Verdict: FILTER (70/30)

Remove SBUX from the QQQ filtration basket. Redistribute 59bps across survivors. Reassess after Q2 FY2026 earnings April 28.

The regression is unambiguous: trailing alpha = -16.3% annualized, idiosyncratic variance = 62% (below 75% target), anti-momentum loading in a momentum-heavy index. The Niccol turnaround is real on the top line — first dual-segment transaction growth in 8 quarters — but the market pays 31x forward for earnings that haven't arrived. Four consecutive EPS misses. Operating margins at 9%. The inflection is a September event. Our 15-week window captures peak margin compression, not the recovery.

The Factor Problem

250-day regression against SPY and MTUM:

Alpha: -16.3% annualized. R-squared: 38%. Idiosyncratic vol: 26.8%.

Three problems compound. First, -16.3% alpha means SBUX is destroying value after removing market and momentum exposures — not rounding error over a short window but 250 trading days of persistent underperformance. Second, 62% idiosyncratic variance sits below the 75% target — over half the variance comes from market beta at 1.58, meaning a 10% market drop translates to ≈16% from beta alone before negative alpha kicks in. Third, the -0.45 momentum loading is systematic drag. QQQ IS momentum — its top holdings are the winners that define the factor. Holding an anti-momentum name in a long-QQQ/short-QQQ basket creates factor exposure that works against you in every scenario where index constituents maintain their trend.

What the Market Prices at $91 / 31x Forward

The 31x forward P/E — identical to QQQ itself — implies NTM EPS of ≈$2.95. Management guides FY2026 at $2.15-$2.40, so the market is pricing FY2027 at roughly $2.95, requiring 23-37% EPS growth. That growth requires four things simultaneously: Green Apron anniversaries lifting ≈260bps of margin drag, tariff and coffee costs normalizing, the $2B cost reduction program showing up in G&A, and comps sustaining at +3-4%.

Each assumption is individually 60-70% probable. All four simultaneously: 13-24%. The market prices the conjunction as likely. That's the mispricing.

Analyst consensus: 23 analysts, Buy, average PT $100.83 (+10.5%). But the March direction is negative — Wolfe Research downgraded to Peer Perform (Mar 9) citing proprietary traffic data showing deterioration near Dutch Bros and 7 Brew locations, and RBC Capital downgraded to Sector Perform (Mar 17) with a $105 PT, arguing labor investments are permanent and cost savings smaller than anticipated. Zero upgrades in March.

The sell-side Q&A reveals where attention lives. On the Q1 call (Jan 28), JPMorgan asked about competitive threats from drive-thru focused chains. Bank of America asked whether additional investments beyond the $500M labor model are coming — and CFO Kathy Smith's answer was telling: "want to make sure got right flexibility" and "not broad-based cost-cutting." That's corporate-speak for "we might spend more." Barclays pivoted entirely to unit growth — the long-term story — which tells you the Street has mentally written off the next 2-3 quarters. Nobody asked about tariffs because the March escalation hadn't happened yet. That's an undigested variable.

Insider Behavior

Five Form 4 filings in March. Three are noise (code F — automatic tax withholding on RSU vests). Two are signal:

Brady Brewer (CEO, International) sold 588 shares at $100.00 on March 9, plus 1,641 shares via a Rule 144 filing. His 10b5-1 plan was adopted December 3, 2025 — one week before the China JV announcement, two months before Investor Day. He knew the calendar and set up the plan to sell restricted stock that vested November 10. Small relative to his 86K share holding, but the timing of plan adoption matters.

Sara Kelly (EVP, Chief Partner Officer) sold 2,500 shares at $97.12 on March 5 — about 4% of her 60K holding. No 10b5-1 plan cited. Discretionary sale.

Both sold $6-9 above today's price. Zero insider buying. In a turnaround where management is publicly confident, nobody is buying stock with their own money.

Options Market Confirms the Skew

OPRA data on the May 1 expiration (34 DTE, spanning April 28 earnings) shows a 6-point put-call IV skew at the money — puts price ≈42% IV versus calls at ≈36%. The ATM straddle at ≈$9.10 implies a total move of roughly 10%, of which approximately 7% is earnings-specific after stripping non-earnings volatility.

The skew is the tell. A 6-point put premium on a $104B name isn't panic — it's informed hedging. The options market sees more downside risk than upside heading into Q2. The stock at 31x says "turnaround works." The options at +6 put skew says "maybe not this quarter."

The Earnings Asymmetry (April 28)

The scenarios where SBUX outperforms QQQ require the turnaround to accelerate ahead of schedule. The scenarios where it underperforms require only that the current trajectory continues. Betting on "trajectory continues" is the higher-probability outcome.

Q2 consensus sits at $0.41 EPS — management guided it as the trough quarter. The real question isn't whether they beat $0.41 but whether they can credibly guide H2 at $1.18-$1.43 (needed for $2.15-$2.40 FY2026). That H2 bridge was set BEFORE March tariff escalation (46% Vietnam, 15% global) and BEFORE the Rewards relaunch backlash. Both are new headwinds the Street hasn't fully modeled.

The Margin Walk

Operating margins at 9.0% vs. 11.9% prior year. The 10-Q is explicit about the decomposition:

Two-thirds of margin compression is labor, one-third is tariffs and coffee. Both persist through our 15-week window. The Green Apron anniversary — the single largest margin recovery lever — doesn't hit until Q4 (September quarter). We're long gone by then.

The Turnaround: Real but Mismatched to Our Window

Credit where earned. The Niccol playbook is showing results:

• First dual-segment transaction growth (Rewards + non-Rewards) in 8 quarters
• Global comps +4%: transaction-driven (+3pp), not ticket
• 650 pilot stores outperform fleet by ≈200bps, all from transactions
• Rewards at record 35.5M (+3% YoY)
• Channel Development +19% (Global Coffee Alliance + RTD)

Investor Day (Jan 29) laid out FY2028 targets: 13.5-15% operating margins, $3.35-$4.00 EPS, 2,000+ net new stores. The path from 9% to 13.5% requires flawless execution on labor efficiency, cost reduction, same-store leverage, and tariff normalization — simultaneously. Market was unimpressed: stock fell on Investor Day. These targets are reasonable on a 2-year horizon but don't help in 15 weeks.

Where I Could Be Wrong

Oversold bounce. RSI at 33 with the stock at 54% of its 52-week range. Mean-reversion in the next 2-4 weeks before earnings could temporarily outperform. But RSI mean-reversion on a name with negative alpha is a bet on noise.

Surprise positive catalyst. China JV closes with better economics, tariff exemption announced, Q2 blowout on Rewards relaunch success. Each individually ≈5-10% probable, collectively ≈15%.

Niccol acceleration. His Chipotle playbook was faster than expected. If the 650 pilot store outperformance scales to the full fleet before September, the margin bridge accelerates. The market is giving him 12-18 months of benefit-of-the-doubt — that grace period hasn't expired.

What Would Flip This to KEEP

1. Q2 margins inflect (April 28). If management shows margin expansion before Green Apron anniversary, cost reduction is pulling forward. That breaks the negative alpha trend.
2. Tariff rollback. If coffee/China merchandise tariffs are reduced, the $100M+ headwind shrinks and back-half relief materializes on schedule.
3. Rewards relaunch works. If Q2 comps accelerate despite backlash, the new program is broadening the funnel as designed.
4. Insider buying. Any C-suite open market purchase would be a strong counter-signal.

None of these are base case. Track for reassessment after April 28.

Edge and Counterparty

This is a $104B name with 23 analysts. The counterparty is every informed participant in the market. Edge on SBUX is structurally small.

P(SBUX underperforms QQQ over 15 weeks): Market ≈50%. Our estimate ≈62%. Edge ≈ 12%.

Three structural reasons SBUX stays overvalued relative to QQQ: index inclusion creates forced buyers (every dollar into QQQ buys 59bps of SBUX regardless of fundamentals), dividend yield at 2.72% anchors income-seeking funds who won't sell until the dividend gets cut, and the Niccol halo from Chipotle creates benefit-of-the-doubt pricing that takes 12-18 months to expire. These are the uninformed counterparties — constrained, anchored, or narrative-driven.

Expected filtration alpha from removing SBUX: +3-6bps over 15 weeks. Small in isolation. Additive across the basket.