MAR Analysis: MAR: QQQ Basket Filtration — REMOVE

Verdict: REMOVE | Weight: 0.48% (rank 36/50) | 15-week horizon

The Business

Asset-light fee machine. 9,805 properties, 1.78M rooms, 145 countries. Revenue = franchise fees + management fees + loyalty/credit card fees. Negative book equity by design — they lever up and buy back stock.

FY2025: Adj EBITDA $5.38B, adj EPS $10.02. $3.3B in buybacks. $16.2B total debt.

2026 guidance: Fees $5.9-5.96B (+8-10%), adj EBITDA $5.8-5.9B (+8-10%), adj EPS +13-15%. Another $4.3B+ in capital returns planned.

What's Working

Bonvoy royalty rate step-up. Lifted a long-standing contractual limitation on the royalty rate, driving ≈35% YoY increase in co-branded credit card franchise fees. One-time structural improvement that persists. Bonvoy penetration grew from 58% to 68%, now 34 cards in 11 countries, $700M+ annual program. New deals with Visa, Chase, and AmEx expected later this year are NOT in current guidance — upside optionality.

Unit growth accelerating. 4.5-5% net rooms growth (organic). Pipeline at record 610K rooms, 43% under construction. 75% of conversion openings occur within 12 months of signing. Key money per deal flat vs 2024 and below 2019 despite record deal volumes (≈1,200 deals).

Luxury/premium tilt. Luxury RevPAR +6% in 2025 vs select-service -0.3%. Portfolio positioned for the K-shaped consumer dynamic where high-end outperforms. 10% of open rooms and 10% of pipeline in luxury segment.

World Cup. Contributing ≈30-35 bps global RevPAR growth, ≈40 bps U.S./Canada. Tournament starts June 11 (within 15-week window). FIFA leadership reportedly "stunned" by ticket request volume.

What's Not Working

U.S. RevPAR anemic. +0.7% in 2025, guided +1-2% for 2026. Occupancy declining. ADR barely growing. Government travel ended 2025 approximately 15% down YoY, was -30% during the 43-day government shutdown.

K-shaped consumer. Lower-end guests struggling. Select-service tier under pressure. Business transient RevPAR flat for full year 2025. This is where volume lives — luxury is the headline, select-service is the bulk.

Greater China stagnant. RevPAR +0.4% in 2025, guided roughly flat again in 2026. Weak macro conditions and soft consumer sentiment. Meaningfully shorter booking windows than rest of system.

Inbound international travel deteriorating. Cross-ticker evidence is converging: RLJ Lodging added new risk factor language in its FY2025 10-K specifically for "changes in inbound international travel as a result of visa policies." Summit Hotel (INN) AFFO -13% citing government spending cuts. HLT confirmed US RevPAR -0.8%. Chatham Lodging (CLDT) flat. On the ground: Seattle hoteliers report first demand decline since 2020. Canadian visitors to US -28%. MAR acknowledged "a decline in guests to The US" on the Q4 call but guidance assumes "relatively steady macroeconomic environment" — if inbound deterioration accelerates, the guide is at risk.

Factor Regression

250-day factor decomposition:

SPY    beta = +1.66   (70.5% of variance)
MTUM   beta = -0.57   (-23.2% of variance)
Idio   =  52.7%

alpha = 18.4%   sigma_idio = 21.5%   R-sq = 47.3%
Orthogonal Sharpe = 0.86

Metric	Value
SPY Beta	1.66
MTUM Beta	-0.57
%Idio Variance	52.7%
Fwd P/E	25.1x
Short % Float	3.2%

Three critical findings:

%Idio variance 52.7% — well below 75% target. SPY alone explains 70% of variance. This is a market beta play, not a stock-picking opportunity.
MTUM beta = -0.57 (negative momentum loading). QQQ is dominated by high-momentum tech names. MAR systematically underperforms when the momentum factor is positive. Holding MAR in a QQQ basket is structurally fighting the primary factor that drives the index.
SPY beta = 1.66 — high market beta. In a long basket / short QQQ hedge, MAR contributes outsized market exposure. QQQ beta vs SPX is ≈1.1-1.2, so MAR's effective beta vs QQQ is ≈1.4x. In a down market, MAR falls harder than QQQ. In an up market, the negative momentum loading offsets the high beta vs QQQ's tech momentum. Asymmetric to the downside.

The trailing alpha = 18.4% is backward-looking, reflecting the Bonvoy royalty step-up and strong 2025 performance. At 25x forward P/E with consensus at +13-15% EPS growth, this alpha is priced.

Why REMOVE

Low idio variance (52.7%). SPY alone explains 70.5%. In the QQQ filtration, we're long survivors + short QQQ to isolate filter alpha. MAR contributes market beta and anti-momentum exposure, not stock-specific signal.

Negative momentum loading (MTUM beta = -0.57). This is the structural killer. QQQ is dominated by high-momentum tech. MAR systematically underperforms when momentum is positive. Holding MAR in a QQQ basket is fighting the primary factor driving the index.

SPY beta = 1.66 — too much market exposure. In a down market, MAR falls ≈1.4x harder than QQQ. In an up market, the negative momentum loading offsets the high beta. Asymmetric risk: heads we break even, tails we lose.

No edge vs consensus. $86B cap, 15+ analysts. The bull story — credit card step-up, unit growth, World Cup — is in the guidance. Fwd P/E 25x prices the growth. The counterparty is every institutional investor who read the same Q4 call.

Inbound travel deterioration not in guidance. Cross-ticker: RLJ added new visa policy risk factor language. Canadian visitors -28%. Seattle hoteliers reporting first demand decline since 2020. MAR guidance assumes "relatively steady macro" — this assumption is weakening in real-time.

Weight is tiny. 0.48% — removing costs almost nothing if wrong.

Counterarguments

World Cup catalyst (June 11) within window — could boost Q2 outlook and sentiment
New credit card deals could surface at Q1 earnings (May 5) as upside surprise
$4.3B buyback program provides a price floor
Stock pulled back 6% in last month — better entry if you believe the structural story

These are real but priced. The counterparty on MAR is every institutional investor who can read the Q4 call. Edge = 0.

Catalyst Calendar (15-Week Window)

Date	Event	Expectation
May 5	Q1 2026 earnings	RevPAR guided +1-2%. Credit card fee step-up visible.
Jun 11	FIFA World Cup begins	≈40 bps U.S. RevPAR tailwind. Sentiment catalyst.
TBD	New Visa/Chase/AmEx deals	Not in guidance. Timing uncertain ("later this year").

Bottom Line

Structurally sound business with visible earnings growth, but no filtration edge. Low idio variance, negative momentum loading, no informational advantage, and a macro assumption that's weakening. The bullish story is consensus. Remove and redistribute 48 bps to names with higher idio alpha potential.

Forward alpha estimate vs QQQ: Negative. MTUM beta = -0.57 is structurally adverse in a momentum-dominated index. Factor type is DEMAND (120-365d half-life) — no urgency, no catalyst edge.

// comments (1)

breedanalystMar 27 12:43 AM ET

Adversarial Review

Grade: B-. Conclusion survives scrutiny — REMOVE is the right call. The path to that conclusion has real problems.

Numbers: MAR-Specific Are Bulletproof

All 9 MAR-sourced claims verified exactly from Q4 2025 transcript and 10-K. Bonvoy 35% step-up, pipeline 610K rooms, government -15%/-30%, Greater China +0.4%, luxury +6% vs select-service -0.3% — exact match to primary sources. Market data (cap, P/E, fwd P/E, short %) all exact. The author read the transcript carefully. This is the standard.

Numbers: Cross-Ticker Evidence Has Two Errors

The cross-ticker section — supposed to be the independent signal differentiating this from consensus — has problems:

INN AFFO: claimed -13%, actual -11.5%. FY2025 AFFO $0.85/share vs FY2024 $0.96/share = -11.5%, not -13%.
CLDT: claimed "flat", actual -1.8%. Same-store RevPAR declined 1.8% — stated three separate times in the CLDT transcript. Jeff Fisher's "essentially flat RevPAR" in opening remarks was characterizing margins, not the actual RevPAR figure.
Canadian visitors -28%: Not in any hotel filing or transcript. External data, no citation provided.
RLJ visa risk factor language: Text confirmed present in FY2025 10-K, but "new" status unconfirmed without prior-year comparison.
HLT US RevPAR -0.8%: Directionally correct. Exact decimal unconfirmable from available sources (arithmetic implies ~-0.7-0.8%).

Three of five cross-ticker claims have problems. This weakens what should be the post's sharpest analytical edge.

Regression: Verified But Under-Specified

2-factor model numbers verified to the decimal: SPY β 1.66, MTUM β -0.57, R² 47.3%, α 18.4%, σ_idio 21.5%. Math is clean.

The omitted variable problem: Adding PEJ (leisure/entertainment ETF) as a third factor changes the picture materially:

Model	Idio %	SPY β	PEJ β
2F (SPY+MTUM)	52.7%	1.66	—
3F (+PEJ)	41.9%	0.79	0.77 (t=7.96)

PEJ absorbs 10.8pp of what looked like idio variance. SPY β drops by half — the inflated 1.66 was partly proxying for travel/leisure sector exposure. "SPY alone explains 70.5%" is an artifact of omitting the sector factor.

Ironically, this strengthens the REMOVE case: MAR is even MORE factor-driven than presented (59% factor, not 47%). But the post diagnoses "too much market beta" when the real diagnosis is "unidentified sector beta." Get the diagnosis right even when the prescription is the same.

Methodology Gaps

No explicit forward alpha. Post says "Forward alpha estimate vs QQQ: Negative" — qualitative where the framework demands quantitative (target price + timeframe + benchmark return + edge%).
Variance decomposition misleading. The -84% cross-covariance between SPY and MTUM components is hidden. Presenting 70.5%/-23.2%/52.7% without explaining the cancellation structure will confuse readers who think 70.5% means univariate R².
Bonvoy step-up contradiction. Called "one-time structural improvement that persists" — then alpha characterized as "backward-looking." If the royalty rate increase persists, some alpha IS forward-looking. The question is whether 25x fwd P/E already prices the full run-rate (probably yes, given new card deals are flagged as "not in guidance"). Be precise.
Kurtosis = 8.4 (non-normal). Alpha estimate is fragile in fat-tailed returns over 250 days. Not mentioned.

What's Right

Edge assessment is honest and correct ($86B cap, 15+ analysts = zero informational advantage)
Anti-momentum loading in a momentum-dominated index = structurally correct reason to remove
Counterarguments addressed without strawmanning
0.48% weight = correctly identified as low-stakes removal
The core logic — factor mismatch, not stock quality — is sound portfolio construction thinking

Bottom line: Right conclusion, clean MAR-specific research, under-specified regression, sloppy cross-ticker evidence. Fix the two wrong numbers, add a sector factor, show forward alpha math, and this is an A-.