Why Has John Keells Holdings Stayed Near Rs.20 Despite Record Earnings?

John Keells Holdings (JKH.N0000) closed at Rs.20.00 on 9 July 2026 — within the same Rs.14–24 nominal band it has occupied, on a subdivision-adjusted basis, for essentially the entire past decade. Over the same period, Group EBITDA has grown from roughly Rs.27bn (FY17) to Rs.80.01bn (FY26, +75% YoY), three flagship projects — City of Dreams Sri Lanka, the West Container Terminal, and BYD/John Keells CG Auto — have all begun contributing profit, and the broader market has re-rated hard: the ASPI is up roughly +256% since December 2019. JKH's own adjusted share price is up only +19.3% (price only) over the identical window.

Having tested twelve specific hypotheses against the CSV trading record, the FY2025/26 Annual Report's audited financial statements, CSE shareholder-register disclosures, and current market data, we find the flat share price is not primarily a story of hidden selling, nefarious crossings, or market neglect of a cheap stock. It is better explained by four things happening at once:

1. A genuine nominal illusion from the July 2024 rights issue and 1:10 subdivision inflated the share count roughly 11-fold, resetting the "look" of the price without changing the underlying enterprise value.
2. Cash conversion has only just turned positive. FY26 free cash flow (audited operating cash flow less capex) was a thin +Rs.6.2bn against Rs.80.0bn of EBITDA — a ~8% conversion rate. The market has been pricing the balance sheet reality (capex-funded, still deleveraging) more than the income-statement headline.
3. The re-rating the bulls are waiting for hasn't been withheld — it's been given, then taken back. JKH's EV/EBITDA multiple compressed from 11.4x (FY24) to 6.2x (FY26) even as EBITDA nearly doubled, which is a market judgment that FY26's earnings — flattered by BYD's opening-year pent-up demand and CODSL's first positive year — are not yet a trusted run-rate.
4. Foreign capital has been a net, if modest, seller (non-resident ownership 44%–42% YoY) inside a country whose sovereign rating is still CCC+ and whose currency has lost roughly a third of its value against JKH's own share price in USD terms since 2019 — a macro overhang no single project can offset.

Crossing transactions are real and material (a confirmed 10 million share crossing was recorded on 7 July 2026, and our volume-anomaly screen flags at least seven multi-hundred-million-share days since 2024 with near-zero price impact) — but the evidence is consistent with block ownership transfers and portfolio rebalancing, not a single suppressive actor. We name what the data shows and label what remains unidentified, per CSE public disclosure.

Ranked by estimated explanatory weight, based on the quantitative and disclosure evidence in this report.

Every event below is drawn from CSE filings, the CSE's company-announcement portal, or the FY2025/26 Annual Report. Where a buyer/seller or exact figure was not disclosed publicly, it is marked undisclosed.

Yes — substantially. JKH's issued share count rose from 1,502,635,957 (pre-rights, Jul 2024) to 16,528,995,520 (post-subdivision) to 17,699,275,116 (31 Mar 2026, after ESOP exercises) — an ~11.8x increase in roughly 20 months. The subdivision alone (1 share → 10) is purely mechanical: it cannot change market capitalisation, only the per-share numerator.

What the annual report itself confirms

JKH's own 10-year summary table (restated for the subdivision) shows the share has traded in a remarkably narrow adjusted band for a full decade: Rs.13.79 (FY17) → Rs.15.96 (FY18) → Rs.15.60 (FY19) → Rs.11.54 (FY20, COVID) → Rs.14.85 (FY21) → Rs.14.50 (FY22, sovereign default) → Rs.14.00 (FY23) → Rs.19.40 (FY24) → Rs.20.20 (FY25) → Rs.18.40 (FY26 year-end). This is not a new phenomenon created by the 2024 corporate actions — JKH has been range-bound on an adjusted basis since at least 2016. The subdivision changed the sticker price from "Rs.194" to "Rs.19.40"; it did not change the multi-year trading range.

Dividend mechanics compounded the confusion

FY2024/25's declared dividend was reported as "Rs.0.50 (final) + Rs.0.10 (interim) = Rs.0.60," but a footnote in the financial statements clarifies the Rs.0.50 final dividend was declared prior to the subdivision — i.e., on the old, 10x-smaller share base. On a like-for-like post-subdivision basis, that Rs.0.50 is equivalent to Rs.0.05. Investors comparing headline dividend-per-share figures across FY24–FY26 without adjusting for this are working from Rs./share amounts that aren't comparable.

Important scope limitation up front: we do not have access to the CSE's official daily crossing-board report for the full 30-month period requested. What follows combines (a) one publicly confirmed, named-and-dated crossing from CSE market commentary, and (b) a volume-anomaly screen we built directly from the trading data — flagging days where volume spiked >5x its trailing 60-day average while the price barely moved, which is the statistical signature of a negotiated block trade rather than order-book price discovery. We label the screen's output as candidates, not confirmed crossings, because we cannot verify counterparties or crossing prices against the official board without that data source.

Confirmed crossing

Candidate high-volume / low-price-impact days (screen, since Jan 2024)

Methodology: candidate = daily volume > 5x trailing 60-trading-day average AND |same-day return| < 1.5%. Screened on subdivision-adjusted volume/price from JKH daily trading data. The 23 Dec 2025 observation — a large block at an exactly unchanged closing price — is the single strongest statistical signature of a negotiated crossing in the entire dataset.

Does this explain why high turnover hasn't moved the price?

Partially, yes. A crossing by definition executes at a negotiated price and does not consume order-book liquidity the way accumulation or distribution would — so a day with Rs.11.5bn of turnover (6 Feb 2024) moving the price only 0.4% is not itself evidence that "the market" rejected upward pressure; it is evidence that a large block changed hands off the order book. Genuine price discovery on JKH's ordinary board volume (excluding these candidate crossing days) shows a materially higher Amihud illiquidity ratio in the earlier years of our sample — see Section 07 — implying that when JKH does trade on pure order-book flow, it moves further per rupee of turnover than the blended average suggests.

Named, disclosed movement among the top 20 shareholders (31 Mar 2025 → 31 Mar 2026)

Source: JKH FY2025/26 Annual Report, "Top Twenty Shareholders" and "Composition of Shareholders" tables (pp.159–160). These reflect the shareholder register, not intraday crossing counterparties — the two data sources are complementary but not identical.

Illiquidity as measured by the Amihud ratio has fallen roughly 13-fold since 2020 — the market absorbs far more turnover per unit of price impact today than it did before the subdivision. This is consistent with one of the subdivision's stated aims (a lower nominal price broadens the practical buyer base) actually working as intended on the liquidity dimension, even though it did nothing for the price level itself.

The cash-flow reality behind the EBITDA headline

Returns and leverage

ROE has structurally declined over the decade — from 11.1% (FY18) to a low of 1.4% (FY25) before recovering to 3.4% (FY26) — even as book value per share has risen every single year (Rs.10.1 → Rs.23.4). JKH transformed from a net-cash conglomerate (FY17: -14.9% net debt/equity, i.e. net cash) into a moderately levered one (FY26: +31.2%) funding CODSL, WCT, and BYD inventory almost entirely from its own balance sheet and the 2024 rights issue.

City of Dreams Sri Lanka (CODSL)

Opened in stages: Cinnamon Life hotel (Oct 2024), casino + Nüwa hotel (2 Aug 2025). WPL (the project company) carries a USD 189m term loan (USD150m long-term + USD39m six-month bridge), roughly 60% of which (~USD90m) matures in March 2031. The rupee's depreciation against this loan cost JKH Rs.1.83bn in net exchange losses within recurring PBT this year alone — a direct, mechanical channel through which currency weakness taxes the very project meant to diversify JKH's earnings base. Gaming tax was raised from 15% to 18% and the citizen entry fee doubled to US$100 in October 2025, both headwinds to the ramp. Management describes performance as "steadily improving" quarter to quarter, with the casino recognising fixed rental income only — the variable rental component (the larger long-run economic prize) activates once operations reach a specified activity threshold, which has not yet been disclosed as met.

West Container Terminal (Colombo West International Terminal / CWIT, Phase 1)

CWIT delivered the clearest unambiguous good-news story in this portfolio: full utilisation of phase-1 capacity within its first year, positive PAT despite phase-1 depreciation, and construction "marginally ahead of schedule." Phase 2 capacity is expected to be capitalised from Q3 FY2026/27; gate-automation equipment has been imported with installation due Q2 FY2026/27. Transportation segment recurring EBITDA grew 32% on this and continued bunkering margin strength.

John Keells CG Auto (JKCG) / BYD

JKCG is the single largest driver of this year's EBITDA growth and the segment most exposed to the "is this repeatable?" question. Vehicle imports were only liberalised from early 2025, meaning FY26 captures a genuine pent-up-demand surge on top of BYD's first-mover positioning. A ~37% new-vehicle-registration share is an extraordinary number for a brand-new entrant, but 20,000 completed vehicle services within the first year (evidence of a real, growing installed base rather than pure order-book hype) and continued showroom/aftersales expansion (Ampara, Rathnapura) and brand extension (the DENZA premium marque) suggest the position has some durability. The open question we cannot resolve from public data: how much of the ~37% share and the margin per vehicle survives once (a) competing NEV/ICE brands respond, (b) the initial liberalisation-driven order backlog clears, and (c) import-duty or EV-incentive policy is revisited.

Key threads

• Currency: the rupee has lost roughly 46% of its US-dollar value since December 2019, with the bulk of that move concentrated in the 2022 sovereign default (LKR/USD jumped from ~200 to ~305 within the fiscal year). JKH's own USD-denominated financial statements show FY26 Group revenue of US$1.68bn and diluted EPS of just US$0.0024 — a reminder of how large the currency translation effect is on any USD-based valuation comparison.
• Sovereign credit: Sri Lanka exited default in December 2024 after restructuring, and Fitch/S&P now both rate it CCC+ (stable). This is real progress, but CCC+ remains deep speculative grade — Fitch's own commentary projects general government debt/GDP near 90% by 2028 and an interest/revenue ratio of ~42%, still far above the CCC-rating median of 16%. This is the backdrop against which foreign institutional investors are deciding whether to return to CSE equities at all, JKH included.
• Interest rates & inflation: inflation was 5.4% YoY in April 2026, near the CBSL's medium-term target, but the central bank raised its policy rate by 100bp in May 2026 on renewed currency and imported-inflation pressure tied to Middle East oil-price pass-through.
• Market-wide sentiment: the ASPI itself pulled back roughly 8–15% from its February 2026 highs (~23,750) to current levels (~21,800) amid escalating Middle East tensions — a broad risk-off that has hit Sri Lankan equities disproportionately hard versus other regional/global indices, per market commentary.
• Tourism & consumption: Cyclone Ditwah disrupted Q3 FY26 (cold-chain, some Leisure activity), but underlying occupancy, supermarket footfall (+14.3%), and outbound-linked travel initiatives (a new JKH-backed Colombo–Ahmedabad flight route) point to a broadly improving consumer and tourism backdrop, independent of JKH's own project-specific execution.

*As at 31 Mar 2026 (Rs.18.40/share); current market cap at Rs.20.00/share and 17.70bn shares is ~Rs.354bn. †FY24 figure not comparable — see reconciliation note in Section 04.

Sum-of-the-parts — directional, not a target price

A full SOTP requires segment-level discount rates and terminal assumptions beyond what public disclosure supports precisely; we present the FY26 recurring EBITDA base by segment (Section 09) as the starting point any reader building their own SOTP should use, with the explicit caveats that (a) CODSL and JKCG both carry first-year-of-scale EBITDA that may not be sustainable at current levels, (b) Financial Services (NTB, Union Assurance) should be valued on an equity-accounted, look-through basis rather than consolidated EBITDA given the group only holds a minority/associate interest in NTB, and (c) any SOTP must deduct WPL's USD189m and the IFC's USD136m holding-company-linked debt before arriving at equity value per share.

These are scenario sketches built from the disclosed FY26 base and stated management assumptions — not price targets, and not a recommendation. Treat the "illustrative value" as a rough EV/EBITDA-multiple sensitivity, not a discounted cash flow output.

Applying these illustrative multiples to a stabilised EBITDA base (which itself is uncertain — see Section 09) rather than the current Rs.80bn figure directly would be double-counting the ramp-up assumption; readers building their own model should form an independent view of "sustainable" EBITDA before applying any multiple here.

Data limitations

• We do not have direct access to the CSE's official daily crossing-board reports for the full 30-month period requested. Section 05's "candidate crossing days" are a statistical screen (volume anomaly + low price impact), not confirmed crossings, except for the single 7 Jul 2026 transaction independently reported by EconomyNext.
• ASPI/S&P SL20 benchmarking (Section 07) uses annual year-end index levels, not a daily series matched to JKH's own trading calendar, because a full daily ASPI history was not available to us. The 2025 year-end point is omitted (not available), and the 2026 comparison point is the latest available level (9 Jul 2026), not a full-year close.
• USD/LKR conversion (Section 11) uses period-spot rates for illustration, not a full daily FX series matched to JKH's own price series.
• Segment-level ROCE, capital-employed, and BYD per-vehicle margin figures are not fully disclosed at the granularity the original brief requested; we report what management has disclosed and flag where a figure is a derived approximation.
• The FY26 vs FY25 financing-activities discrepancy between the Annual Report's MD&A narrative (Rs.45.85bn outflow) and its own audited cash flow statement (Rs.5.88bn outflow) is unresolved in the source document; we used the audited statement throughout.

Source list

• JKH FY2025/26 Annual Report (612 pages) — audited financial statements, MD&A, share information, shareholder register, ESOP disclosures
• JKH daily trading data, 30 Dec 2019 – 9 Jul 2026 (Date, OHLC, trade count, share volume, turnover) — user-supplied CSV, cross-checked for corporate-action breaks
• CSE company announcement portal (cdn.cse.lk) — Rights Issue circular (30 Jul 2024), Sub-division circular (30 Jul 2024), subscription-level letter (11 Oct 2024), various corporate disclosures (2025–2026)
• EconomyNext — daily market reports (ASPI/S&P SL20 levels, crossing report 7 Jul 2026, USD/LKR spot rates, CBSL reserve/policy data)
• Wikipedia, "All Share Price Index" — historical annual ASPI closing levels, 1993–2024
• Melco Resorts & Entertainment, TTG Asia, igamingbusiness.com, Asia Gaming Brief — City of Dreams Sri Lanka opening, investment size, EBITDA disclosure
• John Keells Group / LMD / EconomyNext / Daily Mirror — BYD/JKCG market share, unit sales, aftersales network
• Fitch Ratings, The Business Standard, TradingEconomics — Sri Lanka sovereign credit rating history
• Exchange-rates.org — historical USD/LKR reference rates

Technical appendix — formulas used

• Subdivision adjustment: Price(t < 5-Nov-2024) ÷ 10; Volume(t < 5-Nov-2024) × 10
• Amihud illiquidity: mean( |daily return| ÷ (Turnover in Rs.bn) ) per calendar year
• Annualised volatility: stdev(log returns) × √252
• Candidate crossing screen: Volume > 5 × rolling-60-day mean Volume, AND |return| < 1.5%
• Free cash flow: Net cash from operating activities − (Purchase of PPE + Purchase of intangible assets), per audited Statement of Cash Flows
• USD-adjusted return: (Price(t2)/FX(t2)) ÷ (Price(t1)/FX(t1)) − 1