SL Green's 6.50% Series I Cumulative Redeemable Preferred ($25 liquidation preference, perpetual, 9,200K shares outstanding) yields 7.83% at $20.76. Vornado's comparable preferreds — same sector, same structural seniority, arguably weaker coverage — yield 7.10-7.52%. The gap has persisted while VNO preferreds returned +15-20% over the past year and SLG-PI moved +3.5%. Today's 8-K gives a reason to look again.

What the filing says

Q1 2026 earnings, filed 2026-04-16 (Items 2.02, 7.01). The operating picture:

• 929,264 sf leased — highest Q1 in the company's 28-year history. Average starting rent $105.12/sf (all-time quarterly high). Mark-to-market on replacement leases +16.1%.
• Occupancy 94.4%, up from 93.0% Q4 2025. Company guides 95.0% by year-end.
• Same-store cash NOI +2.6% YoY (excluding lease termination income).
• Harvey AI expanded to 185K sf at One Madison (building now 100% leased). Clay AI 163K sf at 11 Madison. Company discloses 80+ tech tenants seeking 8M+ sf, with 13 AI requirements over 200K sf each.

The credit picture:

• Fixed Charge Coverage 2.20x vs 1.50x minimum (47% headroom). The covenant definition includes preferred dividends in fixed charges.
• Total Debt/Assets 47.7% (max 60%). Secured Indebtedness 27.8% (max 50%). Unencumbered Leverage 44.9% (max 60%). All four covenants in compliance with meaningful room.
• One Madison Avenue refinanced with $1.65B SASB CMBS at 5.81% (181 bps over Treasuries), replacing the $1.171B construction facility. This was the company's largest development risk; it's now permanently financed.
• $2.0B of the $2.4B corporate credit facility extended to June 2031 at 25 bps lower cost.
• Asset sales: 690 Madison Avenue closed in February ($48.5M cash to SLG), 7 Dey Street residential/retail components under contract for $222.6M (Q2 2026 expected close).

Preferred-specific:

• Series I dividend of $0.40625/share declared and paid in cash April 15. Share count unchanged at 9,200K. Annual cost ≈$14.95M against reaffirmed FY 2026 FFO guidance of $4.40-$4.70/share (midpoint ≈$349M): 23x FFO coverage.
• Common dividend reduced to $2.47/share annual (announced March 23) to preserve liquidity. Preferred paid at full rate. Subordination hierarchy working as designed.

What the market thinks

SLG-PI at $20.76, yield 7.83%. RSI 51.5. 18% of 52-week range. 1-year total return +3.5%.

Comparable preferreds:

If SLG-PI traded at VNO-PM's yield, price would be $21.96 — +5.8% from current, plus 7.83% carry = ≈11% total return to year-end. VNO-PM RSI at 92 is stretched; mean reversion from the short side could narrow the gap as well.

SLG common ($41.13, RSI 71, analyst HOLD, consensus PT $52) has partially recognized the leasing picture: +9% 1M, but still -16.9% 1Y. Common momentum is crowded. The preferred has not re-rated.

The gap, numerically: SLG-PI pays 43 bps more than VNO-PM for exposure to the same sector. The price level is consistent with a market that either views the spread as structural (permanent yield premium) or expects convergence only on a slow, uncertain timeline. It is not consistent with a market that views near-term convergence as the central case.

Why the gap exists

Four non-exclusive reasons:

1. Filing-level coverage analysis is not on retail yield screens. Preferred-stock flows cluster around yield-ranked ETFs (PFF, PGX) and income-focused retail. Coverage ratios inside the Series I supplemental disclosures require reading the 10-Q/supplemental; yield screens don't capture 2.20x Fixed Charge Coverage vs 1.50x minimum.

2. SLG-specific credit signaling. The company cut its common dividend to $2.47/share annual (announced March 23) to "retain incremental liquidity." Interpreted conservatively, this reads as management signaling balance-sheet caution — and some preferred holders may be conflating common-dividend stress with preferred-dividend risk. The filing shows the opposite mechanism (preferred held firm while common absorbed the cut), but the initial read of a dividend cut is rarely granular.

3. Liquidity discount. SLG-PI float is ≈$191M (9.2M × $20.76) vs VNO-PM ≈$356M. Thin tape gets less institutional flow and wider bid-ask.

4. Timing of the rerate narrative. VNO reported Q4 2025 in late January and framed the sector recovery explicitly. BXP and ESRT followed. SLG's record Q1 2026 print is today. The preferred market has had 2-3 months to price VNO; SLG-PI hasn't had time yet. If this is the main explanation, convergence should begin within weeks, not months.

The NYC office rerate itself is not idiosyncratic to SLG: Vornado CEO Roth called it the "foothills of best market in 20 years" on the Q4 call. BXP's Doug Linde said "Midtown NY...tightest supply, most landlord-favorable market conditions." Commercial Observer reports >500K sf of AI leasing in NYC in the first months of 2026. SLG is doing the premium-tier execution within a sector that is genuinely re-rating. Investment-grade office CMBS has reopened — SLG's $1.65B One Madison issuance is the largest US office CMBS in 12 months.

Risks

1. Rate backup. Perpetual preferred duration ≈12-15 years. A +50 bp move in 10Y repriced the whole preferred complex; 25-35% of SLG-PI variance is rate-driven. This overwhelms the spread story. No edge on rates.

2. Mamdani tax enacted. Property tax pass-through hits NAVs even if leasing stays strong. Bearish for all NYC office REITs. JLL data shows no current operational impact, but this is a latent tail risk that compresses multiples.

3. VNO preferreds mean-revert hard (RSI 92 → 50). Spread narrows the wrong way — VNO-PM drops toward SLG-PI rather than SLG-PI rising toward VNO-PM.

4. $100M Nov 2026 and $300M May 2027 term loans. Asset sale proceeds (≈$271M) plus $143.9M cash plus revolver capacity cover the 2026 wall. The 2027 overhang is not addressed in this filing.

5. SLG-PI call feature bounds upside. Series I is cumulative redeemable preferred at $25 liquidation preference. Whether it is currently within call protection or callable at par is not established in the Q1 8-K; the prospectus and prior 10-Ks would resolve this. If callable at par, SLG could redeem if it refinances at below 6.50% — a $25 ceiling caps price upside but is itself +20.4% above current, so redemption is a favorable outcome, not a risk. The relevant scenario is that the market prices call risk, limiting how far above par-level the price can run.

6. Coverage drift. If NOI declines or interest expense climbs, Fixed Charge Coverage heads toward 1.50x. Not the central case given the leasing trajectory, but worth monitoring quarterly.

Catalysts

• July 2026 — Q2 earnings 8-K. Covenant check, Q2 leasing volume, occupancy trajectory.
• Q2 2026 — 7 Dey Street closing. Visible ≈$222.6M cash; removes execution discount.
• November 30, 2026 — $100M term loan maturity. Refinanced/repaid cleanly removes near-term overhang.
• December 31, 2026 — YE occupancy print. 95% guide achieved or missed.
• February 2027 — FY 2026 FFO print. $4.40-$4.70 range holds or breaks.

Most of the information flow concentrates July-November. If the thesis works, most of the repricing happens by the end of Q3.

What would change our mind

• Q2 2026 8-K shows a covenant slip (Fixed Charge Coverage below 1.80x) or any secured-debt spike — credit trajectory has reversed.
• 7 Dey Street transaction falls through or delays beyond Q3 — management execution in question; VNO-type catch-up unlikely without proven deleveraging.
• Company announces Series I redemption — upside capped at $25 par (which is actually a good outcome, +20.4% from current, but a different thesis).
• Mamdani's administration passes a property tax bill targeting commercial real estate — multiple compression across the sector; SLG-PI rises less than VNO-PM falls.
• VNO-PM retraces more than 8% without SLG-PI moving — spread compression occurring the wrong way; pricing anomaly being resolved by VNO correction rather than SLG catch-up.

Evidence