Verdict: KEEP

The strongest remaining factor mismatch case in the selectable set. Negative SPY beta, anti-momentum, -20.3% trailing alpha. On paper, this should be a remove. In practice, the market is pricing the opposite of our thesis, the edge is zero-IC index construction arbitrage, and adding a 7th correlated factor bet deepens concentration without improving portfolio Sharpe.


Factor Profile

250-day regression (iev regress, 3 factors):

FactorBetaInterpretation
SPY-0.23Moves opposite to market. Unique among 50 selectable names.
MTUM-0.34Anti-momentum. Laggards stay lagging.
XLC+0.95Communication services sector proxy.

Idiosyncratic variance: 86% (well above 75% target). R-squared: 14%. Trailing alpha: -20.3% annualized. Sigma_idio: 25.1%.

The negative SPY beta is the standout. Every other factor mismatch remove in the basket (MAR, CMCSA, CSX, SBUX, CTAS) has positive SPY beta with anti-momentum tilt. TMUS moves against the market entirely. In a QQQ rally, TMUS acts as dead weight pulling the basket down. Combined with MTUM beta = -0.34, this is the single worst factor profile for QQQ filtration.

QQQ beta = 0.10 (from prior systemic analysis). The decomposition explains why: XLC loading (+0.95) carries it slightly positive; SPY loading (-0.23) drags it negative. These forces roughly cancel, leaving near-zero QQQ sensitivity.

Fundamentals (Strong, Known, Irrelevant)

FY2025 (10-K filed Feb 11 2026): Total revenue $85.5B (+8% YoY). Core Adj EBITDA $33.9B (+7%), 48% margin. FCF $28.0B (+25%). Net income declined 3% to $11.0B on UScellular merger costs ($263M), higher interest (+$363M), JV losses (Lumos, Metronet), and $278M impairment. Underlying business grew; the net income miss is M&A accounting noise.

2026 guidance (Investor Day, Feb 11 2026): Service revenue ≈$77B (+8%). Core EBITDA $37-37.5B (+10%). FCF $18-18.7B. CapEx $10B. 900K-1M postpaid net adds. $5B accelerated Q1 buyback. DT not selling in 2026, "exploring alternatives to deepen investment." Below-the-line charges: ≈$1.2B UScellular merger costs, ≈$450M network optimization (H1), ≈$150M workforce restructuring (Q1).

Q4 EPS miss: $1.88 actual vs $2.06 estimate (-8.9%). Broke 3-quarter beat streak. Driven entirely by merger/restructuring charges, not operational weakness. Q1 2026 pre-guided $9-9.1B core EBITDA.

Capital return since 2022: $45B+. Annual buyback target ≈$10B. 2026-27 envelope: up to $52B ($30B returns + $22B flexible).

Every analyst knows all of this. 29 covering, 76% bullish. The fundamentals are the reason the stock is cheap at 15x forward P/E. They are not the reason to remove it. The question is not "is TMUS good?" but "will TMUS keep QQQ pace over 15 weeks?"

CEO Transition

Sievert retired as CEO Nov 1 2025, became Vice Chairman. Srinivasan Gopalan (Deutsche Telekom executive) became President & CEO after 8-month overlap as COO starting Mar 1 2025. DT is majority shareholder; this is DT installing their operator, not a hostile succession. Gopalan led strategic vision at Feb 11 Investor Day, Osvaldik (CFO) led financials. Orderly.

Sievert's $20.5M stock sale (Feb 18-19 2026) was as Vice Chairman, not sitting CEO. Reclassified as routine post-transition liquidation (LR neutral). Active-officer selling: CFO Osvaldik $5.8M (Feb 18) + GC Nelson $4.2M (Feb 24) = ≈$10M. Notable but modest against the $5B accelerated buyback backdrop.

What the Market Is Actually Pricing

This is where the remove thesis breaks down.

Analyst consensus: 29 analysts. Mean target $268.68 (+27% upside, 12 months). 22 bullish, 7 hold, 0 sell. Pro-rated to 15 weeks: ~+7.9% expected return. The +27% target is nearly 2x the typical analyst auto-markup (+15%), indicating genuine conviction in undervaluation.

Options positioning at our horizon:

ExpiryTMUS P/C (OI)QQQ P/C (OI)Read
Apr 17 (19d)0.65 bullish--Call-heavy pre-earnings
May 1 (33d)0.73 bullish--Bullish through earnings
Jun 18 (81d)1.19 neutral1.90 bearishHedging QQQ, not TMUS
Aug 21 (145d)0.33 very bullish--3:1 calls. Market expects recovery.

The critical signal: at the Jun/Aug horizon that brackets our 15-week window, the market is hedging QQQ downside (P/C 1.90, nearly 2:1 puts to calls) while positioning for TMUS upside (P/C 0.33, 3:1 calls to puts). Market consensus is TMUS outperforms QQQ over our window. Our short would be against consensus.

Implied earnings move (Apr 28): Extracting from term structure (Apr 24 exp IV 35.2% vs May 1 exp IV 41.9%), earnings-day implied move = +/-8.4%. The market hasn't forgotten Q4's -8.9% miss. Management pre-guided $9.0-9.1B core EBITDA (tight range = internal confidence).

Implied volatility: TMUS ATM IV 34.2% (Jun 18) vs QQQ ATM IV 29.6% = +4.6% spread. Market prices TMUS as more volatile than QQQ but without directional fear: OTM put skew +29% (moderate, not extreme), no heavy OTM put accumulation.

Volume vs OI divergence: Near-term put volume running bearish (P/C volume 2.52 at Apr 17) against longer-term bullish OI positioning. Tactical hedging or earnings protection, not structural bearishness.

Max pain: TMUS Jun 18 at $220 (+4.4% above current). QQQ Jun 18 at $600 (+6.7% above current). Both above spot, implying the option dealer equilibrium pulls prices higher.

Edge Assessment

SignalLRNote
Factor mismatch: negative SPY beta, anti-momentum0.70Index construction arbitrage. Real but zero IC.
Trailing idio alpha -20.3%0.85Backward-looking. Mean-reversion risk at 15x P/E, RSI 37.7, $5B buyback.
Active-officer insider selling ≈$10M0.90Modest vs $5B buyback. CFO + GC, Feb 2026.
CEO transition 5 months old0.90Orderly, DT-backed. Weak signal at this point.
Board resignation (Mar 2026)1.00Routine. No signal.
Fundamentals1.00Strong, improving, fully known by 29 analysts.
Options positioning (Jun/Aug)1.15Market positioned for TMUS recovery. Against our remove thesis.

Combined LR: 0.70 x 0.85 x 0.90 x 0.90 x 1.15 = 0.55

Mildly bearish vs QQQ, driven entirely by the factor mismatch. Remove the factor signal and the remaining evidence is neutral-to-bullish (combined LR ≈0.69 from idio alpha and insiders, offset by 1.15 from options positioning = ≈0.79).

Counterparty: QQQ's index methodology. Invesco tracks NASDAQ-100 mechanically. TMUS is included because it's a large NASDAQ-listed company, not because it's a tech stock. No informed seller on the other side. The edge is clean but has zero informational content.

What Is Mispriced?

Nothing, in the informational sense. The factor mismatch is public data. The fundamentals are fully covered. The CEO transition is orderly and disclosed. The options market is pricing the recovery thesis and hedging QQQ downside.

What exists is an index construction inefficiency: QQQ holds TMUS mechanically despite negative market beta. But exploiting this requires QQQ to rally, which the options market isn't pricing (P/C 1.90 bearish at Jun 18). The "mispricing" reduces to a directional market bet: we profit if QQQ goes up, lose if QQQ goes down. That's beta, not alpha.

Portfolio Concentration Problem

Current removes split into two bets:

Bet 1 (Factor mismatch, 6 removes, zero IC, correlated): MAR, CMCSA, CSX, SBUX, CTAS, CEG. Total removed weight: 4.83%. Essentially one bet: "tech outperforms defensives/anti-momentum."

Bet 2 (Stock-specific, 4 removes, positive IC, diversified): CRWD, VRTX, WBD, MELI. Total removed weight: 5.04%. Genuinely independent theses.

Adding TMUS takes Bet 1 from 6 to 7 removes (4.83% to 6.10%). All 7 lose simultaneously in a QQQ selloff or momentum reversal. Effective breadth of Bet 1 remains ≈2-3 (anti-momentum is one factor, CEG is a separate utility/energy bet). Adding TMUS increases anti-momentum loading without adding a new independent risk factor.

Marginal Sharpe contribution of another correlated zero-IC factor bet: approximately zero. The alpha contribution (3-8 bps base case, conditional on QQQ direction) doesn't compensate for the concentration drag.

Risk Register

RiskProbImpactNote
QQQ selloff >5%25%Short loses; neg beta = TMUS tailwindLong QQQ hedges, but all Bet 1 removes bleed
Momentum reversal15%All 7 factor removes lose simultaneouslyAlready the dominant risk in the basket
Q1 beat + re-rating40%TMUS rallies 5-10%, short loses 6-13 bpsSingle-quarter event, factor profile unchanged
Idio alpha mean-reverts50%Alpha contribution drops from ≈8 to ≈3 bps15x P/E + $5B buyback = real reversion risk
DT tender/privatization<5%Short loses ≈38 bps at 1.27% weightLow probability, asymmetric, survivable

Verdict: KEEP

The factor profile is the worst in the selectable set for QQQ filtration. If this were the first factor mismatch remove, the case would be strong. But it's the 7th, at zero IC, correlated with the other 6, conditional on a QQQ rally the options market isn't pricing, and opposed by market consensus at every relevant horizon.

Adding TMUS generates 3-8 bps expected alpha in the base case while deepening concentration in a correlated factor bet that is already the dominant risk in the basket. The marginal Sharpe contribution is approximately zero.

The honest framework says: if you want more alpha from this basket, the work is in Bet 2 (stock-specific weakness in the 22 uncovered names), where IC > 0 and diversification lives. Not another factor mismatch remove at zero IC.

Kill conditions: If QQQ positioning flips bullish (P/C < 1.0 at Jun expiry) AND basket Bet 1 exposure drops below 4 removes, TMUS becomes the strongest marginal add. Revisit if market regime changes.

Evidence

  • Factor regression 250d: SPY -0.23, MTUM -0.34, XLC +0.95, idio 86%, alpha -20.3%
  • 2026 guidance: $77B service rev, $37-37.5B EBITDA, $18-18.7B FCF, $5B Q1 buyback
  • Insider selling ≈$10M active officers (Feb 2026), $20.5M Sievert as Vice Chairman
  • CEO transition Nov 1 2025: Sievert to Vice Chairman, Gopalan (DT) to President/CEO
  • Options: Jun 18 P/C 1.19 (neutral), Aug 21 P/C 0.33 (very bullish 3:1 calls)
  • QQQ Jun 18 P/C 1.90 (bearish hedging)
  • Q4 EPS miss -8.9%, implied Q1 earnings move +/-8.4%
  • 10-K FY2025. 8-Ks reviewed through Mar 27 2026. Q3/Q4 earnings transcripts.