EV and EBITDA, and establish whether the price includes freehold property. A freehold-included deal attaches a 6% to 9% property yield play to an operating multiple, and will always print higher than a leasehold operator of the same quality. The structural logic is explained in the GSE guide to the PropCo/OpCo model in school development.

One further disclosed multiple is worth recording even though it falls outside school operations. In February 2021 Vasta Platform (Nasdaq: VSTA) agreed to acquire Editora Eleva, a K-12 learning-systems provider, from Eleva Educação for BRL 580m payable over five years, and stated in its own release that this corresponded to 16.6x Editora Eleva’s FY2020 EBITDA Published. Curriculum and education-services businesses are adjacent to school operations and are excluded from this directory’s ranges, but a party-stated multiple is rare enough in this sector to note.

The same caution applies to franchise systems. Reuters sources put Primrose Schools at close to USD 2bn against roughly USD 85m of EBITDA on about USD 120m of revenue, an EBITDA margin near 70% and a target above 20x. Those are franchisor royalty economics rather than school operating economics, and the two are not comparable: an operator carries the teaching payroll, the property and the enrolment risk that a franchisor does not. Franchise multiples in the low twenties should never be cited as evidence for what an operating school is worth.

Brazil now supplies the directory’s clearest mid-market anchor. Grupo Salta traded at more than 10x EBITDA on a 32% minority in February 2024 and at an enterprise value of about BRL 5.8bn in November 2025. Annualising the reported nine-month EBITDA of BRL 517m gives roughly BRL 690m, which puts the 2025 trade near 8.4x GSE derived. The annualisation is a GSE assumption, not a disclosed figure, and Brazilian school groups do not earn evenly across the year. Read against the high-teens-to-low-twenties platform range, the gap is the price of scale, premium positioning and hard-currency earnings rather than any difference in education quality.

Frequently asked questions

The open dataset

The full transaction dataset is free to download, reuse and cite under CC BY 4.0. Attribution: Global Services in Education (GSE), International School M&A Transaction Dataset. Current version v31, compiled 1 September 2026, 106 transactions. Version v30 is archived at Zenodo with the permanent identifier 10.5281/zenodo.22168906, which resolves to that exact release. The identifier 10.5281/zenodo.22168905 always resolves to the latest deposited release. Every row carries its version, compile date and this page’s address, so any downloaded copy identifies itself and points back to the current file.

The dataset is the record of account. A few table rows describe more than one dataset entry where the transactions were announced together, so the number of visible rows is slightly lower than the entry count. The CSV always carries them separately.

Dataset versions. The current file is v31, compiled 1 September 2026, covering 106 transactions. Earlier numbered files remain reachable so that anything already cited keeps working, but they are superseded and were compiled against fewer transactions. If you are holding a copy, check the dataset_version and compiled columns against the figures above before relying on it.

For the interpretive companion, the GSE School Valuation Briefing covering underwriting ranges, OpCo/PropCo structuring and negotiation benchmarks, contact GSE to request a copy.

About the author. Greg Parry is Co-Founder and CEO of Global Services in Education (GSE), which develops and operates international schools worldwide. He co-founded the firm in 2011 with Shanna Parry, Co-Founder and Senior Managing Partner. GSE has delivered more than 60 school and education projects across 28 countries, working with investors, developers, governments and operators from feasibility through to opening and ongoing management.

Entries added, version 1.2

BDA Partners, a sector investment bank, published a Southeast Asia K-12 report in August 2025 whose transaction comparables put scaled platforms at 14.8x to 18.1x LTM EBITDA and single schools at 10.6x to 12.0x, on 40th to 60th percentiles. That is a second independent professional source arriving at the same scale-tier structure this directory argues from, alongside Wendel’s through-the-cycle chart, and the same report attributes 19.0x to the 2021 OMERS stake in International Schools Partnership. Two qualifications travel with all of it: BDA is a sell-side advisor marketing the sector it is describing, and its figures are sourced to MergerMarket and Capital IQ, so every number taken from it is graded Reported. The report cites Global Services in Education among its sources. It also records a third policy repricing alongside China 2021 and the UK VAT: Malaysia introduced a 6% sales tax in June 2025 on annual fees above MYR 60,000 for non-Malaysian students.

Eleven entries follow from this material and from Cognita’s announced Chile acquisition.

Entries added, version 1.3

One entry added on 19 August 2026. It carries no price, no stake and no multiple, so it changes none of the ranges in this directory. It is recorded because the transaction evidences a shift in the size of school that institutional buyers will now transact, and because the vendor side is unusually well documented for a private European deal.

Methodology, corrections and disclaimer

Every figure is traced to its earliest identifiable source; a number repeated across twenty outlets counts as one source, not twenty. Figures claiming party disclosure are checked against the primary document itself: the filing, release or prospectus. Figures that fail the check are downgraded or removed, and the entry records the change. Two multiples in circulation have been withdrawn on verification, on Alpha Plus and on NCLE. Structural facts, including OpCo/PropCo treatment, IFRS 16 basis and which properties transferred, are recorded wherever determinable, because they change what a headline number means.

What this directory does not yet cover. Coverage is strongest where disclosure is compelled: listed sellers, rated debt, exchange filings and REIT reporting. It is correspondingly thin where private and family ownership dominates. Known gaps, stated so that no reader mistakes silence for absence: Australian K-12 beyond the listed early-years sector; Indian transactions beyond those recorded above, where reported values are plentiful and party confirmations are rare; Swiss transactions, noting that the established boarding schools are family held and do not trade, so this gap is likely structural rather than temporary; Turkish transactions beyond the two recorded above; Japan and Korea, where no priced K-12 transaction surfaced and where school ownership sits largely with non-profit educational corporations, a structural explanation GSE has not yet verified against the statutes; Egyptian transactions beyond the CIRA disclosures; Vietnamese transactions beyond the EQuest investment and the ACG schools, with the Nguyen Hoang process unresolved; Poland and central Europe, checked without result in this cycle; South African transactions beyond the ADvTECH and Curro disclosures; and Chinese K-12 beyond the regulatory event recorded above.

What this directory records. The scope is fee-paying K-12 schools and school groups, together with special educational needs provision, school property where it trades separately from the operating business, and listed operators carried as comparables. Early-years and childcare businesses are recorded where an operator in scope bought them, because the same buyers are active in both. Universities and tertiary education are excluded as adjacent, which is why Permira and Universidad Europea do not appear and why the AcadeMedia acquisition of IVA Business School was checked and not added. Curriculum and learning-systems businesses are excluded on the same basis and appear in prose only.

From v27 every row in the dataset carries a segment value, so the file can be filtered rather than read as a single population. The 106 entries divide as 77 K-12 fee-paying, 14 school property, 4 early years, 3 special educational needs, 3 listed references, 3 mixed platforms spanning school and either early-years or tertiary operations, and 2 publicly funded voucher operators. Early-years businesses are recorded only where a school operator was the buyer, which is why the AcadeMedia preschool acquisitions and Kids First Group appear and why childcare bought by childcare groups does not. The headline count is therefore closer to a count of school transactions than it was, but it is still not identical to one: school property and listed references are counted too.

The directory deliberately excludes three things: estimated prices for undisclosed deals, valuations attributed to unnamed advisory sources, and any multiple where either the value or the earnings figure fails verification.

The dataset is licensed CC BY 4.0. Cite as: Global Services in Education (GSE), International School M&A Transactions Directory, dataset v30, Zenodo, https://doi.org/10.5281/zenodo.22168906, with the access date. When quoting a figure, quote its grade with it. GSE welcomes corrections supported by documentation via the GSE contact page; corrections are acknowledged to the submitter and reflected in the record.

Two limitations. A Published grade means the figure traces to a party release, filing or exchange announcement. For the larger transactions GSE has opened the source document; for some smaller entries the grade rests on the primary source being named and quoted in reliable reporting rather than on GSE having retrieved the filing itself. Where that distinction matters to an entry, the entry says so. Separately, the dataset records values in the currency of the transaction with no normalised column, because a single conversion date across fifteen years and nine currencies would create a false precision; anyone modelling from the CSV should apply their own rates at their own reference date.

This directory is provided for information only. It is not investment advice, and GSE accepts no liability for decisions made in reliance on it. Figures marked as estimates or reported should be independently verified before use in any valuation or investment decision.