The Rent Roll That Lenders Model as Three Different Portfolios: How In-Place Tenant Creditworthiness, Renewal Probability Cohorts, and Replacement Rent Scenarios Diverge Between Debt-Yield Underwriting and Loss-Mitigation Reserve Sizing

The Structural Problem: One Rent Roll, Three Competing Models

A lender receives a rent roll on day one of underwriting. It contains tenant names, unit counts, in-place rents, lease expiration dates, renewal options, and credit ratings. The document appears uniform, objective, and complete. But over the next 90 days, three separate underwriting teams model the identical rent roll as three fundamentally different portfolios, each with its own assumptions about probability, timing, and replacement cash flow. These models never reconcile. They coexist inside the loan agreement, each driving a different calculation: one for debt service coverage, one for reserve requirements, and one for the loss-mitigation scenarios embedded in the lender's stress tests. The result is a structural divergence that borrowers discover only when market conditions shift or a lease renewal fails.

This divergence is not a modeling error. It is path-dependent. It emerges because underwriting, loan administration, and loss-mitigation teams operate under different constraints, different time horizons, and different regulatory mandates. The debt-yield team optimizes for closing speed and competitive positioning. The reserve-sizing team optimizes for capital adequacy and regulatory compliance. The loss-mitigation team optimizes for downside protection and worst-case cash flow recovery. Each team is rational within its own frame. None of them is wrong. But they cannot simultaneously be right about the same rent roll.

Portfolio One: Debt-Yield Underwriting and the Assumption of Continuity

The debt-yield team models the rent roll as a portfolio of cash flows that continue to maturity. This is the most aggressive of the three frameworks. In-place tenants are weighted by their contractual rent, and the team applies a blended credit adjustment based on aggregate tenant quality. A high-credit tenant (investment-grade or strong balance sheet) might receive a 0 percent haircut to in-place rent. A mid-market tenant might receive 5-10 percent. A weaker credit or single-concept retail operator might receive 15-25 percent. But the critical assumption is that each lease renews at or near its in-place rate, or that the spread between in-place and replacement rent is modest enough to be absorbed by the debt service calculation.

This model produces the loan amount, the debt service coverage ratio, and the loan-to-value that appears in the term sheet. It is the most visible calculation. It drives pricing, commitment, and borrower expectations. And it is structurally optimistic because it assumes renewal without pricing power loss or vacancy cost. The debt-yield team knows this. But their constraint is competitive: if they model aggressive renewal haircuts, they price the loan higher or reduce the loan amount, and the borrower shops elsewhere. The market discipline that should constrain this optimism is weak, because every lender faces the same competitive pressure. So the debt-yield model becomes the market norm, and the norm becomes embedded in the loan agreement as the baseline assumption.

Portfolio Two: Renewal Probability Cohorts and the Assumption of Segmentation

The reserve-sizing team models the rent roll not as a single cash flow stream, but as a series of renewal cohorts, each with its own probability of renewal and its own replacement-rent haircut. Leases expiring in years one through three are segregated from leases expiring in years four through eight. Within each cohort, tenants are grouped by credit quality, and renewal probabilities are assigned based on historical turnover, tenant financial performance, and market conditions. A high-credit tenant in year three might have a 90 percent renewal probability. A mid-market tenant in year six might have a 70 percent renewal probability. A lower-credit tenant in year four might have a 55 percent renewal probability.

For each lease that is modeled as non-renewing, the team calculates a replacement rent scenario. This scenario includes not just the new rent (which may be lower than in-place rent), but also the leasing cost (broker commission, tenant improvement), vacancy loss, and the time required to re-tenant. The replacement rent calculation is where the reserve-sizing model diverges most sharply from the debt-yield model. Whereas debt-yield assumes continuation or modest haircuts, reserve-sizing assumes material replacement-rent compression for leases that do not renew, and it applies a probability weighting to that compression.

This second portfolio model is more conservative than debt-yield, but it is also more complex and less visible to the borrower. It lives inside the reserve schedule, in footnotes, and in the loan administration manual. It is used to size the tenant-renewal reserve, the vacancy reserve, and sometimes the capital reserve. The borrower sees the reserve requirement (e.g., 7.5 percent of gross potential income), but rarely sees the underlying renewal cohort model or the replacement-rent assumptions that drive it. This opacity creates a structural problem: the borrower may believe that the in-place rent will support the debt service, because the debt-yield model says it will. But the reserve-sizing model is simultaneously assuming that some portion of that rent will not renew and will be replaced at a material discount. Both models are contractually binding. Both drive lender behavior. But they are not reconciled.

Portfolio Three: Loss-Mitigation Scenarios and the Assumption of Stress

The third model is embedded in the lender's stress-test framework and in the language of default remedies in the loan agreement. This model assumes that the rent roll can deteriorate faster and more severely than either the debt-yield or reserve-sizing models predict. It models scenarios in which multiple tenants fail to renew, renewal rents compress more than expected, and vacancy persists longer than modeled. It may assume that a key anchor tenant (if the asset is a shopping center or multi-tenant office) experiences financial distress and either vacates or demands a material rent reduction as a condition of renewal.

This third model is used to calculate the lender's loss exposure under various stress scenarios: 10 percent occupancy decline, 15 percent replacement-rent compression, six-month vacancy periods on major tenants. It is also used to determine trigger events for lender intervention, such as occupancy thresholds that activate the lender's right to take control of leasing, to approve tenant mix changes, or to approve capital expenditure budgets. The loss-mitigation model is the most conservative of the three. It assumes that the borrower will not proactively manage the rent roll; instead, it assumes that the lender must be prepared to manage it under stress.

This model is not typically disclosed in full to the borrower. But it is contractually embedded in covenants, default definitions, and lender remedies. When market conditions deteriorate, the lender's loss-mitigation model becomes operative, and the borrower discovers that the lender's assumptions about rent-roll performance were significantly more conservative than the assumptions in the loan term sheet. This creates a second structural problem: the borrower agreed to a loan based on debt-yield assumptions, but the lender has reserved the right to manage the asset based on loss-mitigation assumptions. The gap between these two frameworks is the lender's embedded optionality.

Where the Three Models Collide: The Renewal Year and the Reserve Release

The structural tension between these three models becomes visible in the years immediately before and after a major lease renewal. Suppose a 15 percent weighted-average rent roll is expiring in year four, and the in-place rent is $20 per square foot. The debt-yield model assumes renewal at $20 or at a modest discount to $20. The reserve-sizing model assumes a 60 percent renewal probability and a replacement rent of $17 per square foot for non-renewing leases. The loss-mitigation model assumes a 40 percent renewal probability and a replacement rent of $15 per square foot.

As the renewal year approaches, the borrower and the lender must decide whether to release tenant-renewal reserves that were held at closing. The reserve-sizing model says: "We assumed a 60 percent renewal rate. If we achieve that, reserves can be released." But the lender's loss-mitigation team is now running real-time occupancy and rent-roll data. If market conditions have softened, the loss-mitigation model may say: "Renewal probability is now 45 percent. Replacement rents are now $16. Reserves must remain in place or increase." The result is a reserve-release negotiation that re-surfaces the divergence between the debt-yield and loss-mitigation models. The borrower, who signed the loan based on debt-yield assumptions, believes that reserves should be released if in-place rents hold. The lender, armed with loss-mitigation data, believes that reserves should be held if market conditions have deteriorated relative to stress-test assumptions.

This is not a dispute over facts. Both parties are operating with accurate current data. The dispute is over which model should govern the reserve release. The debt-yield model is backward-looking (it compares current outcomes to closing assumptions). The loss-mitigation model is forward-looking (it compares current market conditions to stress-test scenarios). These two time horizons are structurally incompatible. The loan agreement typically provides language that appears to resolve this tension (e.g., "reserves may be released upon achievement of DSCR of 1.25x"), but that language is itself ambiguous. Does the DSCR calculation use the debt-yield assumptions (continuation rent) or the loss-mitigation assumptions (replacement rent for non-renewing leases)? The loan agreement rarely specifies. The result is that reserve-release disputes are almost always settled through negotiation, not contract language, because the contract language itself is built on three incompatible models.

The Regulatory and Institutional Origins of the Divergence

Why do these three models coexist without reconciliation? The answer lies in the regulatory and institutional structure of commercial real estate lending. The debt-yield model is driven by competitive market positioning and investor appetite for yield. Lenders must offer competitive loan amounts and rates, which requires optimistic rent-roll assumptions. The reserve-sizing model is driven by loan administration and capital adequacy requirements. Lenders must hold adequate reserves to cover losses under stress, which requires conservative assumptions. The loss-mitigation model is driven by the lender's own risk management and by secondary market investors' loss-mitigation concerns. Secondary market buyers demand detailed stress-testing and loss-scenario analysis, which requires extremely conservative assumptions.

These three regulatory and market pressures cannot be reconciled within a single model. So lenders maintain three models. The first model is visible to the borrower and drives the loan term sheet. The second and third models are internal to the lender and drive underwriting approvals, reserve sizing, and risk management. All three are contractually binding, but they operate on different time horizons and with different assumptions. The result is a structural feature of commercial real estate lending: every loan contains embedded optionality for the lender to reframe the rent roll as conditions change.

What I'm watching next

• Tenant-renewal covenant language that explicitly defines which model (debt-yield, reserve-sizing, or loss-mitigation) governs reserve releases and lease-renewal approval rights.
• Loan administration procedures that require documented reconciliation of the three rent-roll models at annual reviews and at major lease-renewal milestones.
• Secondary market investor demands for transparency around replacement-rent assumptions and their gap from debt-yield assumptions, which may force lenders to narrow the divergence or disclose it explicitly to borrowers.