The rent roll on a stabilized small multifamily reads clean until the lender starts counting holdover tenants toward vacancy risk.

Offering new leases before the lender call is standard advice, but the term you choose changes the picture: a 12-month lease signed 11 months before closing still reads better than month-to-month, but a 24-month term with a rent escalation clause gives the lender a forward income argument that offsets their vacancy haircut.

The mechanism you are describing is real and it catches a lot of portfolio owners off guard.

On the holdover point: a month-to-month lease is technically occupied, but many lenders treat it as soft occupancy because the tenant can leave on 30 days notice. Their underwriting model converts that uncertainty into a vacancy penalty, and on a four-unit building the numbers move quickly. If two units get a 25 percent haircut applied to their rent, you lose the equivalent of half a unit's income before the loan is even sized. Debt service coverage ratio (DSCR) is the lender's way of checking whether the property earns enough to cover its mortgage payment with room to spare. A common threshold is 1.25, meaning the property needs to earn $1.25 for every $1.00 of debt payment. Losing income from two haircut units can drop that ratio below the line and shrink what the lender will approve.

On the expense ratio point: lenders apply their own assumed cost percentage to the gross income, often somewhere in the 35 to 45 percent range, because they cannot verify that an efficient self-manager will stay efficient after the loan closes. A seller who has genuinely kept costs lean does not get credit for that in most underwriting models.

The practical prep step you named is correct. A current lease expiration schedule and a documented capital expenditure log are the two documents that do the most work before a lender conversation starts. Renewing month-to-month tenants to 12-month leases before going to market can change how a rent roll reads.

I would confirm the specific vacancy treatment with each lender directly, because it does vary by institution and loan program.

What is the mix of lease terms across your units right now?

The assumption doing the most work in your framework is the 25 percent vacancy haircut on month-to-month units. That figure is not universal, and the spread across lenders is wide enough to matter. Some agency-aligned products treat a current occupant on a month-to-month as equivalent to a signed lease if the tenancy has run continuously for 12 months or more and the rent is current. Others apply the haircut mechanically regardless of tenure. Before you restructure lease schedules around underwriting, it is worth confirming exactly which vacancy treatment your lender uses, in writing, because the deal math changes substantially depending on which model applies.

The expense ratio point is the one I would push harder. Owner-managed fourplexes often run actual expenses at 28 to 32 percent, and a lender floor of 40 percent on a portfolio of three or four buildings is not a rounding error. That delta can suppress the loan proceeds enough to make a cash-out refinance or acquisition financing structurally unworkable even when the actual cash flow would support the debt. The operator has no mechanism to argue the lender out of their floor ratio using trailing actuals alone. The lever that sometimes moves this is a documented third-party management agreement showing market-rate management costs, which gives the lender a credible external benchmark. Whether that applies to a self-managing operator is a question for the lender directly.

The CapEx log recommendation is correct and often underused. A system-by-system capital schedule with dates and amounts shifts the reserve conversation from the lender's assumed worst case to an actual depreciation position, and that changes the reserve requirement the lender books, which feeds back into DSCR.

One risk your post does not address: lease renewal concentration. If multiple leases across the portfolio expire in the same 60-day window, you face simultaneous re-leasing exposure at exactly the moment you are trying to demonstrate stabilization.

How are the current lease expirations distributed across the calendar year, and is there clustering?

Month-to-month isn't the only flag. I've seen lenders ding a 12-month lease signed 10 months ago almost as hard.