COMP filed its FY2025 10-K on February 27. The filing itself is a historical document — it describes the OLD Compass. Nine days into the new fiscal year, Compass closed the Anywhere Real Estate acquisition and became a fundamentally different company. Everything that matters about COMP now is in the subsequent events footnote, not the income statement.

What the Filing Shows

Standalone Compass was genuinely improving. Revenue $7.0B (+24%), Adj EBITDA $293M (4.2% margin, up from negative two years ago), operating cash flow $217M. Commission expense ratio tightened from 82.3% to 81.6%. G&A down to 2.1% of revenue. Three straight years of improving cash flow: -$26M, +$122M, +$217M. The business was working.

Then they bet the house.

The Anywhere Deal

All-stock acquisition of Anywhere Real Estate (formerly Realogy): Coldwell Banker, Century 21, ERA, Sotheby's International Realty, Corcoran, BHGRE. Exchange ratio 1.436x, ≈162M new shares (21.8% dilution). Combined entity: 340,000+ agents, #1 US brokerage by volume.

The financing is the problem. Compass assumed $2.15B of Anywhere senior notes at 6.94% WA (maturing 2029-2030) and issued $1.0B in convertible notes (0.25%, conversion at $15.98). Total gross debt: $3.15B. Annual interest on the senior notes alone: ≈$149M — versus $9M of interest expense in all of FY2025. That's a 16x increase in interest burden overnight.

Combined pre-synergy EBITDA is ≈$640M (Compass $293M + Anywhere ≈$350M guidance). Leverage: ≈5x pre-synergy, ≈3.4x if you believe the full $300M synergy target. Still elevated, but not the 11x figure you'd get using Compass standalone EBITDA — and not existential unless housing crashes AND synergies fail AND the 2029-2030 refi window closes.

The Realogy Parallel

These are literally the same brands. Apollo acquired Realogy (Coldwell Banker, Century 21, Sotheby's) for $8.5B in April 2007. Leverage was similar. Then housing crashed. $3B+ in cumulative losses. Near-bankruptcy. Apollo bought back debt at 10 cents on the dollar, restructured $1.95B in bonds, IPO'd in 2012, doubled its money by 2013.

The critical difference: Apollo bought at the PEAK. Compass is buying at the TROUGH. Mortgage rates stuck at 6%, housing volume depressed, existing home sales at 4.1M (vs 6M+ normalized). If rates decline and volume recovers, leverage works in reverse — EBITDA grows, ratios compress. If housing stays flat, the $149M/yr interest bill grinds.

The pattern isn't "leverage = death." It's "leverage + bad cycle = death." Compass has better cycle timing but identical structural fragility.

Commission Compression: The Dog That Didn't Bark

The 10-K flags NAR settlement commission compression as "highly uncertain." Corroboration says otherwise. Federal Reserve data (May 2025): commissions declined ≈30bps over 25 YEARS. Redfin Q3 2025: buyer's agent commissions actually ROSE from 2.36% to 2.42%. Industry survey: combined buyer+seller commissions up from 5.32% to 5.44% in 2025.

The NAR settlement was supposed to destroy brokerage economics. Eighteen months in, commissions are flat-to-up. Behavioral inertia, buyer agreement friction that protects agent pay, and low transaction volume all working against compression. Long-dated risk, not imminent.

The Regression Kills It

This is where it gets interesting. Default factor regression (SPY, MTUM, XLRE) shows 88% idiosyncratic variance. Looks great. But that model is mis-specified — it's missing the two factors that actually drive this stock.

Add EXPI (direct brokerage peer) and XHB (homebuilders):

COMP 2yr regression:
EXPI    24.4% of variance  ← dominant peer beta
XHB      4.1%              ← housing cycle
SPY      2.6%              ← market
XLRE     1.2%              ← REIT (minor)
Idio    67.7%              ← company-specific

68% idio. Fails the 75% threshold. Nearly a quarter of COMP's return variance is explained by what eXp Realty does on any given day. Another 4-13% is housing cycle. Combined: ≈32% of your "COMP bet" is really a brokerage/housing sector bet you could get cheaper through EXPI or XHB.

The 68% idio component decomposes into latent factors: integration execution, debt refinancing, agent retention, DOJ risk, synergy realization. Every one of these requires deep real estate brokerage expertise that 11 covering analysts have and we don't.

Insider Activity: Zero Signal

yfinance flagged CEO Reffkin "acquiring" 1.15M shares (≈$11M) on February 4. Pulled the actual Form 4. Transaction code M: RSU vesting at $0.00. Code F: tax withholding sale of 639K shares at $11.98. Pure compensation event, not conviction buying.

CFO Wahlers is more telling. His pattern across every Form 4 since November: vest (code M), withhold taxes (code F), sell the rest the next day (code S). $12.03 on January 7. $11.00 in December. He's liquidating every share the moment it vests. Not panic selling — but not holding either.

Zero insider buying at any level. The CFO doesn't want to own this stock.

DOJ Overhang

The deal closed over DOJ staff objections. Compass went over antitrust division head Gail Slater to Deputy AG Todd Blanche. Capitol Forum analysis shows >80% combined market share in Manhattan and Newport Beach, 60-70% in San Francisco, Denver, Boston, DC. Senators Warren and Wyden raised alarm.

The deal is closed. Unwinding a closed merger is nearly unprecedented. But consent decrees, behavioral remedies, or state AG actions could constrain synergy realization in concentrated markets — which happen to be Compass's highest-margin markets.

Verdict

$6.9B market cap. 11 analysts, 73% bullish, mean target $14.09 (+48%). Stock at $9.49 after -22% in a month (but beta is 2.57, so half that move is market sensitivity). RSI 33. ATM IV at 128.7% (234th percentile) — options market pricing extreme uncertainty.

The bull case is real: if integration works, synergies materialize, and housing recovers, this could be a $15B+ company. Consensus already sees this.

The bear case is real: $3.15B in debt on deteriorating brands, $149M/yr interest eating cash flow, franchise integration at 340K-agent scale never attempted, CFO selling, DOJ overhang.

We assign 35% probability to the stock hitting $14 by year-end (vs consensus ≈73% bullish), and 45% probability COMP hits $700M combined Adj EBITDA for FY2026.

Pass. Not because the thesis is bad. Because we have zero edge. The idio component (68%) is in a domain — real estate brokerage operations — where 11 analysts with sector expertise are already doing the work. The factor component (32%) is brokerage/housing beta we can get elsewhere. The insider signal is empty. The filing documents a well-known deal.

If you want housing volume recovery: ZG and STC are purer plays. If you want asset-light brokerage: EXPI. If you want the leveraged integration bet: you need to know more about Coldwell Banker franchise health than the analysts do. We don't.

First real test: Q1 2026 combined earnings, May 7. That's when the pattern collapses.