Kasumigaseki Capital (3498)

Signal panel

Investment thesis

1. Cheap on its own history and on growth. PER 7.1x on guidance; PER band mean 6.88x (±0.37x per σ); PBR 1.97x vs band mean 2.38x (fact sheet, Valuation). Peers trade at 7.1–10.5x forecast PER with implied ROE of ~13–16% against Kasumigaseki's 27% (see Valuation). Reported fact / inference.
2. Conservative guidance record. Actual OP vs initial guidance: FY21 +6.2%, FY22 +15.7%, FY23 +38.8%, FY24 +0.4%, FY25 +14.8%, FY26 +5.2% (fact sheet, Guidance). FY26 initial OP ¥26.5bn → actual ¥27.9bn (presentation p.11). Reported fact.
3. Inventory to hit FY27 is in hand. Inventory (販売用不動産 + 開発事業等支出金) ¥71.2bn at 2026-08-31, up from ¥53.3bn; management states it is sufficient for the FY27 plan (presentation p.3, p.12). AUM plus pipeline ¥1.218tn, +84% YoY (p.13). Company guidance.
4. Balance sheet was reset by the November 2025 offering. Equity ¥86.8bn vs ¥36.2bn; equity ratio 39.1% vs 29.7%; management targets ~30%, which leaves ~¥70bn of borrowing headroom (presentation p.8). This is the main argument against an imminent new equity raise. Company guidance.
5. Returns stay high after the dilution. ROE 27.2% in FY26 (32.5% FY25) with leverage cut; FY29 target ≥25%, aim 30% (presentation p.7). Reported fact / company guidance.

Business and earnings

Kasumigaseki Capital is a developer and asset manager. It buys land, plans a project (hotels, cold-chain and automated logistics, hospice housing, overseas residential), and sells the land into a development fund it organises, booking (1) a land-sale gain, then (2) project-management fees, (3) a success fee on sale of the finished asset to a REIT or private fund, and (4) asset-management fees (presentation p.65). It sponsors its own Kasumigaseki Hotel REIT (presentation p.31). The investors are domestic institutions and funds; the company wins on speed of sourcing (16 hotel sites and 6 logistics sites acquired in FY26, 短信 p.4 (printed 2)) and on owning the operating brand (fav, FAV LUX, seven x seven, BASE LAYER), with newer hotels reaching stabilised RevPAR within six months (presentation p.21).

FY2026/08 (短信 p.1, p.7):

• Sales ¥163.8bn (+69.7%), OP ¥27.9bn (+47.2%), NP ¥16.7bn (+62.9%). OP margin fell to 17.0% from 19.6%; gross margin 33.1% vs 37.8% (presentation p.67). SG&A rose ¥8.8bn, ¥3.6bn of it personnel (p.11).
• One-offs roughly cancel: negative goodwill +¥1,478m (Wakayama Marina City acquisition), impairment −¥811m, bad-debt loss −¥668m (短信 p.7). EPS needs no restatement.
• Silent-partnership (匿名組合) investment gains of ¥6,195m sit inside profit, 37% of NP; cash distributions received were ¥5,637m (短信 p.11), so these are mostly cash-backed. Reported fact.
• Earnings are back-end loaded. Q4 alone was ¥19.7bn of ¥27.9bn OP (71%); Q3 was ¥0.1bn. Several fund sales were disclosed on 2026-08-27/28, days before year end (snapshot pages/report_disclose.html). The company manages and guides annually only (短信 p.2). Quarterly results therefore say little; a year-end sale slipping by a week can move a year.
• Overseas: Miami site sold to a domestic-investor fund within 5 months (presentation p.52); Dubai switched to fund-based development, but ¥12,961m of Dubai development cost was reclassified to fixed assets during FY26 (短信 p.13), i.e. it is now held rather than sold. Inference: a Middle East-driven delay that could become an impairment.
• Data centres and automated warehouses are first projects only (presentation p.37). No revenue yet; not part of this thesis.

Forecast test

• FY2027/08 guidance (2026-10-05): sales ¥220.0bn, OP ¥41.0bn, RP ¥36.0bn, NP ¥24.0bn, EPS ¥976.85, DPS ¥200 (短信 p.1–2). Company guidance.
• IFIS consensus (as of 2026-10-16): OP ¥40.6bn (−1.0% vs guidance), RP ¥36.4bn, NP ¥24.4bn; FY2028/08 OP ¥57.5bn, NP ¥34.7bn. OP estimates fell 1.1% (FY27) and 2.1% (FY28) over 3 months; weekly RP consensus cuts of 0.9%, 2.0% and 1.9% after results (pages/report_topix.html). External estimate.
• Track record: 6/6 beats of initial OP guidance, median +10.5%. FY26's beat (+5.2%) was achieved after an upward revision on 2026-09-17. The beat is habitual, so partly expected, but consensus sits below guidance, so the market is not paying for it.
• Non-operating drag is rising. RP guidance is ¥5.0bn below OP vs ¥2.0bn gap in FY26 — interest expense was ¥2.7bn in FY26 (¥1.3bn FY25) and debt is rising into a 10-year JGB above 3%. Guidance already contains this; further rate rises are the main forecast risk.
• Progress: Not assessable. FY27 began 2026-09-01 and the company gives no half-year guide; Q1 to Q3 have historically been small and volatile (Q3 FY26 OP ¥0.1bn).
• Revision probability: Medium for a beat at year end; low for a mid-year upward revision before Q3, given the Q4-loaded pattern.

Financial strength and capital allocation

• Cash flow. OCF −¥16.8bn (FY26), +¥6.9bn (FY25), −¥8.4bn (FY24). The FY26 outflow is inventory build (−¥26.3bn), prepayments (−¥5.7bn) and tax (−¥8.7bn) (短信 p.3, p.11). ICF −¥31.5bn: PP&E −¥12.2bn, loans −¥11.2bn, acquisitions −¥11.6bn (p.12). Funded by equity ¥34.7bn and net long-term borrowing ¥36.8bn.
• Gate check — negative OCF 2 of 3 years: passes with a funded explanation. For a developer, land for sale is inventory and runs through OCF; the build is backed by equity issued in Nov 2025 and term debt, not by asset disposals. It remains a structural weakness: growth is not self-funding.
• Balance sheet. Cash ¥41.8bn; interest-bearing debt ¥108.0bn including the ¥22.0bn CB; net debt ¥66.2bn; net D/E 0.76x; current ratio 2.36; equity ratio 39.1% (fact sheet; 短信 p.5–6). Current portion of long-term debt jumped to ¥33.8bn from ¥13.6bn (p.6) — refinancing depends on bank appetite for real-estate lending.
• Other current assets rose from ¥4.6bn to ¥22.6bn (短信 p.5), consistent with loans and advances to funds (立替金 +¥4.8bn, loans ¥11.2bn). Inference: more balance-sheet support for its own fund platform, which raises exposure if funds stall. Receivables doubled to ¥4.8bn against +69.7% sales (fact-sheet flag).
• Dilution. Shares outstanding 24.57m vs 19.76m (+24%) after the Nov 2025 offering (短信 p.14). Financing history: public offerings Dec 2021, Dec 2023, Nov 2025; CB Nov 2024 (presentation p.7). The company itself plans for ~11% average annual dilution against a 49% EPS CAGR (presentation p.68). The Oct 2025 offering announcement sent the stock limit-down (Nikkei).
• Capital efficiency. ROE 27.2% = net margin 10.2% × asset turnover 0.74 × leverage ~2.6 (avg). The fall from 32.5% came from lower leverage and margin. IFIS ROIC 9.7% (IFIS definition). Against an assumed cost of equity of ~9% (analyst assumption), the spread is wide. FY29 target ROE ≥25%.
• Returns. DPS ¥165 (payout 24.3%), FY27 ¥200 (20.5%); 6 consecutive increases; no buybacks since 2022. Yield 2.9%. Dividends are covered by earnings, not by FCF.

IR and management credibility

• Mid-term plan 1 (FY22–25): FY25 NP target ¥5.0bn reached in FY24, one year early; FY25 NP ¥10.25bn (presentation p.4).
• Mid-term plan 2 (FY25–29): NP ¥50bn and AUM ¥1.5tn by FY2029/08 (p.4–5). FY27 guidance (NP ¥24bn) is on that path.
• AUM targets met each year: ¥1,239bn → ¥2,329bn → ¥3,061bn → ¥9,476bn incl. development (p.14).
• ROE target and D/E history disclosed (p.7); dilution versus EPS growth disclosed openly (p.68).
• Weakness: single-segment reporting hides the profit split by hotel, logistics, healthcare and overseas, and between land-sale gains and recurring fees.

Valuation and scenarios

Shares 24.57m; BPS ¥3,531.57. Values are 12-month targets on FY2028/08 EPS.

Bear is cross-checked by book: FY27 BPS would be ~¥4,300, so the bear is close to 1.0x PBR.

• What the price implies. ¥6,960 ÷ band-mean 6.88x = EPS ¥1,012, i.e. the market is paying for FY27 guidance and nothing beyond it.
• Peers (Yahoo! Finance Japan, 2026-10-16; PER on company forecast, PBR on actual; implied ROE = PBR ÷ PER, an inference): Sun Frontier (8934) 7.10x / 0.91x / ~13%; Tosei (8923) 10.51x / 1.49x / ~14%; Ichigo (2337) 8.93x / 1.39x / ~16%; Kasumigaseki 7.12x / 1.97x / ~28%. On PER it is the cheapest of the group despite the highest ROE and growth.
• Why the market might be right: earnings depend on continuous sales into self-sponsored funds and a self-sponsored REIT, which work only while capital is cheap; rising JGB yields compress cap rates and fund demand together. Growth needs equity, so shareholders own a shrinking slice. And a single late sale can swing a year.

Bull case

• FY27 beats by ~10% as in 6 of 6 prior years; FY28 consensus moves up.
• No equity raise through FY27 because the 39% equity ratio leaves ¥70bn of debt headroom.
• PBR re-rates toward its 2.38x band mean (¥8,400 at current BPS).

Bear case

• 10-year JGB keeps rising; fund investors pull back; year-end sales slip and FY27 misses.
• Another public offering in late 2026 or 2027 (historical cadence: Dec 2021, Dec 2023, Nov 2024 CB, Nov 2025).
• Dubai assets (¥13.0bn moved to fixed assets) or loans to funds are impaired.
• ¥33.8bn of debt due within a year meets tighter real-estate lending.

Catalysts

What would change the rating

• Upgrade (to higher confidence) if H1 FY27 RP progress is at or above the FY24–26 median at H1, and no new equity is issued by the AGM.
• Downgrade to Hold if a new public offering above ~10% of shares is announced, or FY27 consensus OP falls below ¥37bn.
• Thesis invalidated if FY27 guidance is cut, if an impairment hits Dubai or fund loans, or if fund sales stall (inventory rising while sales fall).

Monitoring

Scorecard

Fact-sheet flags

Computed by factsheet.mjs. Each flag is a prompt to investigate, not a conclusion.

• ●redocf-negative Operating cash flow negative in 2 of the last 3 years (2024/08 -8,446, 2026/08 -16,765)
• ▲amberocf-weak OCF/NP over 2024/08-2026/08 is -0.57 (OCF -18,318, NP 31,962), below 0.5
• ▲amberreceivables-outpacing-sales receivables grew 110.9% in 2026/08 (2,261 to 4,769) against sales growth of 69.7%
• ▲amberfcf-negative Free cash flow -48,243 in 2026/08 (OCF -16,765, ICF -31,478, capex n/a)
• iinfofar-below-52w-high Price 6,960 is -31.8% from the 52-week high 10,200

Sources and limitations

• 2026-10-05 FY2026/08 決算短信: data/3498/2026-10-17/filings/2026-08_FY_短信_20261005.pdf (p.1–2 are the summary pages; later pages are cited by printed number, which is PDF index − 2).
• 2026-10-06 FY2026/08 決算説明資料: data/3498/2026-10-17/filings/2026-08_FY_説明会_20261006.pdf (cited by printed slide number).
• IFIS snapshot 2026-10-17 and fact sheet data/3498/analysis/2026-10-17/factsheet.md; disclosures and news from pages/report_disclose.html, pages/report_topix.html.
• Peer multiples: Sun Frontier, Tosei, Ichigo, 2026-10-16.
• Nov 2025 offering reaction: Nikkei. Rates context: web search, JGB 10-year above 3% since 2026-09 (secondary source, not verified against MoF data).
• Not read: Q3 短信 (in snapshot), FY2025/08 有報 and its 2026-01-30 correction, CB terms. The cause of the 8% fall on 2026-10-15 was not established: it coincided with a broker target cut to ¥10,700 (2026-10-14) and a fund JV announcement with 青山財産ネットワークス (2026-10-15); no negative disclosure was found.
• Peer ROE is inferred from PBR ÷ PER, not taken from filings.
• This is research, not individualised investment advice.