Forestar Group develops residential lots and sells ≈86.5% of them to D.R. Horton, its 62% parent (down from 75% at the 2017 merger) and primary customer. The Q2 FY2026 10-Q (filed April 23) printed a clean beat: revenues +7%, EPS $0.63. One line in the cost-of-sales footnote contradicts the headline.

The filing fact

Land option charges — write-offs when Forestar walks away from land purchase contracts — surged to $6.3M in Q2 versus $0.9M prior year (+600%). H1 charges $7.1M vs $2.0M (+255%). Total controlled lot pipeline contracted from 99,800 to 94,400 in six months (-5.4%). Capitalized interest balance grew to $91.5M (+27% YoY), a structural margin drag that bleeds through cost of sales over 12-24 months as developed lots are sold. Effective tax rate stepped from 22.4% to 26.7% (≈$0.04/share quarterly headwind, attributed to lower stock-comp tax benefits and higher state taxes — persistent, not one-time).

The other side of the ledger: operating cash burn collapsed to -$5.1M from -$469.8M prior year. That's capital discipline — but largely a function of buying less land, the same mechanism we read bearishly on the pipeline.

The cross-ticker contrast

D.R. Horton's own homebuilding segment booked $22.5M in earnest money write-offs in Q2 vs $24.0M prior year — DOWN 6%. Peers ranged from TOL -42% to KBH +47%. Forestar is the +600% outlier. The dollar delta is small ($5.4M of incremental charges); the signal is the rate of change against a parent running flat. The entire land option charge acceleration sits inside DHI's captive subsidiary; the parent's own land posture is stable. The reading: DHI is throttling its captive lot purchases from Forestar while keeping its own development pipeline intact.

The lawsuit

Mississippi PERS filed a derivative complaint in Delaware Chancery on April 29, 2025 (C.A. No. 2025-0465-MTZ). Per Law360 and Bloomberg Law coverage of the complaint, MPERS alleges "at least $700M" in damages from below-market lot sales between FOR and DHI — approximately 38% of FOR's $1.82B equity. The damages figure is from press paraphrase of the complaint; the 10-Q discloses the suit's existence but does not quantify the claim. The +600% vs -6% segment contrast is the legal hook: parent extracts via captive while keeping its own operations stable. Bidirectional optionality — overhang during pendency, but a derivative win returns damages TO Forestar, creating concentrated upside for minority shareholders.

What the market sees

FOR trades at 0.79x book ($35.66 book value per share). Book is largely $2.7B of contracted lot inventory, with $2.09B already under firm DHI orders — the mark is defensible. Short interest 2.6%, options open interest effectively dead (153 contracts on May, 210 on August). No sell-side coverage of the segment-level write-off asymmetry. Q2 was treated as a beat-and-raise: headline revenue growth was inflated by a one-quarter tract sales surge ($42.9M vs $4.1M prior year, composed of 332 acres dumped to non-DHI buyers plus a one-off $9.1M multifamily site to DHI), masking core lot revenue down 4.5% YoY.

Implied probability of the captive throttling thesis is roughly 10% (extended momentum, dead options, no segment-comparison coverage in sell-side notes). Our estimate from filing facts is roughly 35% — the four falsifiable predictions averaging in the mid-60s on individual mechanisms, weighted for the chance the narrative actually crystallizes. Edge of approximately 25 percentage points.

Why the gap

Sub-segment write-off disclosure is not aggregated by sell-side. The captive-throttling fingerprint requires reading FOR's 10-Q alongside DHI's segment-level disclosure — work neither company's research coverage performs systematically. The MPERS lawsuit sits in the legal proceedings footnote, not MD&A. FOR has limited sell-side coverage relative to its $50B parent.

Risks (ranked by impact)

1. Sector beta dominance. Approximately 65-75% of FOR's variance is homebuilder factor. A Fed cut and mortgage rate decline would lift the sector and overwhelm idio alpha. Outright FOR equity does not isolate the captive-throttling thesis; isolating it requires a long/short structure against DHI to hedge sector beta, with borrow availability a real constraint.
2. Vehicle execution. 62% controlled float, options dead, borrow likely 200-400 bps annually. Pair construction depends on FOR borrow availability.
3. Lawsuit bidirectionality. Plaintiff win on the derivative claim returns damages TO Forestar, creating ≈38% of equity in concentrated upside optionality. The lawsuit is not unambiguously bearish.
4. Interpretation ambiguity on tract sales. Q2 surge dressed up the headline beat. Q3 reversion is our base case (73%); if sustained tract sales reflect a deliberate portfolio strategy rather than one-off rationalization, the bear case weakens.

Catalysts

• July 21, 2026: Q3 FY2026 earnings — FOR and DHI report the same week. Three predictions resolve: tract sales < $15M (73%), land option charges > $3M (65%), and DHI parent disclosure on captive lot acquisition pace.
• November 2026: Q4 FY2026 earnings. FY2026 lot deliveries < 14,000 (35%), full-year gross margin < 22.0% (70%).
• Estimated H2 2026 / H1 2027: MPERS lawsuit motion-to-dismiss ruling.

What would change our mind

• Q3 land option charges revert to $1-2M (Q2 was a one-quarter cleanup, not structural pipeline retreat)
• DHI Q3 disclosure shows accelerated lot takedown from FOR (throttling reversal)
• Motion to dismiss granted on the MPERS suit (governance optionality dies)
• Mortgage rates collapse below 6% before July (sector recovery dominates idio thesis)
• Cross-ticker screening of TPH, TMHC, LGIH, RYI, STRL captive land arms shows no comparable +400%+ option charges vs flat-parent fingerprint (would weaken the pattern claim)

Evidence