Agency versus bank debt for stabilized multifamily

Compare an agency and bank proposal in two passes: first establish that each applies to the actual property, then normalize the accepted cash flow and spendable proceeds. An attractive calculation under an inapplicable product is not a financing alternative.

“Agency” is too broad for the comparison sheet. Record the named program, the lender presenting it, the source version, and the unresolved eligibility facts. The example below uses a hypothetical proposal intended for Freddie Mac Conventional Small alongside a hypothetical bank loan. Its financial terms are invented; they are not Freddie Mac’s required ratios, leverage, or rates.

Begin with one factual property file

Assume a 40-unit apartment property with a supported $10 million value and $650,000 of annual property cash flow before the lender-specific adjustments below. Both lenders receive the same rent roll, collections, operating statements, expense evidence, and ownership description on the same date.

The initial submission describes fee-simple ownership. That description remains an item to verify against the title and land documents, not a fact established by the sponsor’s spreadsheet. The acquisition price and other property uses do not change between offers.

Freddie Mac’s April 2026 Conventional Small summary, page two, lists ground leases among ineligible characteristics. It also states that the summary does not replace the Seller/Servicer Guide and that terms can change. This is a specific product screen, not a claim that every Freddie Mac or Fannie Mae product excludes ground leases.

Reconcile the different accepted cash-flow amounts

Suppose Bank B deducts $20,000 for its stated expense adjustment, accepting $630,000. Agency proposal A deducts $50,000 for its stated expense and replacement-reserve adjustments, accepting $600,000. Those deductions are hypothetical and must be reconciled line by line rather than labeled merely “conservative.”

Assume B tests 1.30 times coverage using an 8.00% annual debt constant and a 65% value limit. A tests 1.25 times coverage using a 7.50% annual constant and a 75% value limit. The annual constants represent annual scheduled debt service divided by original principal for this sizing exercise; they are not note rates.

The calculations use unrounded inputs. B is $630,000 ÷ 1.30 ÷ 0.08. A is $600,000 ÷ 1.25 ÷ 0.075. Coverage binds both examples. Other constraints, including product eligibility, can still reduce or eliminate the result.

A accepts less property cash flow but supports more debt because its assumed coverage and payment factors differ. Comparing only the two cash-flow deductions would miss that interaction. Conversely, substituting A’s constant into B’s accepted income would create a third loan neither lender offered.

Translate the note amounts into acquisition cash

Assume B withholds $100,000 of lender-controlled reserves and charges $60,000 in financed or deducted closing costs. A withholds $350,000 and charges $100,000. These amounts are separate from the recurring annual adjustments already used in underwriting; the reserve deposits are closing cash uses, not another annual operating expense.

A’s $342,308 gross advantage becomes approximately $52,308 of additional acquisition funding. This is a closing-availability comparison, not a conclusion that all reserve deposits are permanently lost. Their later use or release belongs on the holding-period cash schedule.

Reopen the comparison when a property fact changes

Now suppose counsel identifies a ground lease that the initial submission omitted. Do not simply add a reserve line and leave A’s $6.4 million result in the approved column. The specific product screen has changed.

The bank is not automatically eligible simply because the agency summary excludes the characteristic. Its collateral, lease term, consent, and underwriting requirements also need review. Until both lenders respond to the corrected fact, the prior numerical ranking is conditional and may no longer describe an available pair of loans.

The revised agency alternative might have different cash-flow treatment, reserves, documents, timing, or expense. Preserve a version history so that old favorable terms do not survive invisibly in the new comparison.

Turn the result into a decision record

Use three separate status columns: fact confirmed, lender accepted, and financially modeled. A title document can confirm a fact without establishing lender acceptance. A lender may accept an asset type while still conditioning its commitment on reports and final underwriting. A completed model establishes neither fact by itself.

Before choosing between the two available proposals, compare amortization, maturity, prepayment, guarantees, reserves, and the planned ownership changes over the same hold. The $52,308 closing difference is small enough that another actual term could reasonably dominate the decision; no fee or rate assumption here determines that outcome.

The completed exhibit should therefore preserve both conclusions: under the stated financial assumptions A supplies modestly more acquisition cash, while a ground-lease finding would require an eligibility reset before that result could be relied on. A program name is useful only when attached to the correct property facts and an identifiable, current lender proposal.