The sponsor's request and the collateral are enough to pre-fill this — but a stress/scenario ask like this always gets one form so you can correct terms and attach documents before I run the numbers.

I found The Modern Apartments (3201 Annrae St, San Diego 92123 — 369 units, built 1959/renovated 2020) and pulled its in-place financials and current agency quotes to pre-fill the loan terms below. If you have an actual term sheet or T12, attach it — the loan terms below are current market quotes, not necessarily what's being offered.

The Modern Apartments — 70% LTV acquisition request fails DSCR before the stress test even starts

DECLINE at the requested $103.3M / 70% LTV — restructure to ≈$76.6M (52% LTV) to clear. The debt yield test binds, not a market assumption: at Day 1, before any rent-growth or cap-rate stress is applied, this loan covers debt service at only 0.96x 1, already below breakeven. Layering in the bear-case rent forecast and a 50 bps exit-cap widening doesn't create the problem — it turns a thin, deteriorating position into an outright refinance failure at maturity.

The ask

Acquisition financing on The Modern Apartments, 369 units in San Diego's Kearny Mesa submarket 2, at 70% LTV — sized here against a lender value of $147.5M 3, producing a request of $103.26M at 6.21% (Fannie Mae Conventional, 10-yr, 75% LTV tier, quoted 9/2/26), 30-year amortization, non-recourse.

Coverage — the deal doesn't clear before you touch a single forward assumption

The haircut is real, not conservative modeling: Kearny Mesa's submarket vacancy runs 7.27% — nearly double this property's T12 actual vacancy of 3.9% 6 — and the submarket is absorbing a heavy 2026 delivery wave (Broadstone Soren, Alexan Camellia, Centrum Phase 6 — several 400+ unit projects within the submarket) 7, so the higher vacancy floor is the right lens, not a haircut for haircut's sake. At the $76.55M counter-offer, DSCR recovers to 1.29x and debt yield lands exactly at the 9.5% floor 5 — that's the number a lender can actually support here.

The stress test you asked for

I ran the RealAI forecasting engine off 26 months of San Diego in-place rent history: base-case rent growth averages +1.5%/yr over a 10-year hold, bear-case averages -2.6%/yr — the engine flags the read as low-confidence because the whole history sits inside one post-2020 regime with no multi-cycle anchor 8. Run through the model as the NOI growth assumption, alongside the requested 50 bps exit-cap widening at the takeout:

• Base case (rent growth ≈ base-case forecast, no incremental cap widening): DSCR climbs from 0.97x in Year 1 past breakeven in Year 3 (1.00x) to 1.09x by Year 10 9. At maturity, NOI grows to $8.43M, the stressed takeout still clears — but by a $128K surplus, 0.15% 9. That's not a cushion, that's a coin flip.
• Bear case (bear-case rent growth + 50 bps cap widening): DSCR never reaches breakeven and declines every year — from 0.93x in Year 1 to 0.76x by Year 10 as NOI falls to $5.75M 10. At maturity, the stressed takeout lender (cap widened to 5.43%) can only support $57.79M against a $86.89M balance — a refi gap of ($29.10M), a 33.5% shortfall 10. The refinance does not clear.

The rent-growth assumption is the swing factor — Day-1 basis and the balance schedule don't move with it, but the entire maturity outcome flips from "clears by a hair" to "$29M short" on that one input.

If it defaults

Recovery cushion is deep on value alone — even at a 30% value decline, net recovery proceeds ($101.2M) still exceed the workout balance, so loss severity is 0% and lender recovers 100% of the outstanding balance 11. But recovery as a % of the original $103.26M par is only 84.2% 11, below the 90% threshold — a function of how little the loan amortizes by year 10, not of the rent scenario. Basis cushion (30%) protects the lender; it doesn't protect the sponsor's equity.

Structure — what would make this financeable

• Proceeds: cut to $76.55M (52% LTV) — sized to the 9.5% debt-yield floor, the binding constraint
• Pricing/amortization: unchanged (6.21%, 30-yr) — the fix is proceeds, not term
• Covenant: DSCR test with cash-sweep trigger given how thin the base-case refi surplus is even at reduced proceeds
• Sponsor: no financials on file — verdict caps at Approve with Structure pending guarantor review 12 regardless of the sizing outcome

Bottom line

At 70% LTV / $103.26M this fails on Day 1 — decline as requested. The sponsor's real ask should be $76.55M, where the deal clears the debt yield floor with a 1.29x DSCR. Even there, don't take comfort in the base case's refinance math: it clears by 15 basis points of surplus, and the bear-case rent forecast alone is enough to turn that into a $29M shortfall. The single diligence item most likely to change this verdict: a signed rent roll showing in-place rents already tracking above the submarket's stressed vacancy assumption, which would tighten the vacancy floor and lift lender NOI closer to the T12 figure.