StoneX Group Inc. (SNEX) — Investment Analysis

Exchange: NASDAQ | Currency: USD | Analysis date: 2026-06-06 Current price: $115.54 (2026-06-05) | Market cap: ~$9.2B | Shares out: 79.25M (post two 3-for-2 splits)

⚠️ Split note: StoneX executed 3-for-2 stock splits in March 2025 and again in March 2026. All per-share figures in this document are stated on the current post-both-splits basis unless explicitly labeled "(pre-split)". Figures quoted in the FY2025 10-K (diluted EPS $5.89) and the Dec-2025 earnings call (BVPS $48.17, TTM EPS $6.70) predate the March-2026 split — divide by 1.5 to compare to today.

1. Executive Summary

Three-sentence thesis. StoneX is a genuinely high-quality, founder-aligned financial-intermediary compounder — a diversified commodities/financial-services broker, clearer, and market-maker that has grown operating revenues ~20%/year, compounded book value ~18%/year, and earned a 16–19% through-cycle ROE while never diluting or distracting shareholders with dividends. The problem is price: at $115.54 the stock trades at ~3.4x book and ~20x trailing / ~16x forward earnings, which fully capitalizes a FY2026 earnings run-rate that is being flattered simultaneously by the RJO Securities acquisition, still-elevated short rates, and a once-in-a-decade precious-metals quarter. On a probability-weighted, through-cycle basis my fair value is ~$81–89, roughly 25–30% below the current price, so despite the quality and the Tweedy Browne signal I rate it WAIT — a great business at a price that offers no margin of safety.

2. Phase 0 — Why Does This Opportunity Exist? (Klarman)

The screen flagged SNEX as "cheap AND a superinvestor bought" (Tweedy Browne increased its position +138% in Q1 2026). Both facts are real, but the "cheap" label requires scrutiny:

1. Relative-cheapness illusion. SNEX looks cheap on a P/E basis (~16x forward) for a business growing earnings 15–30%, and screeners that use the auto-calculated D/E of ~18x mis-classify it. But the forward P/E rests on a peak run-rate (see §5). Adjusted for normalization, it is fairly-to-richly valued.
2. Genuine neglect / complexity. SNEX is a structurally under-covered, hard-to-model business: $132B of gross revenue but only $4.1B of operating revenue, a balance sheet swollen by $45–54B of client and counterparty financing, four segments, and per-share data scrambled by two stock splits in 13 months. This complexity is a legitimate source of mispricing — in both directions.
3. Tweedy Browne's entry was at a lower price. The 52-week range is $70.93–$130.96. Tweedy added in Q1 2026 (calendar), when the stock traded materially below today's $115. A superinvestor buying at $80–95 does not validate a purchase at $115; the margin of safety they captured has largely closed.

Conclusion: The opportunity that existed (a quality compounder trading near book at the 52-week low) has been substantially harvested by the ~36% one-year run. What remains is a quality business at a full price.

3. The Business (from management's own words)

StoneX (formerly INTL FCStone) is a vertically integrated financial-services network connecting mid-market commercial, institutional, and retail clients to global markets across asset classes. Four operating segments (FY2025 segment income, pre-overhead allocation):

Key structural facts (FY2025 10-K MD&A):

• Total revenues $132.4B; operating revenues $4,126.9M (+20%); net operating revenues $2,052.8M (+16%); net income $305.9M (+17%); ROE 15.6% (avg-equity basis).
• Largest non-bank FCM in the US (8th overall after acquiring R.J. O'Brien, "RJO").
• Only non-bank participant in the daily gold/silver/platinum/palladium benchmark fixings; CME-accredited vault holding >$1.2B of metal.
• The franchise is genuinely diversified — Commercial and Institutional each contribute ~40% of segment income, with Retail and Payments each ~12–13%. No single segment dominates, which materially reduces single-point cyclicality versus a pure-play broker.

Earnings driver mix (the crux for valuation): Roughly 30–40% of net operating revenue is now net interest income on client float (interest income $1,734M − interest expense). This is the rate-sensitive engine that has supercharged FY2024–FY2026 results.

4. Phase 1 — Risk Analysis (Inversion: "Where could this die?")

For a leveraged financial intermediary, the dominant risks are not slow margin erosion but sudden, fat-tailed events. Expected loss = P(event) × impact.

Inversion answers:

• How could this lose 50%+ permanently? A large client deficit during an extreme, gapping market (think LME nickel 2022, or a counterparty failure) that exceeds collateral and the firm's reserves, combined with a simultaneous rate-cut cycle that removes the interest-income cushion — the firm survives but earnings power and the multiple both re-rate down hard.
• What makes me sell immediately (non-price)? A material bad-debt/client-deficit charge; net-capital ratio approaching regulatory minimums; loss of a top-3 segment leader or evidence of mis-marked inventory; abandonment of the conservative-risk culture.
• Three-sentence bear case: "SNEX is a leveraged, low-margin broker whose recent record earnings are a confluence of a debt-funded acquisition, peak short rates, and a freak precious-metals/volatility windfall — none of which repeat. Strip those out and you have a 15% ROE business priced at 3.4x book and 20x earnings, with a fat tail of client-default risk that has already nearly wiped out equity once (FY2017). At $115 you are paying a quality multiple for cyclical-peak earnings with no margin of safety."
• Can I state the bear case better than the bears? Yes — the bear case is simply "peak earnings × premium multiple," and it is the base case in my scenario tree.

5. Phase 2 — Financial Analysis

5.1 Through-cycle ROE (the right lens for a financial)

The auto-summary's 12.9% ROE uses year-end equity against a fast-growing equity base, which understates the return. The 10-K's average-monthly-equity methodology is correct:

Through-cycle ROE ≈ 16%, with a clear improving trend post-2017 and an accelerating recent print (Q2-FY2026 quarterly ROE 26.5%; return on tangible book ~32%). This is a legitimately good — not extraordinary — return on equity for a financial. It clears a 12–13% cost of equity but not by the wide margin a Buffett "great business" demands.

5.2 Book-value compounding (the real engine)

SNEX pays no dividend and retains 100% of earnings, reinvesting at ~16% ROE. Book value per share has roughly doubled in ~3 years; BVPS is $34.06 (Mar-2026). This is the correct way to value SNEX: a tax-efficient compounding machine where intrinsic value ≈ book value growing at ROE. The catch is that the multiple on book (3.4x) already pays for several years of that compounding in advance.

5.3 Leverage — read it correctly

The headline "D/E 18x" is an artifact of consolidating $45–54B of client and counterparty financing (repos, securities loaned, payables to clients) onto the balance sheet — these net against client assets and are not corporate risk. The real corporate capital structure:

• Senior secured borrowings $1,159M (incl. $625M Notes due 2032 @ 6.875%) + payable to lenders ~$782M (working facilities) ≈ **$1.9B corporate funding debt** vs equity $2.4–2.7B → corporate D/E ≈ 0.7–0.8x.
• Corporate interest coverage ≈ (pretax $408.8M + corporate interest $77.8M) / $77.8M ≈ 6x. Healthy.

The leverage that matters for a broker is regulatory net capital and the quality of client margining (value-at-risk-based, variation-margin discipline) — both of which management describes as conservative, and which the track record (one bad year in a decade) broadly supports.

5.4 Valuation

Because earnings are at a cyclically-flattered peak, I value SNEX three ways and blend, deliberately normalizing.

(a) Normalized-earnings multiple. Through-cycle normalized EPS ≈ $6.25 (midpoint of TTM $5.74 and a softer-than-run-rate FY2026). A quality-financial multiple of 12–16x → $75–100.

(b) Justified P/B (Gordon: P/B = (ROE − g)/(r − g)).

• ROE 17%, r 11%, g 9% → 4.0x → $136 (bull)
• ROE 16%, r 11.5%, g 8% → 2.3x → $78 (base)
• ROE 15%, r 12%, g 7% → 1.6x → $54 (conservative)

(c) 5-year forward fade-to-exit. Book grows 15%/yr to $69; exit at 2.5x P/B = $171; discount at 11% → **$102 today**.

At $115.54, the stock trades ~30% above conservative fair value. Only the bull P/B (4.0x, requiring 17% ROE sustained and 9% growth at an 11% discount rate) gets to today's price-plus.

5.5 Scenario probability tree (≈1-year horizon)

Probability-weighted fair value ≈ $81, a ~30% discount required from today's price just to reach an even risk/reward. The asymmetry currently points down.

6. Phase 3 — Moat Analysis

SNEX has a narrow moat — real but not wide.

Pricing power: Limited. Rate-per-contract and rate-per-million are competitively set and mean-reverting (FX RPM normalized down from 185 to ~116). The moat is about breadth and trust, not pricing.

Will the moat be wider or narrower in 10 years? Likely modestly wider — the ecosystem/cross-sell flywheel and consolidation of small FCMs (RJO) favor scaled survivors. But it is a competitive, commoditizing industry; this is a "stable-to-slightly-widening narrow moat," which supports a quality multiple only at the right price.

7. Phase 4 — Management & Incentives

• Founder alignment is excellent. Sean O'Connor led the company as CEO from 2002, built it from a sub-$50M-equity micro-cap into a $9B franchise, and transitioned to Executive Vice-Chairman in December 2024, handing CEO to Philip Smith (a long-tenured insider; CFO William Dunaway continues). O'Connor still owns 6.27% (~$575M) of the company.
• Insiders + directors own 11.79% as a group — genuine skin in the game.
• Capital allocation: No dividend, no value-destructive buybacks at high prices; capital is reinvested into the business and into disciplined acquisitions (RJO, Benchmark). The 20-year record of compounding book value ~18%/yr from a tiny base is the strongest evidence of capable, owner-minded allocation. The one blemish (FY2017) was managed through without permanent impairment of the franchise.
• Risk culture: Management repeatedly frames risk conservatively ("we like volatility, not extreme volatility"; VaR-based client margining; $1.2B of fixed SOFR swaps entered to hedge float). For a leveraged broker, this culture is the business.

Munger's incentive test: With 11.8% insider ownership and a no-dividend reinvestment model, management's incentives are well-aligned with long-term per-share book-value growth — exactly what you want. This is a key reason the name belongs on the watch list rather than the reject pile.

8. Phase 5 — Catalysts

No catalyst justifies paying up today. The positive catalysts are real but slow; the negative ones (rate cuts, normalization) are more probable on the relevant horizon. Per the framework, a no-near-term-catalyst quality name requires a 30%+ margin of safety — which is exactly what is absent at $115.

9. Phase 6 — Decision Synthesis

Quality: B+ (16% through-cycle ROE, excellent alignment, improving diversification, one near-death year). Moat: Narrow. Margin of safety at $115: Negative.

This is the textbook situation of a good business at a poor price. The Tweedy Browne signal is informative but their entry was lower; following them in at a 25–40% higher price discards their margin of safety. The framework's decision matrix (price > fair value, no compelling catalyst) yields WAIT.

Entry prices (post-split basis)

Position sizing

0% now. Watch-list only. If it pulls back to the low-$70s (toward where the superinvestor bought and toward ~2x book), initiate a half position (1.5%) and scale toward 3% below ~$62.

Pre-committed sell/avoid triggers (for if/when owned)

1. Any material client-deficit/bad-debt charge (>1 quarter of earnings).
2. Net-capital ratio trending toward regulatory minimums.
3. Departure of the conservative-risk culture or a top segment leader.
4. Evidence the RJO accretion masks underlying organic decline.

What I will NOT act on

A further melt-up on momentum (do not chase); short-term volatility swings; the mere fact that a respected fund owns it.

Appendix — Sources

Per project rules: no Yahoo Finance, no analyst reports/targets used. EODHD MCP returned 401; prices and financials sourced via AlphaVantage; filings via SEC EDGAR.