The Maturation Tax

In FY26 the Indian hospital sector stopped being one trade. Every operator grew. Only some earned. The dividing line was not brand, geography or case mix. It was whether a company was harvesting mature beds or paying for new ones.

Operators harvesting mature assets expanded margin. Fortis lifted consolidated operating EBITDA margin to 22.8% from 20.4% and grew PAT 31.5%. Max held 26.8% at the network level. Operators absorbing ramp costs went the other way. KIMS saw Q4 margin fall to 19.9% from 25.3%, with FY26 PAT down to ₹242cr from ₹415cr. Medanta's reported margin fell 150bps purely on one hospital commissioning. Narayana's Q1 FY27 revenue rose 78% while PAT rose 5%.

That dispersion is not a quality signal. It is an accounting artefact of timing. A new tertiary bed carries consultant retainers, nursing, depreciation and interest from day one, against occupancy that starts near zero and reaches steady state in 24 to 36 months. Section UNT quantifies this maturation tax from company disclosure, and it is large enough to explain most of the spread.

The forward question follows directly. Roughly 15,000 self built beds are coming over FY27 to FY30, increasingly greenfield, increasingly debt funded, and concentrated in the same eight or nine metros. Whether that capital earns depends on two things the sector does not control: how fast beds fill, and what the payer agrees to pay for them.

India runs roughly 15 to 16 hospital beds per 10,000 people against a global median near 29. That gap has been quoted for two decades without translating into private returns, because the binding constraint was never supply. It was ability to pay. What changed over the last five years is the payment mechanism, not the disease burden.

Three funding channels now sit underneath private tertiary demand and they are not equivalent. Which channel fills an incremental bed is the single best predictor of what that bed realises.

The insured pool is compounding faster than hospital revenue, which is why volume growth has been easy to come by. The composition is the problem, not the quantum.

The international patient book is the highest margin volume in the sector and is now material: Fortis at ₹639cr, up 18.5% and 7.8% of hospital revenue; Max at roughly 9% of hospital revenue; Medanta's FY26 international revenue up 33%. It is also the most fragile. Medanta flagged sequential softness from regional flight disruption, and Aster saw UAE and Oman volumes fall on Gulf macro. Treat it as a margin kicker with geopolitical beta, not a growth pillar.

Reported EBITDA margins across this sector are currently not comparable, for two separate reasons. One is definitional, one is cyclical. Both are correctable from disclosure, and correcting them changes the read materially.

Operators publish average revenue per occupied bed on different bases. Fortis states it annualised, at ₹2.51cr per annum, derived from hospital segment revenue. Max states it daily, at ₹77,900, on gross network revenue excluding Max Lab, over total occupied bed days. Medanta states it daily, at ₹66,687, on consolidated income including OPD pharmacy.

Apollo has pushed analysts toward average revenue per patient instead, at ₹1.72 lakh, on the argument that ARPOB is contaminated by falling length of stay from robotics and minimally invasive procedures. Shorter stays mechanically inflate ARPOB with no pricing gain at all. That objection is correct and underappreciated: a chain shifting toward day care will show ARPOB growth that is pure denominator effect.

A commissioned bed is a full cost bed from day one and a revenue bed only once it has patients, consultants and live insurance empanelment. Companies disclose that drag with reasonable candour. The market prices it as though it were structural margin decay.

That conditional carries the whole thesis. KIMS attributed part of its Q4 shortfall specifically to delayed insurance empanelment at new units. Empanelment lag is the most underrated operational variable in the sector, nobody discloses it on a standardised basis, and it sits entirely in the payer's gift. Section RSK sets out where this normalisation framework breaks.

Every major operator is expanding at once. Individually each programme is defensible. Collectively they represent the largest simultaneous private bed addition in the sector's history, concentrated in the same handful of metros.

Two structural points follow. First, the capex mix is shifting toward greenfield, which ramps considerably worse than the brownfield additions that drove FY22 to FY25 returns. Brownfield beds attach to existing consultants, existing empanelment and existing overhead. Greenfield beds attach to none of these. Medanta's Noida unit achieved NABH accreditation inside six months and still ran an EBITDA loss for the year.

Second, the balance sheet is now funding growth in a sector that spent a decade deleveraging. Fortis net debt rose to ₹2,334cr, taking net debt to EBITDA to 1.09x from 0.93x. Narayana's net debt to equity moved to 0.49x with US$117mn and £150mn of foreign currency borrowing after the UK acquisition, and it is seeking a ₹1,500cr enabling resolution. KIMS is running a ₹1,500cr QIP explicitly to retire roughly ₹1,000cr of debt. Global Health's Q4 interest cost rose 78% year on year. Jupiter's hit a record.

The listed set splits into three tiers that behave quite differently. Scale integrators (Apollo, Aster) carry non hospital businesses that dilute headline margin but diversify earnings. Premium metro operators (Max, Fortis, Medanta, Jupiter) run the highest realisation per bed and the tightest catchments. Volume and specialty operators (Narayana, KIMS, Rainbow, Yatharth) trade realisation for throughput or clinical focus.

Read the table against Exhibit 3 and the pattern is clean. The three operators with the weakest earnings prints (KIMS, Narayana, Medanta) are the three carrying the heaviest ramp or acquisition load. The two with the strongest margins (Max, Fortis) added capacity brownfield. Nothing in the FY26 numbers suggests a demand problem anywhere in the sector.

Two structural observations on competition. Realisation per bed differs by roughly 17% across the premium tier once bases are normalised, which is narrower than the reported spread implies and suggests limited pricing differentiation between metro operators. And catchment overlap is rising: Bengaluru, Delhi NCR, Hyderabad and Mumbai now each have three or more listed operators commissioning capacity simultaneously, which is where occupancy dilution will show first.

The consensus risk is government price control. We think that is the wrong risk to size. The Clinical Establishments Act 2010 does empower rate setting, but the provision faces multiple constitutional challenges in the Supreme Court, state high courts have struck down analogous rate fixing (the Bombay High Court on Maharashtra's COVID caps, the Calcutta High Court on the West Bengal commission's orders), and the Court's own threat to impose CGHS rates as an interim measure has recurred since 2024 without execution. CGHS rates sit materially below private tariffs. Imposition would be existential, which is precisely why it keeps not happening.

The risk that is live sits one level down, in bilateral tariff negotiation. In August 2025 the AHPI, representing some 15,200 hospitals, instructed members to suspend cashless services for Bajaj Allianz and Care Health policyholders over refusal to revise tariffs against 7% to 8% medical inflation. It was withdrawn within a week. The precedent matters more than the outcome.

Four transactions in twelve months have changed the sector's structure. The common logic is worth stating: acquired hospitals arrive with occupancy, consultants and live empanelment already in place, which means no maturation tax. That is the direct alternative to the build programmes in Section CAP, and the sector is now running both experiments in parallel.

The Aster and Quality Care merger is the largest of the four and the only one where the arithmetic can be tested from public disclosure, so it is worth one exhibit. Subtracting Aster's standalone FY26 from the published pro forma gives an implied Quality Care margin of 23.0% against Aster's own 20.4%. Quality Care was ascribed 25.2x FY24 adjusted EV to EBITDA in the scheme documents against 36.6x for Aster. On those figures Aster acquired a higher margin business at a lower multiple, in paper.

Share count rose 68.2% and pro forma EBITDA rose 112.6%, so EBITDA per share moves from ₹18.28 to ₹23.09, roughly 26% accretion before any synergy. Set against that: the two networks barely overlap geographically, so cost synergy should be assumed low; four brands are retained across 28 cities; and Blackstone holds 29.71%, a larger stake than the founding Moopen family's 24%, with three of twelve board seats. The deal was well priced. The company is not yet demonstrated. First post merger print is 5 August 2026.

The stance framework is Harvesting, Building, Digesting: a position on the capacity cycle rather than a valuation call. Ratings without a fitted model would be decoration. Section APX sets out the modelling approach for when the company models are built.

Scenarios turn on two independent variables: the pace at which the FY27 to FY30 bed cohort reaches steady state occupancy, and whether ARPOB growth holds above medical cost inflation. Revenue CAGR is sector aggregate for the listed set. Margin is the maturity normalised blended figure from Section UNT.

Probabilities are analyst judgement, not a fitted distribution. The base case is deliberately unexciting: the sector earns its way through the build, but slowly, and multiples compress before earnings catch up. A poor setup for momentum, a reasonable one for accumulation in names where the ramp cost already sits inside the reported base.

Consolidated margin is a lagging indicator in this sector. The variables below move first and are all disclosed quarterly.

Target prices require a fitted model and are deliberately absent here. For the company notes that follow, the approach is a bed cohort build: separate mature units from units commissioned within 36 months, model each cohort on its own occupancy ramp and ARPOB path, and let the blend fall out rather than forecasting a consolidated margin directly. KIMS's disclosed 29.5% mature unit margin and Medanta's 32.4% developing hospital margin excluding Noida are the two cleanest terminal anchors in the sector. Valuation on EV to EBITDA and EV per operating bed, cross checked against a DCF with WACC built from the current risk free rate and an equity risk premium appropriate to the regulatory tail.

Primary sources: company Q4 FY26 and Q1 FY27 results filings, investor presentations and earnings call transcripts via NSE and BSE; the Aster and Quality Care scheme of amalgamation with the NCLT Hyderabad order dated 19 June 2026 and subsequent allotment filings; IRDAI sourced industry premium data via press; Union Budget FY27 health allocations; MoHFW and IBEF sector statistics; broker result updates where cited for derived metrics. All figures are as reported by companies unless marked ^e for derived. No restatement has been applied.