Asbury Automotive Group, Inc. (NYSE: ABG) — A Counter-Cyclical Service-and-Finance Annuity the Market Prices as a Peaking Car Lot

An independent equity-research note. Report date: 2026-06-12. Price reference: $199.48 (2026-06-11 close).

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice. The analysis that follows takes no position and contains no price target outside this block; do your own research.

Verdict: BUY / accumulate-on-weakness at ~$199 — a cyclical-value name where the price already discounts the bear’s earnings trajectory. Fair zone ~$200–250 on a normalized ~$22–25 EPS at 9–10x; cheap into the $170s (below book); rich above ~$285. Conviction: medium. The buyback pays you to wait.

The market is making a category error: it prices ABG as a car-seller at a cyclical peak, when ~71% of its gross profit comes from a counter-cyclical, high-margin annuity (parts & service at a 59% gross margin, plus F&I and the captive TCA underwriter at 93%). The stock trades at ~7x trailing / ~8–9x normalized earnings and 0.96x book — the 1.25th percentile of its own ten-year history, the cheapest clean EV/EBITDA (~6.7x) in the franchised-dealer group, while running the group’s highest operating margin. At $199 the embedded expectation is a bear-leaning base case: new-vehicle gross-profit-per-unit (GPU) reverting fully through management’s own $2,500–3,000 band, zero credit for the ~8%/yr sub-book buyback, the multi-decade roll-up runway, or the service/TCA annuity that demonstrably dampens the cycle (TCA segment income was flat YoY in Q1-2026 while dealership income fell ~25%). The framing is contrarian/cyclical-value, not quality-compounder: this is a well-run operator (disciplined ~3–5x-EBITDA M&A, hard-asset-backed, counter-cyclical buyback, gain-generating divestiture recycling) in a structurally average, capital-intensive, low-return industry (~8.6% ROIC, barely above WACC, negative tangible book). You are not buying a moat; you are buying a cheap, self-shrinking cash machine where the downside (~$120–150) is real but bounded and the path to ~$250+ requires only stabilization plus the buyback, not heroics.

What keeps conviction at medium rather than high: earnings are genuinely above-normalized and still reverting (Q1-2026 same-store new GPU $3,061 vs. a $5,583 COVID peak; full reversion to ~$2,750 strips ~$5/share, ~20%, off EPS), TTM GAAP EPS (~$28) is additionally flattered by ~$5/share of one-time divestiture gains, and three execution risks are stacked into one year (a CEO handoff, a company-wide Tekion DMS migration peaking mid-2026, and the Herb Chambers integration). The honest read is “statistically cheap, on a number that is still falling” — the value-trap risk is real, which is why the entry matters and why the elevated 10.7% short interest is directionally right on the near-term trajectory even as it under-weights the terminal value. Flips decisively bullish if new GPU stabilizes ≥$3,000 for three-plus quarters (management raised this view in Q1-2026) while the share count visibly shrinks ~8%/yr. Flips bearish if new GPU breaks below $2,500 and keeps falling, or if direct-sales reform spreads from EV-only makers to the legacy OEMs and begins dismantling the franchise-law moat.

One-liner: “Priced as a peaking car lot; it’s mostly a counter-cyclical service annuity that’s quietly buying itself in at below book.”

1. Executive Summary

Asbury Automotive Group is the smallest of the six public US franchised-auto-retail consolidators (AutoNation, Lithia, Group 1, Penske, Sonic, Asbury) — 223 new-vehicle franchises across 36 brands at 171 locations in 15 states, plus 39 collision centers and Total Care Auto (“TCA”), an in-house F&I product underwriter. Revenue has roughly tripled, from ~$7.2B (2019) to $18.0B (2025), almost entirely via acquisition: Park Place (2020), the transformational Larry H. Miller + TCA deal (Dec-2021, ~$3.2B), Jim Koons (Dec-2023, $1.5B) and Herb Chambers (2025, $1.76B).

The defining fact of the business is its profit architecture: vehicle sales are ~82% of revenue but only ~29% of gross profit. The real economics sit in two back-end pools — parts & service (47.9% of gross profit at a 58.7% margin) and F&I/TCA (23.4% of gross profit at a 93% margin) — which together throw off ~71% of gross profit on ~18% of revenue. Parts & service is a counter-cyclical, vehicle-parc-tied annuity defended by warranty/recall captivity; F&I, with TCA’s vertical integration, captures underwriting margin and ~$1B of insurance float that most peers cede to third parties. Wrapped around these annuities is a large, low-margin, hyper-cyclical vehicle-distribution engine.

That cyclicality is the whole investment debate. New-vehicle GPU exploded during the 2021–22 chip-shortage inventory drought and has been mean-reverting ever since: $4,462 (2021) → $5,583 peak (2022) → $4,702 (2023) → $3,697 (2024) → $3,432 (2025) → $3,061 same-store (Q1-2026). Group diluted EPS whipsawed in lockstep: $9.55 (2019) → $44.61 peak (2022) → $28.74 (2023) → $21.50 (2024) → $25.13 (2025). Management has explicitly guided that new GPU “will eventually stabilize in the $2,500–$3,000 range” — i.e., there is more reversion to come — and in Q1-2026 raised that view toward “closer to $3,000.” On a normalized GPU and stripping the recurring one-time divestiture gains (which inflate the ~$28 TTM EPS by ~$5/share), the honest earnings base is the low-$20s, partially defended by improving used-vehicle GPU, durable F&I per-unit revenue, a growing service annuity and a hidden, conservatively-deferred TCA income backlog.

On the balance sheet, the most important analytical move is to separate the debt. A naïve ~$5.4B “total debt” screen overstates leverage by ~$2B: $2.03B is self-liquidating floor-plan inventory financing (offset by inventory), not corporate leverage. Honest corporate net leverage is ~3.2x (within management’s 2.5–3.5x target, guided back below 3.0x by year-end 2026). Returns are ordinary — ROE ~13–14%, ROIC ~8.6% (barely above WACC) — and tangible book equity is negative (~−$487M), because $4.4B of acquired goodwill + franchise rights exceeds total equity, against which the company has impaired ~$140M/yr for three straight years.

Capital allocation is the bright spot: disciplined M&A (franchises bought at ~3–5x EBITDA with ~half the price hard-asset-backed by inventory and owned real estate), gain-generating divestiture recycling ($566M of proceeds in 2025), and a counter-cyclical, price-sensitive buyback that has shrunk the share count ~17% (22.4M → ~18.6M) and re-accelerated to $147M in Q1-2026 as the stock fell below book. The incentive plan is per-share-oriented (Adjusted EPS is the dominant LTI metric) but conspicuously lacks any ROIC hurdle — the one alignment gap for a debt-funded roll-up. Insider ownership is low (<1%), but a brand-new director made the only open-market purchases in 18 months (two code-P buys, ~$130k combined, on weakness below book), and activist Impactive Capital holds 6.5%.

On valuation, ABG is the cheapest of the franchised six on trailing P/E (7.05x) and the only one trading below book (0.96x, 1.25th percentile of its own history), despite the highest operating margin. The embedded expectation at $199 is a bear-leaning base case that prices full GPU reversion with no credit for the buyback, the consolidation runway or the annuity. Scenario value zones bracket roughly $120–150 (bear), $190–250 (base) and $280–360 (bull), with the current price hugging the bear/base boundary. The elevated short interest (10.7%) is sector-wide (Lithia 16.6%, Sonic 23.1%, Penske 13.5% are all higher) and expresses a thesis on the franchised-dealer model — directionally right on the near-term earnings trajectory, arguably too pessimistic on terminal value. This article takes no position in its body; the analysis below lays out the cyclical-normalization debate, the annuity, the capital-allocation discipline, and the structural risks, and leaves the judgment to the reader (and to Claude’s Take above).

2. Business Overview

What ABG does. Asbury Automotive Group is one of the largest franchised automotive retailers in the United States. As of 2025-12-31 it owned and operated 223 new-vehicle franchises representing 36 brands at 171 dealership locations, 39 collision centers, and Total Care Auto, Powered by Asbury (“TCA”) — its in-house finance-and-insurance (F&I) product underwriter — across 15 states (FY2025 10-K, Item 1, p.11). The model spans the full ownership lifecycle: sell a new or used vehicle, finance it and attach aftermarket protection products (F&I), then capture the recurring repair/maintenance/parts (P&S) annuity over the vehicle’s life. Revenue has roughly tripled from ~$7.2B (2019) to $18.0B (2025), almost entirely via M&A (Park Place 2020; Larry H. Miller + TCA Dec-2021; Jim Koons Dec-2023; Herb Chambers 2025).

Two reportable segments (FACT, 10-K Note 20):

• Dealerships — the store network. FY2025 external revenue $17,672.9M, segment operating income $858.6M.
• TCA — the captive F&I product company/reinsurer acquired with Larry H. Miller. FY2025 external revenue $326.1M, segment operating income $79.8M (a ~24% segment operating margin — far above the dealership segment’s ~4.8%). TCA carries $1,024.3M of segment assets (vs. $10,389.5M for Dealerships) and $536.6M of the group’s $2,281.3M goodwill — i.e., ABG paid up for this vertical-integration asset.

Revenue vs. gross-profit mix — the central fact of the business (FACT, FY2025 10-K MD&A, p.40–41):

The punchline: vehicle sales are ~82% of revenue but only ~29% of gross profit; P&S + F&I are ~18% of revenue but ~71% of gross profit. P&S alone (47.9% of GP at a 58.7% gross margin) is the single largest profit pool — a high-margin, vehicle-parc-tied annuity. F&I, net runs a ~93% gross margin because the cost of sales is essentially chargeback reserves; F&I gross profit per vehicle retailed was $2,214 (2025), up 1% YoY. This is the standard franchised-dealer profit architecture, and ABG is a clean expression of it.

Recurring vs. cyclical. P&S is the most defensive line — counter-cyclical (when new-car affordability falls, the existing fleet ages and needs more service; warranty/recall work is non-discretionary) and grew +6% in 2025 and +13% in 2024 even as new-vehicle gross profit fell (-3% in 2025, -9% in 2024). F&I/TCA revenue is sticky (premiums recognized over multi-year contract lives; ~$243.6M current deferred revenue). New- and used-vehicle gross profit is the cyclical, GPU-driven swing factor: new-vehicle gross profit per unit normalized from a COVID peak to $3,432 (2025) vs. $3,697 (2024), and group diluted EPS has whipsawed $9.55 (2019) → peak $44.61 (2022) → $28.74 (2023) → $21.50 (2024) → $25.13 (2025).

Brand/OEM mix (FACT, 10-K p.38): new-vehicle revenue mix is 40% import, 32% luxury, 28% domestic. The luxury skew (Park Place — Mercedes/Lexus/Porsche in Dallas; Herb Chambers — luxury in New England) matters: luxury new-vehicle GPU was $6,814 at a 9.1% gross margin in 2025, vs. import $2,302 (5.7%) and domestic — a structurally richer mix that also carries higher-dollar service work. Geography is concentrated in the Sun Belt / fast-growth states (Florida, Texas, Georgia, the Carolinas) plus the Larry H. Miller Mountain-West footprint (Utah/Colorado/Arizona) and now New England (Herb Chambers).

Clicklane is ABG’s proprietary omni-channel digital retail platform — end-to-end online purchase (trade appraisal, financing, F&I, home delivery) bolted onto the physical franchise network. Management positions it as the digital front-door to the dealership; it is an enabling tool, not a separate profit center, and is not a demonstrable moat.

Verdict: A scaled, well-diversified franchised retailer whose economics are dominated by two high-margin recurring pools (P&S, F&I/TCA) bolted onto a large, low-margin, cyclical vehicle-sales engine. The mix is the quality; the vehicle sales are the cyclicality.

3. Industry Dynamics

Structure: a regulated, fragmented oligop-of-locals. US franchised new-vehicle retail rests on a 50-state franchise-law system: state laws require new vehicles of a given brand to be sold through independently-owned franchised dealers, restrict OEMs from selling new vehicles directly to consumers, and restrict a competitor from relocating or opening a same-brand store inside an existing dealer’s protected market area (FACT, 10-K Item 1A, p.~38–39: “State automotive franchise laws restrict competitors from relocating their stores or establishing new stores of a particular vehicle brand within a specified area that is served by our dealership of the same vehicle brand”). This is the defining structural feature — it confers local-market exclusivity per brand and is the reason the franchised channel exists at all.

Market size & fragmentation → roll-up runway. There are ~16,000+ franchised new-car dealers in the US (the peer KMX report cites >18,000 franchised new-car rooftops). The six public consolidators — AutoNation (AN), Lithia (LAD), Group 1 (GPI), Penske (PAG), Sonic (SAH), Asbury (ABG) — combined still hold only a low-double-digit % of total franchised new-vehicle units. This is the structural opportunity: a highly fragmented, family-owned-dealer industry being slowly consolidated by acquirers who can buy single-store/small-group operators at 3–5x EBITDA, layer on scale (shared back office, F&I product, floorplan terms, data) and reprice the multiple. ABG’s own history (revenue 7.2B→18.0B in six years, almost all acquired) is this thesis in motion. Runway is genuinely long — decades — and acquisitions are typically immediately accretive because private multiples sit well below the public group’s.

The profit pools (where the durable economics live):

• Parts & service / fixed ops — the annuity. Tied to the vehicle parc (units-in-operation), and counter-cyclical: an aging fleet, deferred new purchases, and especially warranty/recall work (which by franchise law can typically only be performed by an authorized franchised dealer of that brand) keep service bays full when sales soften. Independent shops increasingly can’t service modern (software-defined, sensor-laden) vehicles — a widening captive-aftermarket advantage. This is the highest-quality revenue in the value chain.
• F&I — high-margin attach at point of sale (financing arranged via third parties for a flat/commission, plus aftermarket products: VSC, GAP, prepaid maintenance). ~93% gross margin, ~$2,200 per vehicle. Vertical integration of the product (TCA) lets a dealer capture underwriting margin + investment float that most peers cede to third-party providers.
• New & used vehicle gross — the commoditized, cyclical, price-transparent core. New-vehicle GPU is normalizing hard from the 2021–22 supply-shock windfall back toward pre-COVID ~$2,000–2,500 levels (ABG new GPU $3,432 in 2025, still above pre-COVID, still falling). This is where the cycle lives.

Threats.

• Direct-sales / EV circumvention. Tesla, Rivian, Lucid and other EV-only makers have been permitted in several states to bypass franchise laws and sell directly (FACT, 10-K: “certain electric vehicle manufacturers have been permitted to circumvent the state automotive franchise laws of several states”). If legacy OEMs ever win the legal/political fight to sell direct (they have tried), the dealer channel’s regulatory moat erodes. So far the franchise lobby (NADA + state dealer associations) has successfully defended the system for decades; it remains the single biggest structural risk and the one most outside ABG’s control.
• Online used-car disruptors (Carvana, CarMax). These attack the used and F&I pools, not the franchised new-car/warranty-service core. The peer cross-read is instructive: Carvana grew used units ~43% in a year on a ~1.5% share of a ~38M-unit fragmented market; CarMax holds only ~3.6% of its addressable used segment and that share is falling. Used retail is genuinely commoditizing and price-transparent — a real margin headwind to the dealers’ used operations — but it does not touch the new-vehicle franchise grant or the warranty/recall service annuity.
• Interest-rate sensitivity (two channels). (i) Floorplan financing — inventory is debt-financed; ABG carries $2.03B of floorplan notes payable and paid $91.2M floorplan interest in 2025. (ii) Consumer affordability — higher rates + post-COVID price inflation have pushed monthly payments to stress levels, suppressing new-unit demand (ABG same-store new units down ~6–9% in recent quarters).
• Tariffs. New 10-K risk factor flags “new tariffs or trade restrictions on imported vehicles or parts” — material given ABG’s 40% import mix.

Capital cycle (Marathon lens). The retail layer is not in a classic capital-flood/bust like the used-car-e-commerce layer was (2020–21 SPAC flood → 2023 bust: Shift Ch.11, Vroom wound down). Instead, the franchised-retail capital cycle is a slow, regulation-dampened consolidation: supply (rooftops) is fixed/declining, new entry is legally blocked, and capital is being deployed into acquiring existing supply rather than building new capacity. That is the favorable side of the capital cycle — returns are not being competed away by greenfield over-build because the franchise system prevents it. The unfavorable side: the earnings are riding down from a once-in-a-generation supply-shock GPU peak, so reported ROIC/EPS are mean-reverting regardless of how disciplined the operators are. The honest read is a structurally-protected channel experiencing a cyclical earnings normalization, not a capital-destroying over-supplied industry.

Verdict: a structurally good industry by the standards of retail — defended by franchise law (high regulatory barrier to entry), anchored by a counter-cyclical high-margin service/parts annuity, and offering a multi-decade fragmented roll-up runway at accretive private multiples. The qualifiers are real and not cosmetic: it is capital-intensive (inventory + real estate + floorplan debt), low-margin at the consolidated line (~17% gross, ~5% operating, ~3% net), cyclical (vehicle GPU normalization, rate-sensitive demand), and carries a tail risk that the very franchise-law moat could be legislated away by direct-sales reform. Good industry, mediocre standalone economics, real but partly-borrowed moat.

4. Competitive Position

Does ABG have a moat? Mostly an industry-wide regulatory one it shares with every peer, plus a few thin company-specific edges. In Greenwald’s taxonomy the genuine, ABG-specific advantage is local economies-of-scale + customer-captivity in fixed ops, and a modest vertical-integration edge in F&I (TCA). It is a well-run operator in a protected-but-commoditized industry, not a wide-moat franchise. Pressure-testing each candidate:

(a) Franchise-law regulatory protection — REAL but NOT ABG-specific. The local-market exclusivity and direct-sales ban is the strongest barrier in the business, but every franchised dealer enjoys it identically. It protects the channel against OEMs and new entrants; it does nothing to differentiate ABG from AutoNation, Lithia, Group 1, Penske or Sonic, who hold the same brand grants in their own territories. In Greenwald terms this is an industry barrier to entry, not a firm-level competitive advantage. It explains why the industry earns acceptable returns; it does not explain why ABG should out-earn its peers. Verdict: shared moat, not a differentiator.

(b) Local-market scale/density + brand portfolio — REAL but modest and local. Competitive advantage in this business is fundamentally local (Greenwald: dominate a local market, not the national one). Within a metro, density lets a dealer spread advertising, share inventory, build a service-customer base and earn better OEM allocation. ABG has built genuine local clusters (Dallas via Park Place; Atlanta/Jacksonville via Coggin; Mountain West via LHM; New England via Herb Chambers) with a luxury skew. But these local-scale economics are available to any consolidator that achieves the same density, and ABG is the smallest of the big six by revenue. Verdict: a real but replicable, geographically-bounded edge — not a durable national moat.

© Captive-aftermarket / fixed-ops switching advantage — REAL and durable, but industry-wide. Warranty and recall work on a given brand must (largely) be done by that brand’s franchised dealer; modern vehicle complexity (OEM diagnostic software, proprietary parts, technician certification) makes independents progressively unable to compete for technical repair. This is genuine customer captivity and is why P&S earns a ~59% gross margin and grew through the cycle. But, again, it is a feature every franchised dealer shares. ABG executes it well (and Tekion/CP-count/cycle-time initiatives aim to lift retention) but does not own a structural edge here over GPI or AN. Verdict: durable captivity, industry-wide not ABG-specific.

(d) TCA vertical integration — the genuine ABG differentiator, and the most interesting one. Acquired with LHM (Dec-2021), TCA lets ABG manufacture/underwrite its own F&I products (VSC, GAP, prepaid maintenance) and retain the underwriting margin + investment float that most peers hand to third-party providers (e.g., national VSC underwriters) for a flat commission. TCA earned $79.8M segment operating income on $326.1M external revenue (~24% margin) in 2025, holds $1.0B of segment assets (the reserve/float pool), and recognizes premium over multi-year contract lives. This is a structural, hard-to-replicate edge: it captures a margin pool peers forfeit, smooths earnings (it is the counter-cyclical segment — TCA operating income was flat at $21.2M in Q1-2026 vs $21.1M Q1-2025 while the dealership segment fell from $208.5M to $158.2M), and builds an insurance-float asset. It is the closest thing ABG has to a proprietary advantage. Caveats: it required a large goodwill outlay ($536.6M), the float carries underwriting/reserve risk, and Lithia (DFC) and others are building comparable captives — so it is an edge ABG got to earlier and runs at scale, not a permanent exclusive. Verdict: a genuine, financially-visible, company-specific differentiator — the best part of the moat case, but narrow and increasingly contested.

(e) Clicklane — NOT a moat. A capable omni-channel tool that improves the customer experience and may lower cost-to-serve, but it is not a network-effect or switching-cost asset; Carvana/CarMax and every peer have comparable digital capability, and there is no evidence Clicklane drives durable share gains or pricing power. Verdict: table-stakes technology, not a competitive advantage.

Does the edge show up in the numbers? Yes — at the margin line, modestly. Across the public franchised group, ABG has historically run the highest dealership operating margin and the cheapest multiple:

ABG’s margin lead is real but narrow (47bps over GPI, ~6bps over AN) and is explained by (i) the luxury/import brand skew (higher GPU, higher-dollar service), (ii) the TCA F&I-margin capture, and (iii) genuine cost discipline (SG&A/gross-profit historically among the lowest in the group, ~64–65% in 2025 — though it spiked to ~67% in early 2026 on the Tekion migration). It is not a moat-sized gap; a 0.5-point operating-margin lead in a 5%-margin business is good execution, not a fortress. ROIC is structurally modest and cyclical (cyclical-peak net margin ~3%; the business consumes capital in inventory, real estate and floorplan debt — $3.59B total debt + $2.03B floorplan).

Greenwald market-share-stability test. Share among the consolidators shuffles with M&A rather than organic share-shift — there is no evidence of a stable, defended ABG share gain at peers’ expense from a structural advantage; gains come from buying rooftops. That is the signature of a fragmented industry being consolidated, not of a firm with a demand-side moat.

Verdict: no wide, ABG-specific moat — a well-run operator inside an industry-wide regulatory moat. The durable barriers (franchise-law exclusivity, warranty/recall captivity, fixed-ops annuity) are real but shared by every franchised dealer. ABG’s own edges are (i) a richer luxury/import brand mix, (ii) disciplined cost control yielding a slim group-leading operating margin, and (iii) the TCA vertical-integration of F&I, which is the single genuine company-specific differentiator — visible in the margin, the float asset and the counter-cyclical earnings stability — but narrow and increasingly imitated. The honest characterization: a good business model (high-margin recurring fixed-ops + F&I) and a strong operator in a structurally-protected but cyclical, capital-intensive, low-margin industry — not a high-ROIC compounder with a proprietary moat. The investment case rests on consolidation runway + capital allocation + cyclical-trough pricing, not on a durable competitive advantage.

5. Growth History and Forward Opportunities

The headline growth is real but almost entirely acquired, and the organic core is in a normalization-driven decline — the single most important distinction in reading ABG’s trajectory. Revenue compounded from $7.21B (2019) to $18.0B (2025), a ~16.5% revenue CAGR, but this is a roll-up’s arithmetic, not organic momentum:

Organic (same-store) reality. Strip the acquisitions and the underlying business is flat-to-down. In FY2025, consolidated revenue rose 4.7% but was driven by Herb Chambers (~$2.9B of acquired annualized revenue) net of $566.5M of divestitures (24 franchises sold). On a same-store basis: new-vehicle gross profit fell ~8% (GPU −11%), used-vehicle revenue fell ~4%, and F&I revenue fell ~2%; only same-store parts & service grew (+3% revenue, +5% gross profit ex-recon). A reader looking only at consolidated revenue/EPS growth would entirely miss that the core, ex-acquisition franchise is contracting as the GPU windfall deflates. This is a critical interpretive point for valuation: the growth optics are M&A-financed; the organic engine is normalizing down.

The quality of growth, by source:

• Vehicle units (cyclical, low-quality): same-store new- and used-unit volumes have been soft (new units down ~6–9% in recent quarters) on affordability stress (new-vehicle ASP >$52K, elevated rates). This is the low-quality, price-taking, commoditized layer.
• Parts & service (the high-quality growth): +6% (2025) / +13% (2024) total, mid-single-digit same-store, on an aging, more-complex car parc (warranty work +19% in 2025) — recurring, counter-cyclical, 59%-margin. This is the only consistently-growing organic stream and the highest-quality growth in the business.
• F&I/TCA (durable): F&I per-vehicle revenue has held remarkably flat ($2,197–$2,351) through the entire cycle — pricing durability — and TCA’s deferred-premium book ($828M, +8% YoY) is a growing, conservatively-recognized income backlog.

Forward opportunities (ranked by quality and reliability):

1. The consolidation runway — the primary, most durable lever. A fragmented ~16,000±rooftop industry (the public six hold only a low-double-digit combined share) consolidating at 3–5x-EBITDA private multiples versus ABG’s own ~7x public multiple makes every disciplined tuck-in immediately accretive. The franchise-law moat protects the channel from greenfield over-build, so this is a decades-long, supply-constrained runway. ABG has executed it competently and repeatedly.
2. Tekion DMS cost-out — the clearest company-specific catalyst. The migration off CDK to Tekion (>50% of stores converted by Q1-2026, full conversion targeted fall-2026, peak disruption mid-2026) is a 2026 SG&A drag that management frames as flipping to a structural SG&A/gross-profit step-down in 2027. Early evidence at converted Koons stores (+21% gross-dollars-per-technician, +16% advisor productivity YoY) is encouraging but management-sourced — treat as hypothesis until it shows in consolidated SG&A.
3. Fixed-ops / service capacity — adding service bays and technicians to harvest the aging parc; the highest-return organic reinvestment available.
4. The buyback as a per-share “growth” lever — at ~8%/yr share shrink, per-share metrics can compound high-single-digit with zero operational improvement.
5. TCA penetration — extending the captive F&I underwriter across the acquired (Koons, Herb Chambers) and legacy footprint, capturing margin currently ceded to third-party providers.

Verdict: low-quality consolidated growth (M&A-financed, masking an organic decline as GPU normalizes), with a high-quality minority (parts & service, F&I/TCA) and two genuine forward levers — the long consolidation runway and the Tekion self-help. The growth that matters for value creation is the per-share kind (buyback + accretive tuck-ins + Tekion cost-out), not the revenue kind. Investors should anchor on same-store gross profit and per-share metrics, never on headline revenue/EPS growth, which is acquisition arithmetic sitting on a normalizing earnings base.

6. Financial Quality

All figures reconcile to ABG’s FY2025 10-K (filed 2026-02-20, fiscal year ended 2025-12-31), the FY2024 10-K (2025-02-26), the FY2022 10-K (2023-03-01), and the Q1-2026 10-Q (filed 2026-05-01). Per-unit data is from the MD&A operational tables; GPU/SG&A guidance from the Q4-2025 (2026-02) and Q1-2026 (2026-05) earnings calls.

6.1 Revenue & gross-profit composition — a low-margin distributor whose profit pool sits in the back end

ABG is a $18.0B-revenue franchised auto retailer, but the revenue line badly overstates the economic size of the business. The four reporting lines and their economics in FY2025 (consolidated income statement, 10-K p. 64):

The central fact: vehicle sales are 82% of revenue but only 29% of gross profit. New vehicles, more than half of revenue, throw off a 6.5% margin. The real profit pool is the high-margin back end — P&S (48% of GP at 59% margin) and F&I (23% of GP at 93% margin) together are 71% of gross profit on 18% of revenue. This is the structural reason auto retailers should be valued on gross profit and EBITDA, not revenue. The model is a distribution business (new/used vehicles, where ABG has explicitly no cost advantage vs. peers — 10-K p. 27) wrapped around two genuinely attractive annuity-like streams: a recurring, recession-resilient service business and a near-pure-margin F&I product engine.

Consolidated gross margin was 17.1% in FY2025 vs. 17.2% in FY2024 and 18.6% in FY2023 — a slow grind lower as the COVID-era new-vehicle margin bubble deflates, partly offset by used and P&S mix.

Same-store vs. acquired. FY2025 growth was almost entirely acquired, not organic. Total revenue rose 4.7% ($17,188.6M → $18,000.0M), driven by the Herb Chambers acquisition (closed July 2025; $2.9B of acquired annualized revenue) net of $566.5M of divestitures (24 franchises sold). On a same-store basis the underlying business was flat-to-down: same-store new-vehicle gross profit fell 8% (GPU −11%), same-store used revenue fell 4%, and same-store F&I revenue fell 2%. Only same-store P&S grew (+3% revenue, +5% GP ex-recon). Interpretation: organic gross profit is contracting; the consolidated growth optics are M&A-financed. This matters enormously for valuation — a reader looking only at consolidated revenue/EPS growth would miss that the core, ex-acquisition franchise is in a normalization-driven decline.

6.2 The GPU normalization story (CENTRAL to the thesis)

ABG’s earnings are dominated by new-vehicle gross profit per unit (GPU), which inflated grotesquely during the 2021–22 chip-shortage inventory drought (dealers had pricing power because there were no cars) and has been mean-reverting ever since. The trajectory, taken directly from the MD&A operational tables across four 10-Ks (all-store, full-year):

Several conclusions, and they cut in different directions:

(1) New GPU is the swing factor and it is still falling. From a $5,583 peak (2022) the all-store figure has fallen ~38% to $3,432 (2025) and the same-store Q1-2026 print was $3,061 (the all-store $3,371 is flattered by the Herb Chambers luxury mix). Management has repeatedly and explicitly guided that new-vehicle profitability “will eventually stabilize in the $2,500 to $3,000 range” (Q4-2025 call) — i.e., they are telling investors there is more reversion to come. At Q1-2026’s same-store $3,061 the business is still above the top of its own normalized range. This is the single most important number in the thesis: 2024–25 EPS is NOT yet at a normalized GPU level. Quantified sensitivity: ABG sold ~181,000 new units in 2025; moving the new GPU from $3,432 to the $2,750 midpoint of management’s range strips ~$124M of pre-tax gross profit, ≈$93M after tax, ≈$5.00 of EPS off an 18.6M share base — against FY2025 reported diluted EPS of $25.13. That is a ~20% headwind to reported EPS still embedded if management’s own normalization call is right, before any offset.

(2) The offsets are real but smaller. Used-vehicle GPU has improved ($1,517 in 2024 → $1,672 in 2025 → $1,828 same-store in Q1-2026, +12% YoY), the result of a deliberate “profitability over units” sourcing discipline — six sequential quarterly GPU increases in the last seven quarters. F&I PVR has held remarkably flat ($2,197–$2,351) through the entire cycle — evidence of genuine pricing durability in the back end. And P&S gross profit grew 10% in 2025 (warranty +19%, customer-pay +8%) as an aging, more-complex car parc drives service demand. These are the durable, defensible streams; they partially cushion the new-GPU air pocket but do not fully offset it.

(3) Net verdict on earnings level: above-normalized, with reversion not complete. Pre-COVID (2019) ABG earned ~$9.55 diluted EPS on ~$7.2B revenue. The business is structurally larger and better-mixed now (LHM, Park Place, Herb Chambers added ~$6B+ of revenue; F&I/TCA captive insurance built out; used and P&S more disciplined), so a return to 2019 EPS is not the right normalization anchor. But FY2025’s $25.13 (and TTM $28.29) sits on new GPUs still ~$400–700/unit above management’s own stated stabilization band. A defensible normalized EPS is materially below the reported/TTM figure — call it the low-$20s once new GPU completes its glide to ~$2,750 and assuming used/P&S hold. The bull case requires the new GPU to stabilize at the high end ($3,000) AND used/P&S/Tekion cost-out to keep building; the bear case is new GPU breaks below $2,500 in a recession while volumes also fall.

6.3 Capital structure — SEPARATING THE DEBT (the single most important QoE move)

A naïve screen (e.g., a standard data terminal) shows ABG with ~$5.4B “total debt” and concludes the company is heavily levered. That figure conflates two economically opposite things and overstates real leverage by ~$2B. The honest decomposition (10-K balance sheet p. 65 and Debt note 14, p. 96):

(a) Floor plan ($2.03B) is NOT real leverage. It finances new- and used-vehicle inventory, is collateralized by and offset by that inventory, and self-liquidates as cars sell. ABG even runs floor-plan “offset accounts” (cash parked to net against it). It is a real interest cost — floor plan is tied to SOFR, and floor-plan interest expense was $91.2M in FY2025 (vs. $89.9M in 2024 but only $9.6M in 2023, when manufacturer floor-plan assistance and lower balances/rates masked it). So floor plan belongs in the cost structure analysis (a SOFR-sensitive expense, partly reimbursed by OEM floor-plan assistance) but not in net-debt/leverage.

(b) Real corporate net leverage. Gross corporate debt $3,593.4M, less $40.4M cash = ~$3,553M net (the company holds almost no cash — it sweeps to the floor-plan offset; “liquidity” of $927M at YE2025 is offset accounts + revolver/used-line availability, not balance-sheet cash). The four senior notes ($2.25B at a blended ~4.7% fixed, well-laddered 2028/2029/2030/2032) are the cheap, fixed core; the ~$1.2B of real-estate term loans are floating (SOFR-based) and amortizing, secured by owned dealership real estate (PP&E net is $3.07B). ABG reports a “transaction-adjusted net leverage ratio” of 3.2x at YE2025 (and again at Q1-2026), up from 2.9x at YE2024 — the increase entirely attributable to the debt-funded Herb Chambers deal. That sits at the upper end of management’s stated 2.5x–3.5x target range. My own crude check (net corporate debt $3,553M / adjusted EBITDA ~$1,084M [operating income $860.6M + D&A $82.4M + asset impairments $141.0M]) = ~3.3x, corroborating the company’s 3.2x. The honest corporate net-leverage figure is ~3.2x — NOT the ~$5.4B / >4x a conflated total-debt screen implies. Management has guided to getting below 3.0x by summer/year-end 2026 from divestiture proceeds (Q4-2025/Q1-2026 calls), balanced against opportunistic buybacks.

© TCA / finance-related. TCA (the captive F&I underwriter) carries its own restricted cash/investment portfolio (Investments $414.7M on the balance sheet, plus restricted cash) set aside to satisfy future insurance claims, funded by deferred premium ($828.2M total deferred revenue). This is ring-fenced (the Landcar TCA subsidiaries are non-guarantors of the senior notes) and economically distinct from corporate leverage — it is an insurance float, not debt.

6.4 Margins, returns, FCF, SBC, interest drag

Margins. FY2025: gross 17.1%, operating 4.8% ($860.6M), net 2.7% ($492.0M). Operating margin is structurally thin and cyclically compressing — it peaked at ~8.2% in 2022 ($1,272.6M op income on $15.4B) and has roughly halved as GPU normalized. The 2.7% net margin is characteristic of the sector (a distributor’s economics); the question is never the headline margin but the return on capital the back-end annuities generate.

Returns. ROE is healthy: net income $492.0M / average equity (($3,891.9M+$3,502.1M)/2 = $3,697M) = ~13.3% (end-of-period ~12.6%). The ~14.5% figure cited elsewhere reflects FY2024 or an adjusted base; on FY2025 GAAP it is ~13%. ROIC is the more revealing — and more sobering — number. NOPAT ($860.6M × (1−25.7%) = ~$639M) / invested capital (equity $3,891.9M + corporate debt $3,572M − cash $40.4M = ~$7,424M) = ~8.6% (≈10% if asset impairments are added back to NOPAT). That is barely above a plausible WACC. The reason is the M&A-built balance sheet: goodwill ($2,281.3M) + intangible franchise rights ($2,097.6M) = $4,378.9M of acquired intangibles, which EXCEED total shareholders’ equity of $3,891.9M — tangible book equity is NEGATIVE (~−$487M). Greenwald’s lens: the franchise rights are a genuine regulatory/relationship barrier (state franchise laws), but ABG paid full price for them via Park Place/LHM/Herb Chambers, so the economic-value-creation test is whether incremental ROIC on those deals beats the cost of capital — at ~8.6% consolidated ROIC, the roll-up is creating only thin value above cost, and a recession-driven GPU air-pocket would push it below WACC.

FCF. Operating cash flow was a strong $775.2M in FY2025 (up from $671.2M in 2024 and $313.0M in 2023 — the 2023 trough reflected a $575.7M inventory build as supply normalized). Capex was $186.0M (ex-real-estate) + $19.3M real estate = $205.3M. Simple FCF (OCF − total capex) = ~$570M; OCF − maintenance capex ≈ $589M. Management’s own “adjusted free cash flow” was $465M for FY2025 and $120M in Q1-2026. Caveat on OCF quality: OCF is inflated by the accounting geography of floor plan — trade floor-plan changes run through operating cash flow (−$6.8M in 2025, but +$154.8M in 2024 and +$144.1M in 2023, which flattered those years’ OCF) while non-trade floor plan runs through financing. A clean read strips floor-plan-trade swings; doing so, “adjusted cash flow from operations” was actually $651.4M (2025) / $688.4M (2024) / $705.3M (2023) per the 10-K — i.e., on a normalized basis operating cash generation has been declining, the opposite of the reported OCF trend. This is a genuine QoE nuance: reported OCF rose 2023→2025 largely because the 2023 inventory rebuild reversed, not because the business got more cash-generative.

SBC is immaterial and conservatively handled: $27.7M in FY2025 (0.15% of revenue, ~5.6% of net income), expensed straight-line, forfeitures recognized as incurred. There is no dilution story — share count is shrinking, not growing (see Capital Allocation).

Interest drag. Two distinct buckets, both SOFR-exposed: floor-plan interest $91.2M and other interest expense $187.5M (FY2025), $278.7M combined vs. $269.0M in 2024 and only $165.7M in 2023. The 10-K discloses $2.62B of variable-rate debt (floor plan + used line + revolver + floating mortgages); a 100bp rate move = $26.2M of annual interest. The senior-note core is fixed and cheap (~4.7%), so the rate sensitivity is concentrated in floor plan and the real-estate term loans. As rates fall, this is a modest tailwind; the Herb Chambers RE facility (SOFR+2%) added ~$537M of floating drag in H2-2025.

6.5 Quality-of-earnings flags

(1) TCA insurance accounting is CONSERVATIVE, not aggressive — a positive QoE finding. TCA (the captive F&I underwriter acquired with LHM/Landcar in 2021) sells extended service contracts, GAP and protection products. Premium is deferred and recognized ratably over the policy life, with acquisition costs capitalized and amortized in step. Total deferred revenue was $828.2M at YE2025 (up from $766.0M) — a large and growing liability that defers, rather than front-loads, income. Management quantifies the drag explicitly: the “non-cash TCA deferral headwind” reduced adjusted EPS by $0.31 in Q4-2025 and $0.26 ($7M pre-tax) in Q1-2026. Translation: as TCA’s in-force book grows, current reported earnings are being suppressed by deferral, building a backlog of recognized-later income. This is the opposite of an earnings-quality red flag — it is a hidden, conservatively-deferred asset. The risk is on the claims side (reserve adequacy on the in-force book), not revenue timing.

(2) Gains on dealership divestitures inflate “other income” and EPS — strip them. ABG books divestiture gains in non-operating income: $80.2M pre-tax in FY2025 (24 franchises sold for $566.5M; vs. only $8.6M in 2024 and $13.5M in 2023), and a much larger $125.8M pre-tax in Q1-2026 (vs. $4.1M in Q1-2025). These are real cash-generating capital-allocation actions, but they are non-recurring and should be excluded from run-rate earnings. Management does adjust them out (Q1-2026 adjusted net income $102M / adjusted EPS $5.37 excludes the $94M after-tax divestiture gain, plus ~$5M Tekion implementation, ~$3M weather, ~$1M duplicate-DMS costs — vs. reported Q1-2026 diluted EPS of $9.87).

(3) TTM EPS $28.24/$28.29 is OVERSTATED by the Q1-2026 divestiture gain — reconciliation. TTM diluted EPS = FY2025 $25.13 − Q1-2025 $6.71 + Q1-2026 $9.87 = $28.29 (matches the third-party aggregator’s $28.24). The entire $3.16 step-up over FY2025 is explained by the Q1-2026 divestiture gain: the after-tax $94M gain ÷ ~19.0M diluted shares = ~$4.95 of EPS — i.e., TTM EPS is flattered by ~$5 of one-time gains, not a sign of underlying acceleration. Stripping it, Q1-2026 adjusted EPS was $5.37 (and Q1-2026 was additionally hit by ~$0.56 of weather and the $0.26 TCA-deferral headwind). The honest underlying run-rate is below FY2025’s $25.13, not above the TTM $28.29. Anyone valuing ABG off TTM GAAP EPS is anchoring to a one-time-gain-inflated number on top of a still-above-normalized GPU — a double overstatement.

(4) Asset impairments are a recurring “non-recurring” item. ABG took $141.0M of franchise-rights/goodwill impairments in FY2025, $149.5M in 2024, and $117.2M in 2023 — ~$140M/yr for three straight years. These are non-cash and management adjusts them out, but their persistence is itself a signal: the company is regularly writing down intangibles it paid for in acquisitions, consistent with the thin ROIC finding — i.e., some acquired franchise rights have not earned their purchase price.

(5) Net income vs. cash flow. Reported OCF ($775.2M) comfortably exceeds net income ($492.0M), with the gap explained by non-cash D&A ($82.4M), impairments ($141.0M), SBC ($27.7M), loaner-vehicle amortization ($56.8M) and deferred taxes ($28.2M), less the divestiture gain (−$80.2M). No accrual red flag — but recall that the trend in clean OCF is down, masked by floor-plan geography.

Verdict: A low-margin cyclical with M&A-inflated GAAP — economics improve only modestly with scale, and reported earnings are above-normalized

ABG is not a business whose unit economics improve materially with scale. It is a thin-margin (2.7% net, 4.8% operating) vehicle distributor whose genuine economic value sits in two back-end annuities — P&S (recurring, recession-resilient, 59% margin) and F&I/TCA (near-pure-margin, conservatively deferred). Scale brings some procurement, fixed-cost-absorption, and Tekion-driven SG&A leverage (management’s clearest forward catalyst: a back-half-2026/2027 cost-out story), and it diversifies brand/geographic risk — but consolidated ROIC of ~8.6% (barely above WACC), negative tangible equity, and ~$140M/yr of recurring intangible impairments show the roll-up is paying near-full price for the franchise rights and creating only thin value above cost. On earnings level, the verdict is clear and directional: FY2025 reported EPS of $25.13 — and especially the TTM $28.29 — overstates normalized power. New-vehicle GPU ($3,432 all-store FY2025; $3,061 same-store Q1-2026) is still above management’s own $2,500–$3,000 stabilization range, embedding a ~$5/share (~20%) reported-EPS headwind still to come; TTM EPS is additionally inflated by ~$5/share of one-time Q1-2026 divestiture gains. The honest, normalized earnings base is the low-$20s, partially defended by improving used GPU, durable F&I PVR, a growing service annuity, the hidden TCA deferral backlog, and a real ~$570M of FCF supporting an aggressive buyback. The leverage scare is overstated — true corporate net leverage is ~3.2x, not the >4x a $5.4B conflated-debt screen implies, and it is on a path back below 3.0x. This is a competently-run, well-financed cyclical at a cyclically-elevated (if normalizing) earnings level — the economics are adequate, not compounding.

7. Capital Allocation

Verdict (preview): Disciplined and rational, with one structural caveat — the moat in the prices paid and capital recycled, not in the returns earned. Asbury allocates capital the way a Marathon “capital-returns” investor would want a consolidator to: it buys private dealership groups at low single-digit EBITDA multiples (plus separately-valued real estate), recycles lower-return stores out via gain-generating divestitures, shrinks the share count opportunistically when the stock is cheap, and de-levers on a schedule. The caveat is that the consolidated returns this produces are ordinary — ROIC ~8.6%, ROE ~14.5%, and the stock trades below tangible-adjusted book — so management is a skilled buyer of mediocre assets, not a compounder of capital at high rates. It is the favorable side of the asset-growth anomaly (buying below replacement-ish value, not bidding returns down), but it is not a high-return machine.

M&A track record — multiples paid vs. value created

ABG’s growth is overwhelmingly acquired. The transformational sequence:

(Source: FY2025 10-K acquisition note, p.80; FY2023 10-K acquisition note.)

Two things stand out, both favorable under a Greenwald/Marathon lens:

1. The intangible premium is consistent and contained (~44% of price), and the balance is hard-asset-backed. In both Koons and Herb Chambers, roughly half the purchase price is inventory plus owned real estate (Herb Chambers: P&E $605.5M + inventory $372.1M = $977.6M of $1,761.8M; Koons: P&E $420.0M + inventory $311.6M + assets-held-for-sale $103.8M = $835.4M of $1,500.0M). ABG is buying franchises at a stated ~3–5x EBITDA and paying separately for the dirt — it is not paying a momentum multiple for blue-sky goodwill. This is the disciplined-acquirer signature: the premium over tangible assets (the franchise rights + goodwill) is the capitalized excess return of the franchise, and at these multiples the implied earnings yield on the operating business is high.

2. It is consistently debt-funded post-2021, and management de-levers on a clock. Herb Chambers lifted transaction-adjusted net leverage from 2.9x (YE2024) to 3.2x (YE2025) — within the 2.5x–3.5x target band — and management guides back below 3.0x by mid/late-2026, funded by divestiture proceeds and FCF (Q4-2025 call; FY2025 10-K MD&A). The 2021 LHM/TCA deal was the only one to use equity (shares 19.3M→22.4M); every deal since has avoided dilution.

But the value created is unremarkable. The cumulative consequence of all this M&A is $2,281.3M goodwill + $2,097.6M intangible franchise rights = $4,378.9M of intangibles, ~38% of the $11,618.2M balance sheet (FY2025 BS, p.64), against which the business earns ROIC of only ~8.6% and ROE ~14.5%. High returns have not been bid down by ABG’s buying — but they were never high to begin with. This is the honest read: ABG is a competent roll-up of a structurally average industry (auto retail), and the discipline shows up in the price paid and the real-estate backing, not in a rising return on capital. The stock trading below book (~0.95x) is the market’s verdict that the acquired franchise rights may not be worth their carrying value through-cycle — a live impairment risk (note Q4-2025 booked an $87M non-cash asset impairment, the canary).

The buyback — opportunistic and counter-cyclical, but modest in scale

The prompt’s flag is correct and important: the XBRL PaymentsForRepurchaseOfCommonStock tag captures only the RSU net-share-settlement line (“Repurchases of common stock, including amounts associated with net share settlements”) of $12.8M (2025), $10.2M (2024), $11.4M (2023). The real buyback is the separate cash-flow line “Purchase of treasury stock”:

(Source: FY2025 10-K cash-flow statement p.69 + share-repurchase disclosure; Q1-2026 call.)

Three-year program spend ~$550M (2023–25) plus $147M in Q1-2026. Remaining authorization: $175.9M as of 2025-12-31 (under the $400.0M authorization set 2024-05-15, itself a $256.2M increase). Diluted share count fell 22.4M (2022) → 19.6M (2025) → ~18.6M now, ~17% reduction — a genuine per-share tailwind.

Timing is the right kind — counter-cyclical and price-sensitive, not procyclical empire-building. Management is explicit (Q4-2025, Q1-2026 calls) that buyback pace is “dictated by share price, leverage profile, economic conditions, and trade-offs with strategic tuck-in acquisitions.” The 2023 buyback was the largest ($267.7M) and at the lowest average price (~$203); as the stock ran to $230s in 2025 they slowed (down to $99.9M); and as it fell back toward $182–217 in Q1-2026 they re-accelerated to $147M in a single quarter. With the stock at ~$199 (below book) and the valuation_index at the 1.25th percentile of its own 10-yr P/B history, continued aggressive repurchase here would be value-accretive. The honest caveat: at ~$465M adjusted FCF and a 3.2x leverage profile mid-de-lever, the scale of buyback is capped — this is a “balance it as we go” program, not a Sloan-style share cannon.

Divestitures as a capital-recycling tool — a genuine, recurring discipline

A distinguishing and favorable feature: ABG actively sells lower-return stores, generating cash and gains, and redeploys into buybacks/de-leverage/luxury mix. FY2025 divestiture proceeds were $566.5M (vs $196.3M 2024, $30.7M 2023); Q1-2026 divested 10 dealerships + a collision center (~$600M annualized revenue), directing $147M of proceeds to buybacks and the rest to debt. Management frames this as avoiding “low-return CapEx” and lifting the luxury mix from ~32% to ~36% (Q4-2025 call). This is precisely the supply-side discipline Marathon prizes — pruning capital from low-return uses rather than growing assets for growth’s sake. (Skeptic’s note: the divestiture gains recur and flatter GAAP earnings — strip them when normalizing; management’s adjusted figures correctly back them out, e.g. the $26M Q4 divestiture gain excluded from adjusted NI.)

Capex — rising, with the Tekion bet on top

Capex (ex-real estate) $186.0M (2025), $162.6M (2024), $142.3M (2023); real-estate purchases $19.3M (2025), $145.6M (2024). 2026 capex guided to ~$250M (and again 2027) — a step-up driven by facility upgrades, service-bay expansion, and the Tekion DMS migration (replacing CDK). Tekion is the live operational bet: >50% of stores converted by Q1-2026, full conversion targeted by fall-2026, with integration cost/disruption peaking Q2-2026. The early evidence is encouraging but management-sourced (treat as hypothesis): at the Koons stores converted summer-2025, gross-dollars-per-technician +21% YoY and productivity-per-service-advisor +16% YoY (Q1-2026 call). If real and durable, Tekion lowers the fixed-ops cost-to-serve — the highest-margin profit pool — and is a rare organic return-enhancing use of capital in an otherwise M&A-driven model.

Priority stack and verdict

The revealed capital-allocation priority order: (1) fund/maintain the floor-plan working-capital engine; (2) tuck-in M&A at disciplined multiples when available; (3) de-lever to <3.0x; (4) opportunistic, price-sensitive buybacks; (5) capex/Tekion. No dividend — appropriate for a sub-book, FCF-recycling consolidator. Verdict: management has allocated capital intelligently — disciplined acquisition multiples, hard-asset backing, gain-generating divestiture recycling, and counter-cyclical buybacks are all the right behaviors. The reason this does not translate into a high-return compounder is the industry, not the allocator: even well-bought auto-retail assets earn ~8–9% ROIC. ABG is doing a good job in a structurally average business — exactly the Greenwald “good operator, no demand moat” profile.

Compensation & Incentive Alignment

Incentive metrics (FY2025 proxy, DEF 14A 2026-03-24). The proxy’s stated “Most Important Performance Measures” are Adjusted EBITDA, Adjusted EPS, and Adjusted Operating Margin (p.49). Concretely:

• Annual cash incentive: EBITDA, with target EBITDA scaled by USAAS (used-store-as-adjusted same-store, a same-store sales-growth gate — “appropriate to increase target EBITDA at higher levels of USAAS”). So the annual bonus rewards both profit and same-store performance, not raw revenue/size.
• Long-term incentive (LTI): 60% performance share units (PSUs) + 40% time-vested RSUs. PSU metrics: Absolute Adjusted EPS (70% weight) + Adjusted EPS Growth Relative to Peers (30% weight), with an absolute TSR ±10% modifier (TSR up >10% adds 10pts; down >10% subtracts 10pts; down >20% caps payout at 100%). The relative component is capped at 100% of target. FY2025 actual PSU payout = 100% of target (Absolute EPS 127%×70% = 89%; relative EPS-growth 0–45% band = 11%; TSR modifier 0%).

Read on alignment. The metrics are per-share-oriented (Adjusted EPS is the dominant LTI metric, 70% PSU weight), which is good — it means buybacks and disciplined share issuance directly help executives, and there is no pure-revenue/size metric rewarding empire-building. The TSR modifier ties some pay to actual shareholder outcomes. The Marathon-relevant gap: there is NO ROIC / return-on-capital metric anywhere in the plan (the proxy contains zero instances of “ROIC,” “return on invested capital,” or “relative TSR” as a primary metric). For a debt-funded roll-up whose central risk is over-paying for franchise rights, the absence of a return-on-capital hurdle is the one genuine alignment weakness — EPS can be grown by accretive-looking but low-ROIC acquisitions financed with cheap floor-plan/real-estate debt. Adjusted Operating Margin partly compensates (ABG’s 5.6–5.8% adjusted op margin is “highest amongst the Automotive Peer Group”), but margin ≠ capital efficiency.

Pay quantum (Summary Compensation Table, FY2025). CEO David Hult total $10,735,903 (2025), up from $8,999,130 (2024) and $8,138,627 (2023) — base $1.3M flat, stock awards $7.0M, non-equity incentive $2.4M. Incoming CEO Daniel Clara (COO) $3,251,100 (2025); CFO Michael Welch $3,040,163. By S&P MidCap standards this is moderate, not egregious — the CEO’s pay is ~2% of net income and heavily equity-weighted (~65% stock). Hult’s Executive-Chairman package (per 2025-12-08 8-K) steps the base down sharply ($750k 2026 → $525k 2027 → $300k 2028) — a sensible, declining sunset, not a golden parachute.

Insider ownership — LOW (a real negative). Per the FY2026 proxy Security-Ownership table (the authoritative source; disregard any third-party-aggregator figure as a stale artifact): directors and executive officers as a group (15 persons) own only 145,318 shares — well under 1%. CEO Hult holds just 55,420 shares; incoming CEO Clara only 3,816; CFO Welch 9,122. This is a hired-management company, not an owner-operator — alignment rests on the comp plan, not on skin-in-the-game. The shareholder base is institutional/activist: BlackRock 11.0%, Impactive Capital LP 6.5% (an activist — variant-perception signal), Dimensional 5.7%, Eminence Capital 5.0%. Impactive’s presence as a top-5 holder suggests external pressure on capital discipline/portfolio optimization, which may partly explain the aggressive divestiture-and-buyback cadence.

Board. 11 directors, 10 independent (after DiSantis joined 1-Mar-2026). Reasonable governance housekeeping in flight — the 2026 proxy includes a proposal to eliminate supermajority charter provisions (a shareholder-friendly de-classification of voting thresholds).

SEC Filings Sweep & Insider Read

8-K material-event timeline (last ~2 years)

• 2023-12-11 — Jim Koons Dealerships acquisition (~$1.5B) closed.
• 2024-05-15 (8-K) — Board increased buyback authorization by $256.2M to $400.0M total.
• 2025-12-08 — CEO TRANSITION announced. David W. Hult to transition from President & CEO to Executive Chairman effective after the May-2026 Annual Meeting (“Transition Date”); Hult notified the Company 2025-12-04. Board elected Daniel E. Clara (COO) as incoming President & CEO, effective the same Transition Date. Clara, age 45, joined ABG in 2002 (24-year insider), COO since Feb-2025, prior SVP Operations 2020–25. Clara’s CEO compensatory arrangement was “not yet determined” at filing (to be disclosed by 8-K/A). (Source: 8-K 2025-12-08, Item 5.02.) — This is an internal-promotion, continuity transition, not a disruptive outside hire; the architect of the strategy (Hult) stays as Executive Chairman through ≥2027.
• 2026-02-04 (8-K, Item 5.02) — B. Christopher DiSantis appointed to the Board, effective 2026-03-01 (bringing the board to 11, 10 independent).
• 2025 (multiple) — Herb Chambers acquisition ($1,761.8M, debt-funded) and the Q4-2025/Q1-2026 divestiture program (13 stores, ~$750M annualized revenue) executed.
• Recurring quarterly earnings 8-Ks; 2025-05/2026-05 annual-meeting voting results (8-K).

Form 4 insider read (sampled across 2025-01 → 2026-05)

The corpus is dominated by routine, non-conviction codes: A (grants, e.g. Clara +4,532 sh on CEO appointment 2026-05-06; directors Reddin/Alsfine +932 sh each Feb-2026), F (tax-withholding on vesting — Welch, Briesemeister), G (gifts — Hult), and small S (discretionary sells by directors: Morrison 800 sh @ $255.61, James 625 sh @ $244.10, Calloway 400 sh @ $254.40, Milstein 1,132 sh @ $235.26 — all in 2025 at $235–256, none under a 10b5-1 plan, i.e. discretionary but small).

The lone conviction signal is genuine and is a CLUSTER OF TWO, not one. New director B. Christopher DiSantis made TWO open-market code-P PURCHASES:

• 500 shares @ $202.30 on 2026-03-11 (~$101k) — first transaction, made within ~10 days of joining the board (1-Mar-2026), alongside his initial grant.
• 157 shares @ $182.31 on 2026-05-22 (~$28.6k) — a second buy as the stock fell further toward the 52-wk low.

Both are discretionary (not 10b5-1). ~$130k combined is modest in dollars but directionally meaningful: it is the only open-market buying by any insider in the trailing 18 months, made by a brand-new director adding on weakness — the textbook “fresh eyes, buying the dip below book” signal. No other officer or director bought on the open market. Net insider activity otherwise = routine grants/withholding plus small director trims — i.e., mildly negative-to-neutral aside from the DiSantis cluster, which stands out precisely because everything else is non-conviction.

Insider read verdict: the DiSantis purchases are a real, if small, contrarian signal; they are not corroborated by any other insider buying, and they sit against a backdrop of very low aggregate insider ownership (<1%). Treat as a modest positive, not a thesis-maker.

8. Changes and Headwinds — Last Two Years

The trailing two years carry an unusually dense stack of strategic, operational and leadership changes — most corrective or constructive, but several stacking execution risk into 2026.

Strategic / M&A:

• Herb Chambers acquisition (closed Jul-2025, $1,761.8M): the largest deal since LHM, adding a New England luxury cluster (~$2.9B annualized revenue), lifting the luxury mix from ~32% to ~36% and the import/luxury skew that drives ABG’s GPU and service-dollar advantage. Debt-funded (floor plan + revolver + a $546.5M real-estate facility), it pushed transaction-adjusted net leverage from 2.9x (YE2024) to 3.2x.
• Active divestiture program: $566.5M of divestiture proceeds in 2025 (24 franchises) and ~10 stores + a collision center in Q1-2026, generating large book gains ($80.2M pre-tax in FY2025; $125.8M in Q1-2026). Management frames this as pruning low-return stores, funding deleveraging and buybacks, and upgrading the brand mix — a genuine capital-recycling discipline (and a recurring GAAP-EPS flatterer to strip when normalizing).
• Jim Koons (Dec-2023, $1.5B): the prior transformational deal (large MD/VA/DE metro group), fully in the base from 2024.

Operational:

• Tekion DMS migration (in flight): replacing CDK with Tekion across the network — a significant operational bet. >50% of stores converted by Q1-2026; full conversion targeted fall-2026; duplicate-system costs and a productivity dip peak in Q2-2026 (a ~$5M+ quarterly SG&A drag and a ~$0.56 Q1-2026 EPS headwind from weather/implementation combined). The thesis: a 2026 cost drag that becomes a 2027 structural SG&A step-down.
• New-vehicle GPU normalization (the dominant headwind): down ~38% from the 2022 peak and, per management, not finished — though the Q1-2026 view was raised toward “closer to $3,000.” Used GPU, by contrast, has risen for six of the last seven quarters on a deliberate profit-over-volume sourcing discipline.
• TCA deferral headwind: as the in-force F&I book grows, conservative ratable premium recognition is currently suppressing reported EPS (~$0.26–$0.31/quarter) — a near-term optical drag that builds a later-recognized income backlog.

Leadership / governance:

• CEO transition (announced 2025-12-08): David W. Hult (CEO 8.5 years) moves to Executive Chairman after the May-2026 Annual Meeting; Daniel E. Clara (COO since Feb-2025, a 24-year ABG insider who joined in 2002, age 45) becomes President & CEO. This is an internal-promotion continuity transition — the strategy architect stays on the board through ≥2027 with a declining-base sunset — not a disruptive outside hire (a favorable contrast to peer CarMax importing a hotel executive). Clara’s CEO comp package remained to be disclosed at filing.
• Board refresh: B. Christopher DiSantis joined the board 2026-03-01 (board now 11, 10 independent) and immediately made the only open-market insider purchases in 18 months. The 2026 proxy proposes eliminating supermajority charter provisions (shareholder-friendly).
• Activist presence: Impactive Capital holds 6.5% (top-5 holder) — plausibly part of the pressure behind the aggressive divestiture-and-buyback cadence.

Macro / sector headwinds:

• Affordability and rates: new-vehicle ASP >$52K and elevated financing rates are suppressing unit demand now; ~$2.6B of variable-rate debt (floor plan + real-estate term loans) means ~$26M of annual interest per 100bp — a two-sided rate sensitivity (relief if rates fall, pain if higher-for-longer).
• Tariffs: the FY2025 10-K added a risk factor on tariffs/trade restrictions on imported vehicles and parts — material given the 40% import mix.
• EV / direct-sales: EV-only makers continue to bypass franchise law in several states; EV adoption also structurally threatens the high-margin service annuity over the long run.

Verdict: on net, the changes strengthen the franchise but concentrate execution risk into 2026. The brand-mix upgrade (Herb Chambers/divestitures), the disciplined deleveraging, the counter-cyclical buyback, the continuity CEO handoff and the Tekion self-help are all constructive and corrective. The headwinds — GPU still normalizing, the Tekion disruption peaking mid-2026, affordability/rate pressure on volumes, and three execution risks (CEO + DMS + Herb Chambers integration) running simultaneously — are real and largely cyclical/transitional rather than structural. The one genuinely structural overhang (direct-sales reform) did not materially worsen in the period. This is a company doing sensible things into a cyclical earnings down-glide, not one whose thesis broke.

9. Risk Analysis

Catastrophic-loss / total-loss assessment. The risk of permanent capital impairment is low. ABG is profitable through-cycle (even the trough years earned ~$21–25 EPS), generates ~$570M of real FCF, carries no near-term refinancing wall, and its $2.03B floor-plan “debt” is self-liquidating against inventory. The real assets (owned dealership real estate carried at cost) provide a downside floor that divestiture multiples (well above book contribution) corroborate. A total loss would require a simultaneous, sustained collapse of new-vehicle GPU and volumes and the service annuity and a refinancing crisis — implausible short of a structural dismantling of the franchise system. The realistic downside is a cyclical one (earnings and multiple de-rate together toward the ~$120–150 bear zone, ~25–40% below the current price), not a permanent one.

10. Valuation Discussion

Market data as of 2026-06-11 close: ABG $199.48, ~18.6M diluted shares (Q1-2026 10-Q), market cap ~$3.71B. Multiples reconcile to the FY2025 10-K, Q1-2026 10-Q, and an own-history valuation-percentile index (pulled 2026-06-11). Peer multiples from third-party market data (2026-06-11) — reconciled to filings; treated as orientation, not authority. NO price target, NO buy/sell — this section is embedded-expectations and scenario work only.

10.1 Which multiple, and the EV trap

Auto-dealer valuation has one recurring error that must be cleared first: the enterprise-value figure most screens report is meaningless because it sweeps in floor-plan financing. Aggregator EV for ABG runs ~$9B, but that conflates ~$2.03B of self-liquidating floor-plan notes (inventory financing, offset by the inventory itself) and the TCA insurance float with genuine corporate leverage. The honest capital structure is market cap ~$3.71B + corporate net debt ~$3.55B = corporate EV ~$7.26B (the senior notes + real-estate term loans, net of de-minimis cash). All EV/EBITDA work below uses that corrected corporate EV, and the cleaner gauges are P/E and P/B, which sidestep the floor-plan distortion entirely.

The second error is valuing off TTM GAAP EPS of ~$28.24, which is doubly inflated: ~$4.95/share of one-time Q1-2026 divestiture gains and a new-vehicle GPU still ~$300–700/unit above management’s own stabilization band. The valuation has to be done on normalized EPS, not reported.

10.2 Where ABG trades — own history and the peer group

Own history (own-history valuation percentiles, vs. ABG’s own trailing ~10-year distribution):

The standout is P/B at the 1.25th percentile — ABG has essentially never been cheaper on book value in its public history, and it trades below book ($199.48 vs. BVPS ~$207). The composite sits in the bottom sixth of its own 10-year range. On its own past, ABG is priced for a bad outcome.

Peer group (franchised-dealer consolidators + used-only disruptors, third-party market data, 2026-06-11):

*EV/EBITDA “screen” uses the aggregator EV that includes floor plan — directionally useful for relative ranking across the franchised group (all are distorted the same way) but not an absolute multiple. ABG’s corrected corporate EV/adjusted-EBITDA is ~$7.26B / ~$1.08B ≈ 6.7x, the cheapest in the group on a clean basis.

The verdict from the table: on trailing P/E, ABG is the cheapest of the six franchised consolidators (7.05x vs. a 10–13x peer cluster, Sonic’s 26.6x distorted by a depressed earnings base). This is the central valuation puzzle the bull case is built on — ABG has historically run the highest dealership operating margin in the group yet trades at the lowest P/E. On forward P/E the gap narrows sharply (ABG 6.74x vs. GPI 6.84x, AN 7.99x, LAD 7.66x) because the Street’s forward estimates already bake in the GPU normalization and Tekion cost-out — i.e., the trailing-multiple “cheapness” is partly an artifact of above-normalized trailing earnings, exactly the warning. The franchised group as a whole trades at a steep discount (7–13x P/E, ~10–15x distorted EV/EBITDA) to the used-only disruptors KMX (31x) and CVNA (40x P/E, 22x EV/EBITDA), reflecting the market’s willingness to pay a large premium for the growth narrative of online used retail over the cash-generative, cyclical, ex-growth franchised model — even though the franchised dealers earn far higher returns on capital and throw off real free cash flow today.

Net positioning: ABG sits at the cheap end on every backward-looking metric and at the bottom of its own decade-long range on P/B. The question valuation must answer is whether that discount is (a) a value opportunity the market is missing, or (b) the market correctly capitalizing earnings that are about to fall.

10.3 The reverse-DCF / embedded-expectations read

At $199.48 on ~18.6M shares, the market cap is ~$3.71B. Against the honest normalized earnings base, the market is paying roughly 8–9x normalized earnings and ~0.96x book for ABG. What is embedded in that?

• On a simple capitalized-earnings basis: at ~8.5x a normalized ~$23 EPS, the price implies the market expects little-to-no real earnings growth — essentially a low-/no-growth cyclical that returns its earnings yield (~12% on normalized EPS) to holders via buybacks, with a modest haircut for cyclicality and the direct-sales tail risk. The market is not capitalizing the consolidation runway, the Tekion cost-out, or the TCA float build — those are zero-weighted or negatively weighted in the price.
• On a dividend/FCF-discount basis: ABG pays no dividend but generated ~$570M of FCF — a ~15% FCF yield on the $3.71B cap. For that yield to be “fair” rather than cheap, the market must believe FCF is at a cyclical peak and will fall materially (GPU reversion + recession) and that the buyback can’t durably compound per-share value. That is a coherent bear view, but it is an assumption, not a fact — and it is the assumption the 10.7% short interest is expressing.
• Implied terminal multiple: even a re-rate to a still-modest 10x normalized EPS (below where AN/GPI/LAD trade today on trailing earnings) on ~$23 normalized EPS implies ~$230/share before any buyback accretion — i.e., the current price embeds both full GPU reversion and a permanent below-peer multiple. The market is underwriting the bear-leaning end of the range as the base case.

10.4 Bear / Base / Bull on NORMALIZED EPS

All three scenarios strip the ~$4.95 one-time Q1-2026 divestiture gain and normalize new-vehicle GPU per management’s own framework; they then apply a sector-appropriate multiple band to produce an embedded-expectations value zone (not a target). Buyback accretion is treated as an explicit, separable lever.

Reading the scenarios. The current $199.48 sits at the bottom of the base zone / top of the bear zone — i.e., the market is pricing close to a bear-leaning base case (full GPU reversion to the middle of the band, no multiple re-rate, no credit for the buyback or consolidation runway). The asymmetry the bull leans on: the downside to a genuine bear (~$120–150) is ~25–40%, while the upside to a base/bull re-rate (~$250–360) is ~25–80%, and the base case alone — management’s own raised GPU view holding — supports the current price-to-modestly-above without heroic assumptions. The bear’s rebuttal: the multiple has been low for years for structural reasons (cyclicality, capital intensity, negative tangible book, direct-sales tail risk), so a re-rate is the least reliable leg, and a recession would hit volumes and GPU and P&S simultaneously.

10.5 Sub-book (P/B < 1.0x): justified discount or value signal?

ABG trades at 0.96x book (1.25th percentile of its own history). Two competing readings, both partly right:

• Why sub-book can be justified: (i) Tangible book is negative (~−$487M; goodwill $2,281M + franchise rights $2,098M > equity $3,892M). The “book” being discounted is ~$4.4B of acquired intangibles, and ABG has impaired ~$140M/yr of them for three straight years — the market is rationally skeptical that the franchise rights are worth carrying value. (ii) Earnings are above-normalized and ROE (~13%) is cyclically flattered; on normalized EPS, ROE compresses toward ~10–11%, and a business earning ~10% on equity with ~8.6% ROIC (barely above WACC) should trade near book, not at a premium. (iii) Capital intensity (inventory, real estate, floorplan) means book is “real” but low-returning.
• Why it can be a value signal: (i) The real estate ($3.07B net PP&E) is carried at cost and is likely worth more than book — owned dealership land in Sun Belt/luxury metros has appreciated; there is hidden asset value the divestiture multiples (sold at “attractive multiples” well above book contribution) corroborate. (ii) Buying back stock below book at a ~13% ROE is mathematically accretive to book value per share — every share retired at 0.96x book adds to remaining-holder book. (iii) 0.96x book on a business generating a 15% FCF yield is not a distressed multiple; it is a cheap one if earnings merely stabilize.

Net: the sub-book multiple is not a pure mispricing — negative tangible book and above-normalized earnings justify a discount to the historical average — but the 1.25th-percentile extreme overshoots what those structural factors warrant if management’s raised GPU-stabilization view (Q1-2026) proves correct. It is a discount that is partly earned and partly opportunity; the swing factor is, once again, where new-vehicle GPU settles.

10.6 The buyback math — the key bull lever, quantified

This is where sub-book + low-multiple + real FCF compound. At ~$199/share and ~$570M FCF, if ABG directed even ~$300M/yr to repurchases (it bought $147M in Q1-2026 alone, and $100M in all of FY2025 — the pace is accelerating, Q1-2026 call), it retires ~1.5M shares/yr against an 18.6M-share base — ~8% of shares outstanding per year. Mechanics:

• Per-share accretion: retiring 8% of the float lifts EPS ~8–9% with no operational improvement at all. On a flat ~$23 normalized EPS, a single year of that pace pushes EPS toward ~$25; two years toward ~$27 — entirely from share shrink. At a constant multiple, that is ~8–9%/yr of price appreciation manufactured from the buyback, on top of any operational or multiple-driven gains.
• Because it’s below book and at ~7x earnings, the accretion is doubly powerful: buying at 0.96x book grows BVPS; buying at a ~14% earnings yield (1/7x) is buying back the company’s own equity at a far higher return than reinvestment in 3–5x-EBITDA tuck-ins arguably offers — management explicitly framed Q1-2026 buybacks as exploiting a “price-to-value dislocation.”
• The constraint: leverage. ABG is at 3.2x (upper end of its 2.5–3.5x target) and has guided to <3.0x by year-end 2026. The buyback competes with deleveraging and tuck-in M&A for the same FCF + divestiture proceeds. The pace is therefore cyclical — it accelerates when the stock is cheap and leverage permits (as in Q1-2026) and pauses when leverage is tight. So the buyback is a powerful but not unlimited compounding lever; it is the single clearest mechanism by which the embedded-expectations gap could close even if the multiple never re-rates.

Verdict: priced for full GPU reversion with no credit for the buyback, consolidation runway, or annuity — cheap on normalized earnings if the bear case doesn’t fully land

ABG trades at ~7x trailing / ~8–9x normalized earnings, ~0.96x book (1.25th percentile of its own history), and the cheapest clean corporate EV/EBITDA (~6.7x) in the franchised-dealer group — despite running the group’s highest operating margin. The embedded expectation at $199 is a bear-leaning base case: new-vehicle GPU reverting through management’s band, no multiple re-rate, and zero value ascribed to the ~8%/yr buyback shrink, the multi-decade consolidation runway, the counter-cyclical P&S annuity, or the TCA float build. The discount is partly earned — negative tangible book, above-normalized trailing earnings, ~8.6% ROIC barely above WACC, capital intensity, and the direct-sales tail risk all justify a below-market multiple — but the extreme (near-cheapest-ever P/B, sub-book on a 15% FCF yield) overshoots if management’s raised Q1-2026 GPU view (“closer to $3,000”) holds. The scenario zones bracket ~$120–150 (bear) / ~$190–250 (base) / ~$280–360 (bull), with the current price hugging the bear/base boundary — a setup where the downside is real but bounded and the upside requires only stabilization plus the self-funding buyback, not heroics.

11. Variant Perception

11.1 Consensus

The consensus view on ABG (and the franchised-dealer group broadly) is “cheap cyclical, but cheap for a reason — earnings are at/above a once-in-a-generation GPU peak and have further to fall, the model is structurally threatened by direct-sales/online used, and you don’t pay up for a low-ROIC, capital-intensive, ex-growth roll-up no matter how low the multiple.” The Street prices it accordingly: a 7x trailing / ~7x forward P/E that converges with peers on forward estimates (i.e., the Street already models the GPU normalization), no multiple-re-rate baked in, and a sector-wide elevated short interest (ABG 10.7%, but LAD 16.6%, SAH 23.1%, PAG 13.5%, GPI 11.2%, AN 9.8% — see below). Consensus is neither euphoric nor distressed; it is “value-trap-aware” — willing to call it statistically cheap but unwilling to underwrite a re-rate.

11.2 The strongest BULL case

1. Cheapest of a group it out-margins. ABG trades at the lowest P/E and cleanest EV/EBITDA in the franchised six while historically earning the highest dealership operating margin (luxury/import mix + TCA capture + cost discipline). On any mean-reversion of the relative multiple toward AN/GPI (~11–12x), the re-rate alone is ~50%+.
2. The buyback compounds per-share value at sub-book. ~$570M FCF on a $3.7B cap funds an ~8%/yr share shrink, accretive to both EPS and BVPS because it’s executed below book and at a ~14% earnings yield. Management is leaning in explicitly ($147M in Q1-2026 alone) on the “price-to-value dislocation.” This manufactures high-single-digit annual per-share growth with zero operational improvement.
3. The annuity is underappreciated and counter-cyclical. ~71% of gross profit is P&S (59% margin, aging-car-parc tailwind, warranty/recall captivity) + F&I/TCA (93% margin, conservatively deferred — a hidden income backlog). TCA segment income was flat YoY in Q1-2026 while dealership income fell — proof of cycle-dampening. The market values ABG as a car-seller; it is mostly a service-and-finance annuity wrapped in a car-seller.
4. Self-help: Tekion cost-out. The DMS migration is a 2026 SG&A drag (duplicate costs, productivity dip) that flips to a structural SG&A/GP step-down in 2027 — a visible, company-specific earnings catalyst independent of the cycle (converted Koons stores already show +21% gross/technician, +16% advisor productivity).
5. Disciplined, long roll-up runway. A fragmented ~16,000±rooftop industry consolidating at 3–5x-EBITDA private multiples vs. ABG’s own ~7x public multiple — every accretive tuck-in is immediately value-creating, and the franchise-law moat protects the channel from greenfield over-build (favorable Marathon capital-cycle position).

11.3 The strongest BEAR case (what the 10.7% short interest is expressing)

1. Earnings are at a cyclical/GPU peak still reverting. New-vehicle GPU ($3,061 same-store Q1-2026) remains above management’s own $2,500–3,000 band; full reversion strips ~$5/share (~20%) off EPS, and TTM GAAP EPS is additionally flattered by ~$5 of one-time divestiture gains. The “7x P/E” is on a number that is going down — the value-trap signature.
2. Structural EV / direct-sales threat. Tesla/Rivian/Lucid already bypass franchise law in several states; if legacy OEMs ever win direct-sales reform, the channel’s regulatory moat — the entire reason the business earns acceptable returns — erodes. EV adoption also structurally shrinks the high-margin service annuity (fewer moving parts, less maintenance).
3. Rate-sensitive on both ends. Affordability (new-vehicle ASP >$52K) is suppressing volume now; floorplan + real-estate term loans carry ~$2.6B of variable-rate debt ($26M per 100bp). A higher-for-longer or recession scenario hits demand and financing cost simultaneously.
4. Debt-funded serial M&A on negative tangible book. Tangible equity is ~−$487M; ROIC is ~8.6%, barely above WACC; ~$140M/yr of recurring intangible impairments show some acquired franchise rights haven’t earned their price. The roll-up creates only thin value above cost — and leverage at 3.2x (upper end of target) limits the buyback when the stock is cheapest.
5. Execution/transition risk. A simultaneous CEO transition (Hult → Clara), a company-wide DMS migration (peak disruption mid-2026), and a large recent acquisition to integrate (Herb Chambers) — three execution risks stacked into one year.

Is the elevated short interest right? It is sector-wide, not an ABG-specific indictment — LAD (16.6%), SAH (23.1%), PAG (13.5%) all carry higher short interest than ABG’s 10.7%. The shorts are expressing a thesis on the franchised-dealer model (GPU peak + direct-sales threat + rate sensitivity), not a company-specific flaw at Asbury. On that thesis the shorts are directionally correct on the near-term earnings trajectory (GPU is still reverting; 2026 has real headwinds) but appear too pessimistic on the terminal value — they discount the buyback’s per-share compounding, the counter-cyclical annuity’s cycle-dampening, and the fact that the multiple already prices full reversion. The short case is a timing/level bet that earnings haven’t bottomed, not a solvency or franchise-impairment bet; the crowded, sector-wide positioning is itself a contrarian signal if GPU stabilizes at management’s raised (~$3,000) view rather than collapsing.

11.4 The 3–5 assumptions that matter most

1. Where new-vehicle GPU settles ($2,500 bear vs. ~$3,000 mgmt-raised base) — the single biggest swing on EPS.
2. Whether the P&S/F&I/TCA annuity holds and compounds (mid-single-digit P&S growth; TCA float build) — the cycle-dampener and the quality of the earnings.
3. Whether the buyback pace is sustained (FCF + divestiture proceeds directed to repurchase vs. consumed by deleveraging/M&A) — the per-share compounding engine.
4. Whether direct-sales reform spreads beyond EV-only makers to legacy OEMs — the binary tail risk to the entire moat.
5. Whether Tekion delivers the 2027 SG&A step-down — the company-specific self-help catalyst.

11.5 What would falsify each side

Falsifies the BULL: new-vehicle GPU breaks below $2,500 and keeps falling; P&S customer-pay growth turns negative on a sustained (not weather/DMS) basis; the buyback stalls because leverage stays >3.2x; a major state passes broad legacy-OEM direct-sales reform; Tekion savings fail to materialize in 2027 (SG&A/GP stays ~67%). Any two of these and ABG is a cheap business getting cheaper for good reason.

Falsifies the BEAR: new-vehicle GPU stabilizes at/above $3,000 for 3+ quarters (as management’s Q1-2026 view suggests); P&S resumes mid-single-digit growth post-Tekion; the share count visibly shrinks ~8%/yr; the multiple re-rates toward the AN/GPI ~11x as the GPU-peak fear fades; divestitures keep printing above-book multiples (validating hidden real-estate value). Any two of these and the value-trap label is wrong — it’s just value.

The crux: both sides agree ABG is statistically cheap; they disagree on whether the earnings base is peak (bear) or trough-ish/normalizing (bull). The evidence — management’s raised GPU guidance, six of seven quarters of rising used GPU, a counter-cyclical TCA, and an accelerating sub-book buyback — tilts the terminal-value argument toward the bull, while the near-term earnings trajectory (still-reverting GPU, 2026 Tekion drag, soft volumes) belongs to the bear. The market is pricing the bear’s trajectory as the terminal state — which is the variant-perception gap.

12. Fact vs. Interpretation

13. Open Questions

1. Where does new-vehicle GPU actually settle — $2,500 (bear) or ~$3,000 (management’s raised view)? The single biggest determinant of normalized EPS; unresolved until 3–4 more quarters of same-store data.
2. Does Tekion deliver the 2027 SG&A step-down? The +21%/+16% productivity figures at converted Koons stores are management-sourced and early; will not be confirmable in consolidated SG&A/gross-profit until 2027.
3. Is the TCA in-force reserve adequately funded? Conservative revenue deferral is established; claims/loss experience on the captive F&I book is not externally visible — a hidden risk on a $1B float.
4. What is incoming CEO Clara’s compensation package and any strategic shift? Not disclosed at filing; an internal-continuity hire suggests little change, but unconfirmed.
5. What is the true through-cycle normalized EPS once the COVID distortions, divestiture gains, and Tekion drag all clear — and how much of the 2021–22 GPU lift was structural (better mix, inventory discipline, F&I capture) versus a pure supply-shock windfall?
6. Will the buyback pace hold if leverage stays at the upper end and tuck-in M&A competes for the same capital? The per-share compounding thesis depends on it.
7. Hidden real-estate value: owned dealership land is carried at cost; the gap to market value (corroborated by above-book divestiture multiples) is real but unquantified in the filings.

14. What Must Be True (bull and bear, each with a falsification test)

BULL case — what must be true:

1. New-vehicle GPU stabilizes at/above ~$3,000 (management’s raised view), not $2,500 — so normalized EPS holds in the mid-$20s rather than collapsing to the high-teens.
2. The counter-cyclical annuity (P&S + F&I/TCA, ~71% of gross profit) keeps compounding mid-single-digit and dampens the vehicle cycle as designed.
3. The sub-book buyback continues at ~8%/yr share shrink, manufacturing high-single-digit per-share growth and accreting book value per share.
4. The consolidation runway and Tekion cost-out add incremental, accretive value; the multiple eventually re-rates toward the AN/GPI ~11x as the GPU-peak fear fades.

Falsification test for the bull: new-vehicle GPU breaks below $2,500 and keeps falling; or same-store P&S customer-pay growth turns sustainably negative; or the buyback stalls because leverage stays >3.2x; or a major state passes broad legacy-OEM direct-sales reform. Any two of these and ABG is a cheap business getting cheaper for good reason.

BEAR case — what must be true:

1. Current earnings are at a cyclical/GPU peak with material further reversion, so the ~7x P/E sits on a number that is going down (the value-trap signature).
2. The franchise-law moat is fragile — direct-sales reform spreads, and/or EV adoption structurally shrinks the service annuity.
3. The debt-funded roll-up creates only thin value above cost (~8.6% ROIC ≈ WACC; recurring intangible impairments; negative tangible book), so the low multiple is permanent and deserved.
4. A recession hits volumes, GPU and P&S simultaneously, overwhelming the annuity’s cushion.

Falsification test for the bear: new-vehicle GPU stabilizes at/above $3,000 for 3+ quarters; and/or the share count visibly shrinks ~8%/yr while divestitures keep printing above-book multiples; and/or the multiple re-rates toward peers as the peak-earnings fear fades. Any two of these and the value-trap label is wrong — it’s just value.

The crux: both sides agree ABG is statistically cheap; they disagree on whether the earnings base is peak (bear) or trough-ish/normalizing (bull). The evidence — management’s raised GPU view, six-of-seven quarters of rising used GPU, a counter-cyclical TCA, and an accelerating sub-book buyback — tilts the terminal-value argument toward the bull, while the near-term trajectory (still-reverting GPU, 2026 Tekion drag, soft volumes) belongs to the bear. The market is pricing the bear’s trajectory as the terminal state — that is the variant-perception gap.

15. Source Appendix

Primary — SEC filings (EDGAR, CIK 0001144980), mirrored locally:

• ABG Form 10-K, FY2025 (filed 2026-02-20, period 2025-12-31) — business, segments, MD&A gross-profit/GPU tables, debt note, acquisition note, revenue recognition, risk factors.
• ABG Form 10-K, FY2024 (2025-02-26), FY2023 (2024-02-29), FY2022 (2023-03-01), FY2021 (2022-03-01) — multi-year GPU, segment, and capital-structure history.
• ABG Form 10-Q, Q1-2026 (filed 2026-05-01, period 2026-03-31) — same-store GPU, divestiture gain, leverage, buyback.
• ABG DEF 14A proxy statements (2026-03-24, 2025-04-02) — compensation, incentive metrics, security ownership, board.
• ABG Form 8-K: 2025-12-08 (CEO transition), 2026-02-04 (DiSantis board appointment), 2024-05-15 (buyback authorization), plus M&A / earnings 8-Ks.
• ABG Form 8-K/A — Herb Chambers / acquisition pro-forma.
• ABG Form 4 insider filings (2025–2026) — including DiSantis code-P purchases (2026-03-11, 2026-05-22).

Primary — quantitative data:

• SEC EDGAR XBRL company facts (revenue, net income, operating income, diluted EPS, diluted shares, stockholders’ equity, goodwill, operating cash flow) — authoritative for US-filer financial series.
• Third-party market-data providers — price, market cap, EV, peer multiples (AN, LAD, GPI, PAG, SAH, KMX, CVNA); reconciled to filings, treated as orientation not authority.
• Market-data fundamentals (sector/industry/IPO/employees) and an own-history valuation-percentile index (P/E, P/B, P/S percentiles; composite 16.3rd percentile); third-party signal, reconciled to filings.

Primary — management commentary (treated as hypothesis, validated against filings):

• ABG earnings-call transcripts: Q1-2026 (2026-04-28), Q4-2025 (2026-02-05), Q3-2025 (2025-10-28), Q2-2025 (2025-07-29), Q2-2022 (2022-07-29) — GPU outlook ($2,500–3,000 stabilization, raised to ~$3,000 in Q1-2026), SG&A/Tekion commentary, capital-allocation framing, divestiture and buyback cadence.
• Financial news aggregators — Q1-2026 earnings reaction; DiSantis insider-buy item.

Secondary — peer/industry context:

• CarMax (KMX) and Carvana (CVNA) public filings and disclosures — used-car-retail industry structure, fragmentation, capital-cycle and direct-sales framing; context for the disruptor threat and peer valuation.

Analytical frameworks: Greenwald & Kahn, Competition Demystified (barriers to entry; moat taxonomy; market-share-stability and ROIC tests); Chancellor / Marathon, Capital Returns (supply-side capital-cycle analysis; asset-growth anomaly).

All non-obvious facts are cited to the above. Facts, interpretations, assumptions and open questions are labeled throughout. Management commentary is treated as a hypothesis requiring external validation, never as evidence on its own.

APPENDIX A — Standard Diligence Questionnaire

Asbury Automotive Group, Inc. (NYSE: ABG). Supplemental to the research memo; grounded in the same evidence base (FY2025 10-K, Q1-2026 10-Q, DEF 14A 2026-03-24, earnings transcripts, EDGAR XBRL). Fact / Interpretation / Assumption labels applied where material.

General

What thoughtful questions have other investors asked about this company? The dominant questions: (1) Where does new-vehicle gross-profit-per-unit (GPU) normalize, and is current EPS therefore at a cyclical peak? (2) Is the franchise-law moat durable against EV direct-sales and OEM pressure? (3) Can the parts & service + F&I/TCA annuity dampen the vehicle cycle enough to justify owning the equity through normalization? (4) Is the debt-funded roll-up creating value (ROIC vs. WACC) or just buying revenue? (5) Why does the lowest-multiple name run the highest operating margin in the group? These map directly to the memo’s variant-perception section.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: Above-normalized and still reverting. New GPU is ~38% off its 2022 peak ($5,583 → $3,061 same-store Q1-2026) but still above management’s $2,500–3,000 stabilization band; full reversion strips ~$5/share (~20%). TTM GAAP EPS (~$28) is additionally flattered by ~$5/share of one-time divestiture gains. Normalized EPS is the low-$20s. Driven by the external environment or internal actions? Both. The GPU cycle is external (supply-shock windfall deflating). Internal actions — used-GPU profit-over-volume discipline, M&A mix upgrade to luxury, the Tekion cost-out, and the buyback — are partially offsetting and improving the quality of the earnings base. How stable are revenues? Consolidated revenue is stable-to-growing (acquisition-driven); same-store revenue is flat-to-down as GPU normalizes. Parts & service (~14% of revenue, ~48% of gross profit) is the stable, counter-cyclical core; vehicle sales are the volatile swing. Outlook for products/services? New/used vehicle gross is normalizing; P&S grows with an aging, more-complex car parc; F&I/TCA is durable and a growing deferred-income backlog. How big will this market be? US franchised auto retail is a multi-hundred-billion-dollar, mature, GDP-ish-growth market, domestic, highly fragmented (~16,000+ rooftops) — shrinking in rooftop count (consolidating) but stable in dollars, with a long roll-up runway.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Interpretation: Stable-to-consolidating at the franchised layer (regulation-dampened, blocked greenfield entry); intensifying at the used/F&I layer (Carvana/CarMax, price transparency). How profitable is the business (ROIC, ROE)? ROE ~13–14%; ROIC ~8.6%, barely above WACC — ordinary, depressed by M&A-built goodwill/franchise rights. Fact (inputs) / Interpretation (WACC comparison). How profitable is the industry — competitors, barriers to entry? Low consolidated margins (~3% net) but acceptable returns thanks to franchise-law barriers; six public consolidators + thousands of private operators; entry barred by franchise law (high regulatory barrier). Can the business be easily understood? Yes — a distribution business wrapped around two high-margin annuities; the only subtlety is separating floor-plan from corporate debt and normalizing GPU. Can it be undermined by foreign low-cost labor? No — local, service-based, regulation-protected; immune to offshoring (though tariffs on imported vehicles/parts are a cost risk given the 40% import mix). Do brands matter? Yes — OEM brand mix (40% import / 32% luxury / 28% domestic) drives GPU and service-dollar economics; luxury skew is a margin advantage. ABG’s own brand is not a moat. What is the nature of competition? Local-market, per-brand exclusivity (franchise law) limits direct same-brand competition; competition is for acquisitions, for service retention, and on used-vehicle price. Customers’ switching costs? Low on vehicle purchase (price-transparent); meaningfully higher on warranty/recall service (must use a franchised dealer of the brand) — the captive-aftermarket advantage.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? Owned dealership real estate (~$3.07B net PP&E carried at cost) is likely worth more than book — hidden value corroborated by above-book divestiture multiples. TCA’s deferred-income backlog ($828M deferred premium) is a conservatively-recognized future-income asset. Off-balance-sheet liabilities? None material beyond standard operating leases; TCA insurance reserves are on-balance-sheet (ring-fenced float). Floor-plan notes ($2.03B) are on-balance-sheet but self-liquidating against inventory. How conservative is the accounting? Interpretation: Net conservative on the back end — TCA premium is deferred over policy life (suppressing current EPS); divestiture gains are correctly excluded from adjusted figures; ~$140M/yr intangible impairments are taken promptly. SBC is immaterial (~0.15% of revenue). How CapEx-hungry is the business? Moderately — capex ~$186M (2025), guided ~$250M (2026–27) on facilities + Tekion; plus large working-capital/inventory and real-estate intensity. Not a light-capital model, but FCF is healthy (~$570M).

Capital Allocation & Management

How much FCF, and how is it used? ~$570M FCF (2025); philosophy: maintain floor-plan engine → disciplined tuck-in M&A (3–5x EBITDA) → de-lever to <3.0x → counter-cyclical price-sensitive buyback → capex/Tekion. No dividend (appropriate for a sub-book recycler). Significant acquisitions recently? Yes — Herb Chambers (Jul-2025, $1.76B); Koons (Dec-2023, $1.5B); LHM+TCA (Dec-2021, ~$3.2B). Disciplined multiples, ~half hard-asset-backed. Buying back shares? Yes — ~17% share-count reduction (22.4M → ~18.6M); $267.7M/$183M/$99.9M (2023–25) + $147M Q1-2026; counter-cyclical (largest at lowest prices); $175.9M remaining authorization. Issuing large amounts of new shares to insiders? No — SBC immaterial; share count shrinking. The only equity issuance was the 2021 LHM deal. Compensation policy? LTI dominated by Absolute Adjusted EPS (70% of PSU) + relative EPS growth + TSR modifier; annual bonus on EBITDA gated by used same-store sales. Per-share-oriented (good) but no ROIC metric (the alignment gap for a debt-funded roll-up). CEO pay ~$10.7M (moderate, ~65% equity). Motivations of management? Hired professional managers (insider ownership <1%); alignment via the comp plan, not skin-in-the-game; activist Impactive (6.5%) provides external capital-discipline pressure. Internal-continuity CEO transition (Hult → Clara).

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a standard US C-corp, single share class, NYSE-listed, issues a 1099. Dividend policy? None — capital returned via buyback. How profitable is the business? Thin margins (2.7% net, 4.8% operating) but ~13–14% ROE; profit concentrated in the high-margin back end (P&S 59% GM, F&I 93% GM). Is net income diverging from cash from operations? OCF ($775M) exceeds NI ($492M) on non-cash D&A/impairments/SBC — no accrual red flag. Caveat: clean OCF (ex floor-plan-trade geography) has been flat-to-declining, masked by reported OCF; and divestiture gains inflate GAAP NI (strip them).

Risks & Downside

What factors would cause the stock to decline? New GPU breaking below $2,500; a recession hitting volumes + GPU + P&S; direct-sales/franchise-law reform; higher-for-longer rates; failed Tekion cost-out; buyback stalling on leverage. Risk of a catastrophic loss? Low — profitable through-cycle, ~$570M FCF, no near-term refi wall, self-liquidating floor plan, real-estate downside floor. Chance of a total loss? Very low — would require simultaneous collapse of GPU, volumes, the service annuity and a refinancing crisis, implausible short of a structural dismantling of the franchise system. The realistic downside is cyclical (earnings + multiple de-rate to ~$120–150), not permanent.

Recent News & Events

Has the business environment changed recently? Yes — GPU normalization ongoing (but management raised its stabilization view in Q1-2026); affordability/rate pressure on volumes; tariffs added as a risk factor. Significant acquisitions? Herb Chambers ($1.76B, Jul-2025), funded by debt; offset by an active divestiture program ($566M proceeds 2025). Change in accounting policies? None material; TCA deferral mechanics ongoing. Recent changes — new markets, facilities, management? New England entry (Herb Chambers); Tekion DMS migration network-wide; CEO transition (Hult → Clara, effective after May-2026 annual meeting); new independent director (DiSantis) who made the only open-market insider buys in 18 months.

APPENDIX B — Source Appendix

Primary — SEC filings (EDGAR, CIK 0001144980), mirrored locally:

• ABG Form 10-K, FY2025 (filed 2026-02-20, period 2025-12-31) — business, segments, MD&A gross-profit/GPU tables, debt note, acquisition note, revenue recognition, risk factors.
• ABG Form 10-K, FY2024 (2025-02-26), FY2023 (2024-02-29), FY2022 (2023-03-01), FY2021 (2022-03-01) — multi-year GPU, segment, and capital-structure history.
• ABG Form 10-Q, Q1-2026 (filed 2026-05-01, period 2026-03-31) — same-store GPU, divestiture gain, leverage, buyback.
• ABG DEF 14A proxy statements (2026-03-24, 2025-04-02) — compensation, incentive metrics, security ownership, board.
• ABG Form 8-K: 2025-12-08 (CEO transition), 2026-02-04 (DiSantis board appointment), 2024-05-15 (buyback authorization), plus M&A / earnings 8-Ks.
• ABG Form 8-K/A — Herb Chambers / acquisition pro-forma.
• ABG Form 4 insider filings (2025–2026) — including DiSantis code-P purchases (2026-03-11, 2026-05-22).

Primary — quantitative data:

• SEC EDGAR XBRL company facts (revenue, net income, operating income, diluted EPS, diluted shares, stockholders’ equity, goodwill, operating cash flow) — authoritative for US-filer financial series.
• Third-party market-data providers — price, market cap, EV, peer multiples (AN, LAD, GPI, PAG, SAH, KMX, CVNA); reconciled to filings, treated as orientation not authority.
• Market-data fundamentals (sector/industry/IPO/employees) and an own-history valuation-percentile index (P/E, P/B, P/S percentiles; composite 16.3rd percentile); third-party signal, reconciled to filings.

Primary — management commentary (treated as hypothesis, validated against filings):

• ABG earnings-call transcripts: Q1-2026 (2026-04-28), Q4-2025 (2026-02-05), Q3-2025 (2025-10-28), Q2-2025 (2025-07-29), Q2-2022 (2022-07-29) — GPU outlook ($2,500–3,000 stabilization, raised to ~$3,000 in Q1-2026), SG&A/Tekion commentary, capital-allocation framing, divestiture and buyback cadence.
• Financial news aggregators — Q1-2026 earnings reaction; DiSantis insider-buy item.

Secondary — peer/industry context:

• CarMax (KMX) and Carvana (CVNA) public filings and disclosures — used-car-retail industry structure, fragmentation, capital-cycle and direct-sales framing; context for the disruptor threat and peer valuation.

Analytical frameworks: Greenwald & Kahn, Competition Demystified (barriers to entry; moat taxonomy; market-share-stability and ROIC tests); Chancellor / Marathon, Capital Returns (supply-side capital-cycle analysis; asset-growth anomaly).

All non-obvious facts are cited to the above. Facts, interpretations, assumptions and open questions are labeled throughout. Management commentary is treated as a hypothesis requiring external validation, never as evidence on its own.