Phase Space AI

Valuation

Inter Parfums [IPAR]

Inter Parfums [IPAR] — Valuation

Spot $126.50 · 32.028m shares · market cap $4,051.5m · net CASH $54.0m · NCI $223.3m (book) · EV $4,220.8m · TTM revenue $1,494.6m (to 2026-03-31) · EV/Sales 2.82x · EV/EBIT 15.7x on TTM EBIT of $269.4m

EV includes the noncontrolling interest, and that is the single most consequential construction choice in this memo. IPAR consolidates 100% of Interparfums SA (68% of net sales) while owning 72% of it. EBIT in the denominator is 100%; market cap is the 72% claim. Excluding NCI — which the screen did — understates EV by $223.3m and the EV/EBIT multiple by 0.8x (14.9x vs 15.7x). Book value is used as a floor; the minority's Euronext market value exceeds book, so the true adjustment is larger. Stated, sourced, and flagged as a floor rather than presented as exact.

EV is stated LEASE-EXCLUSIVE. Net cash = cash $80m + short-term investments $157m − total debt $183m = +$54.0m. Both asset legs read explicitly rather than through cashAndShortTermInvestments, per the INSM precedent.

---

Company state: A (mature and structurally stable) — the cleanest A in this cluster

Test applied: profitable; operating margin low-variance across ≥5 years; no structural regime change.

This is what STATE A actually looks like. Gross margin has moved 0.3pp in five years (63.3–63.9%). Operating margin has sat in a 17.9–19.1% band for four consecutive years. No transformative acquisition (the Goutal and Off-White IP purchases are single-brand and small). No accounting-basis break. Twenty annual periods of AV history, all profitable.

The 20-year record contains one genuine regime change and it is behind the window: the 2020 COVID trough at 13.0% and a revenue base of $539m, from which the business roughly tripled. The 2021–2025 window is post-normalisation and internally consistent.

Strong seasonality is present and must not be mistaken for instability. Quarterly operating margin runs Q1 ~21–22%, Q2 ~18%, Q3 ~25%, Q4 ~6–10% — Q4 carries the holiday A&P load. Annualising any single quarter produces nonsense, which is exactly what AV's OVERVIEW.OperatingMarginTTM of 21.5% does (it is Q1'26 alone).

evidence_grade: A−. IPAR's disclosure is the best of the three by a distance: every licence with its expiry date, brand-level revenue percentages, royalty expense in dollars and as a percentage of sales, promotion and advertising in dollars and as a percentage of sales with a forward target, NCI reconciled at both income and equity level, top customer named. The only material gaps are volume/price and the owned/licensed revenue split.

---

Terminal margin: 17.5%, below the FY2025 actual, because management guided A&P up

valuation.md rule 2 requires an explicit causal bridge for a below-current terminal margin. Here it is not an inference — it is the company's own quantified forward guidance.

Opex bridge reconciled from 10-K lines, FY2025 (% of net sales of $1,489.4m):

Terminal bridge (year 5):

What this terminal margin assumes about A&P, stated as the brief requires: it assumes A&P RISES 1.2pp to 21.0% of revenue and never levers, because that is what the company has told the market it intends to do, and because every licence agreement separately mandates minimum advertising expenditure as a contractual term. A&P here is not discretionary opex that can be flexed for margin — it is a covenant. The only expense line I have levered is non-A&P, non-royalty SG&A, by 0.7pp, and that is supported by a specific disclosed fact (headcount already carried for revenue not yet booked).

Constraints checked: - m_EBIT,T (17.5%) ≤ m_gross,T (63.7%) ✔ — 46.2pp of headroom. - Terminal is 0.7pp below FY2025 and 0.5pp below TTM. The DATA_DEFECTS.md warning is answered: it is below because management guided the largest opex line up by 1.2pp of revenue in the same document that reported the actual. A terminal margin at or above 18.2% would require ignoring the company's stated intent. - Not monotonicity-violating: rule 2 requires V_bull > V_base > V_bear, not m_bull > m_base > m_bear. A bull case pairs a lower margin (full 21% A&P) with higher revenue from the 2027–2030 launch pipeline.

---

Implied-path test (primary output — terminal value ~78% of EV)

reverse_dcf.py --spot 126.50 --shares 32.028 --net-cash -169 --revenue 1494.6
               --years 5 --wacc 0.09 --terminal-margin 0.175 --exit-multiple 14.0
               --fcf-margin 0.133 --hist-cagr 0.111

--net-cash -169 = the net of +$54.0m of genuine net cash less $223.3m of NCI, so the tool solves against the correct $4,220.8m EV rather than the screen's $4,047.0m.

--fcf-margin passed as +0.133 (TTM FCF $199.0m / revenue $1,494.6m = +13.31%). This is the largest FCF margin of the three names and therefore the largest correction — DATA_DEFECTS.md measures the omission bias at +3.57pp at a 15% FCF margin. The sign is firmly positive, so omitting it would have overstated required CAGR by roughly 3pp and suppressed a true PASS. Trailing and forward signs agree: capex is $23.4m against $222.3m of CFO, there is no capex cycle, and the company pays a $104m/yr dividend out of it. Basis: trailing TTM, and it is the right basis here.

Required revenue CAGR: 6.7% Demonstrated: 11.1% (FY2022 $1,086.7m → FY2025 $1,489.4m = 11.1% ✔) Margin: demonstrated − required = +4.4pp

Verdict: PASS. The price does not require acceleration — it requires roughly 60% of what the company has already delivered.

But the honest qualifier, which the screen itself flagged: the latest quarter grew 1.8%, not 11.1%, and FY2025 grew 2.5%. The 11.1% three-year CAGR is a backward-looking figure built substantially on the 2022–2023 post-COVID reflation. The required 6.7% sits above the current run rate, not below it. So the correct statement is: the price requires materially less than the three-year record and materially more than the last four quarters. That is a genuine tension and I am not resolving it in the model's favour.

The bridge between the two is the launch pipeline, and it is dated: Goutal (2026), Longchamp (2027), Off-White (2027), David Beckham (April 2028), Nautica (January 2030). Those are contractual, dated, disclosed revenue additions — the strongest forward evidence available on any of the three names in this cluster.

Implied compression: IPAR trades at 15.7x TTM EV/EBIT (NCI-inclusive). Solving at 14.0x assumes −10.8% compression — the mildest of the three, appropriate for a business decelerating from 11% to ~7%.

Sensitivity over the exit multiple

* the 10.0% figure printed at 14.0x in the flat-5%-FCF sensitivity sweep becomes 6.7% at the true 13.31% FCF margin — a 3.3pp swing from the interim-FCF parameter alone, which is the largest single-parameter sensitivity in this cluster and a direct confirmation of the DATA_DEFECTS.md magnitude table.

Flip point sits between 12x and 13x — the most robust of the three names. IPAR passes across a 2x range of exit multiples against the three-year record.

What steady_state_check.py flagged

steady_state_check.py --ticker IPAR --terminal-margin 0.175 --exit-multiple 14.0

• MARGIN RAMP: 2020 margin 13.0%, latest 18.2%, moved +5.2pp; "latest is window peak: False"; terminal 17.5% = −0.7pp vs latest, −1.6pp vs peak. No flag raised. The tool accepts that a terminal margin set below the latest print is not an extrapolated ramp endpoint — correct, and this is the tool working as designed.
• ROIC measured 17.2%; warranted multiple at steady-state ROIC 9.3x; terminal reinvestment rate implied by g/ROIC 17.4%; SCALING CONTAMINATION 1.00x.
• FLAG RAISED — EXIT_MULTIPLE_UNWARRANTED_AT_STEADY_STATE: "14.0x is 1.5x the 9.3x warranted at a 17% steady-state ROIC. It is justifiable only if today's returns persist forever, which is the assumption under test."

My response: this is the weakest of the three flags — 1.50x versus 1.77x (SN) and 1.83x (TPB) — and it is the one with the best counter-evidence, because IPAR's own 1,509-day EV/Sales history has never traded below 1.81x and its ROIC of 17.2% is a 20-year-stable licensed-royalty model rather than a scaling-phase artifact (SCALING CONTAMINATION 1.00x confirms no ramp). The identity's g=3% assumption is also least violent here: IPAR's required path is 6.7%, not 14–20%, so the gap between the terminal-growth assumption and the underwritten path is smallest. The flag is nonetheless accepted as live: at 9.3x the name fails by ~7pp.

---

12-month target: $140.75, +11.3% to spot (38th percentile of its own 6-year history)

Multiple history (own, EV/Sales, as-known TTM revenue lagged 45 days, split-adjusted bars):

• 1,509 trading days, 2020-07-27 → 2026-07-29 — six years, one continuous regime.
• min 1.81x · p25 2.55x · median 2.98x · p75 3.44x · max 4.42x
• current 2.82x = 38th percentile.

IDENTIFIED, not UNIDENTIFIED. The window is six years, spans COVID recovery and the 2022–23 inflation cycle, has genuine dispersion (1.81x–4.42x, a 2.4x range), contains no divestiture or separation that breaks the revenue denominator, and the SPLITS endpoint confirms no split inside it (the most recent was 2008). This is a valid reference class with information in the dimension being matched — the condition valuation.md rule 6 requires and which SN's series fails.

Build: - NTM revenue = 5/12 × FY2026 consensus $1,503.0m + 7/12 × FY2027 consensus $1,617.8m = $1,569.6m (AV EARNINGS_ESTIMATES, 3–5 analysts — thin coverage, stated). - Anchor multiple 2.98x = the 50th percentile of its own six-year history, against a current 38th percentile. The name is trading below its own median and I am anchoring at the median, not above it. - EV = 2.98 × $1,569.6m = $4,677.4m; less net cash $54.0m, plus NCI $223.3m → equity $4,508.1m; ÷ 32.028m shares. - Target $140.75, +11.3% to spot.

Band: at the p25 of 2.55x, $119.3 (−5.7%); at the p75 of 3.44x, $163.3 (+29.1%).

Named product events inside 12 months (see IPAR_Catalyst_Calendar.md): the Goutal 2026 launch (IPAR's first year of commercial use after acquiring the IP), an 11th fragrance in the Solférino collection with international distribution ramp, two new Lanvin initiatives in late 2026, and — the ones that matter as risk — the Anna Sui, Graff and Moncler licence expiries on 2026-12-31.

Estimate-revision direction (AV, 30-day trailing): FY2026 EPS 0 up / 4 down; FY2027 0 up / 0 down; Q2'26 and Q3'26 0 up / 0 down. Negative on the current year, neutral elsewhere. Consistent with the 1.8% latest-quarter print.

Sanity band: no external professional target was retrieved, so the check valuation.md calls for is not performed. Recorded as absent rather than asserted.

---

Downside case with a named cause (Type: MEASURED — does not reject the name)

Named cause: non-renewal of a top-four licence.

The four largest brands — Jimmy Choo 17%, Coach 15%, Montblanc 15%, GUESS 12% — are 59% of sales and all four are licensed. GUESS runs to 2048 and Coach to 2031, both freshly renewed, so the near-term exposures are Jimmy Choo (2031) and Montblanc (2030). Neither is inside the horizon, which is why this is a downside case rather than a base case.

The mechanism is real and has happened: Dunhill was discontinued in 2024 and the FY2025 10-K names it as the cause of the US revenue decline.

Quantified: loss of Jimmy Choo at renewal removes 17% of revenue (~$254m) at above-average margin, and the fixed cost base — retained headcount, distribution, the Paris and US infrastructure — does not shrink proportionally. Assume revenue $1,241m at a 14% operating margin (operating deleverage of ~3.5pp) = EBIT $174m. At a 12x multiple (a licensor-dependent business that has just demonstrated licensor risk deserves less than 14x): EV $2,086m, plus net cash $54m, less NCI $223m → equity $1,917m, $59.9/share, −52.7%.

Nearer term and less severe: the Lanvin repurchase right on 2027-07-01 lets the seller buy back the Lanvin brand names and trademarks for €70m (~$82m) at a fixed price. If Lanvin has become valuable, IPAR loses it at a 2021-struck price; if it has not, the option lapses. That is a dated, one-sided call written against IPAR's own owned-brand portfolio, and it is worth stating that this is the only dated, quantified structural risk of the three names in this cluster.