Bank Debt: Recourse vs Nonrecourse, Bad-Boy Carveouts, and Loan Covenants

Key Takeaways

The Recourse Question the Term Sheet Doesn't Answer

Bank debt term sheets typically describe themselves as "nonrecourse" or "limited recourse." The categories are real but incomplete. Every nonrecourse bank loan includes bad-boy carveouts that convert nonrecourse exposure to full recourse under specified conditions. The question that matters in underwriting is not "is the loan recourse" but "what specifically triggers recourse, what is the practitioner-realistic probability of those triggers firing, and what does it cost to negotiate them out."

This article works through the recourse taxonomy with illustrative pricing ranges: the bps rate difference between nonrecourse, partial recourse, and full recourse, the carveout categories (loss-only vs springing), the burn-off thresholds that release recourse over time, and the covenant breach question that is genuinely separate from the bad-boy framework. The audience is institutional borrowers negotiating bank debt term sheets, capital markets analysts pricing the recourse-equivalent cost difference between bank and agency, and sponsors facing the tighter recourse terms that regional banks have often required on small-balance bank loans in 2025–2026.

Recourse, Nonrecourse, and Bad-Boy Carveouts

Three structures dominate the institutional CRE bank debt market:

• Full recourse. The borrower (and typically a personal guarantor on smaller deals) is personally liable for the full loan balance plus accrued interest plus enforcement costs. Default puts the entire borrower balance sheet at risk, not just the property collateral. Common on bank construction debt for less-capitalized sponsors and on smaller balance loans (<$10M).

• Partial / burn-off recourse. The borrower is recourse during the construction or value-add period and burns off to nonrecourse at stabilization, typically conditioned on hitting DSCR, occupancy, and debt yield thresholds. The construction component of many bank loans uses this structure; the permanent component is fully nonrecourse with carveouts.

• Nonrecourse with carveouts. The borrower is not personally liable for the loan balance under normal circumstances; the lender's remedy is the collateral property. But specified events ("bad boy" acts) convert nonrecourse exposure to either loss-only recourse (the lender recovers actual losses from the bad act) or full springing recourse (the entire loan balance becomes personally recourse). Life co loans and CMBS typically use this structure as the standard; bank loans vary and are frequently partly or fully recourse.

The carveout structure has been the subject of significant legal commentary because the boundary between loss-only and springing recourse varies by lender and is routinely negotiated. Practitioner guidance fromLP Legaldescribes two categories, "loss" events ("above the line") and "full recourse" events ("below the line"), and notes that lender-drafted provisions often expand those categories in ways that materially increase guarantor exposure.

The Recourse Pricing Discount

Recourse is a priced trade-off. A borrower who accepts partial or full recourse typically receives a lower rate than on comparable nonrecourse terms, because the lender gains a claim beyond the collateral. The ranges below are illustrative assumptions, not a published benchmark; actual quotes vary by lender, sponsor, and asset:

For a $40M, 5-year, interest-only bank loan, a 100-bps rate difference between nonrecourse and full recourse pricing is $400K per year in interest cost, $2M over the 5-year term (undiscounted; $1.5M to $3.0M across the 75–150 bps range). The discount is real money; sponsors who treat recourse as an abstract risk rather than a priced trade-off systematically accept too much liability for too little pricing benefit.

The institutional context: per Federal Reserve H.8 weekly consolidated banking data (Table 1, commercial real estate loans, percent change at a seasonally adjusted annual rate, break adjusted), CRE loan growth at U.S. commercial banks slowed to 1.8% in 2024 and 1.6% in 2025, from 4.3%, 11.3%, and 4.8% in 2021, 2022, and 2023, and was 3.0% in the first quarter and 3.2% in the second quarter of 2026. Bank construction debt often comes with full or partial recourse.

The Bad-Boy Carveout Taxonomy

The bad-boy carveout list in a typical institutional nonrecourse loan often runs to roughly 8–20 items (illustrative; it varies by lender) in two categories: loss-only (the lender recovers actual losses from the act) and full springing (the entire loan becomes personally recourse).

Standard loss-only carveouts:

• Fraud, intentional misrepresentation, or material omission in loan documents or financial statements
• Willful misconduct in property management or operations
• Misappropriation of rents, security deposits, or insurance proceeds
• Failure to maintain insurance coverage required by the loan documents
• Failure to pay property taxes or other obligations that create senior liens
• Environmental violations or failure to remediate known environmental conditions
• Waste of the property (deferred maintenance, intentional damage)
• Removal of property collateral without lender consent
• Damages from unpermitted alterations
• Failure to comply with anti-money-laundering or sanctions regulations

Full springing recourse triggers:

• Voluntary bankruptcy filing by the borrower
• Collusion with creditors to file involuntary bankruptcy against the borrower
• Transfer of the property without lender consent ("due-on-sale" violations)
• Transfer of equity interests in the borrower beyond permitted thresholds
• Unpermitted additional debt or liens on the property or the borrower
• Certain SPE (single-purpose entity) covenant breaches, often limited to those that cause substantive consolidation
• Certain covenant breaches, in some lender forms (negotiable; covered below)

The loss-only category compensates the lender for the specific damages from the bad act; the springing recourse category puts the entire loan balance on the borrower's personal balance sheet. The distinction is enormous: a $40M loan with a $500K loss-only event costs the borrower $500K; the same loan with a full-springing-recourse event puts $40M of personal liability on the borrower (or guarantor) overnight.

Springing Recourse Triggers

Four of the most commonly negotiated springing recourse triggers in institutional CRE bank debt:

• 1. Voluntary bankruptcy filing. The borrower (or any constituent member exceeding threshold equity) filing voluntary bankruptcy triggers full springing recourse to the guarantor. This is the carveout that prevents sponsors from using bankruptcy as a negotiation tool with the lender. Some SPE structures bifurcate the borrower's bankruptcy from the constituent members' bankruptcy to limit the trigger's scope.

• 2. Unpermitted transfer. Selling or transferring beneficial interest in the property (or in the borrower entity itself) without lender consent. The threshold for "transfer" varies; some loans have de minimis exceptions (transfers below an agreed equity percentage, intra-family transfers, partner-to-partner within an existing structure). Sponsors who add or remove LPs mid-hold need to verify the permitted- transfer provisions before doing so.

• 3. Subordinate financing. Layering on mezzanine, preferred equity, or other subordinate debt that creates a lien or claim against the property or borrower entity, without lender consent. Permitted subordinate debt structures are negotiated in loan documents; unpermitted layering triggers full springing recourse.

• 4. SPE covenant breach. The borrower entity must remain a single-purpose entity, with no other business activities, no commingled assets, and limited indebtedness beyond the loan. Lenders commonly make SPE breaches full-recourse events, but borrowers often negotiate to limit full recourse to breaches that cause substantive consolidation, with non-material SPE lapses treated as loss-only. Practitioner guidance from LP Legal (linked above) notes that routine actions and administrative errors relating to SPE separateness requirements can inadvertently breach these covenants. The SPE covenant exists to prevent substantive consolidation in bankruptcy: the lender wants confidence that the borrower's bankruptcy estate is limited to the property.

Lender forms commonly include other full-recourse events as well, such as receiver appointments, admissions of insolvency, and interference with the lender's remedies; borrowers typically negotiate these down to voluntary or bad-faith conduct.

Burn-Off Provisions

Bank construction debt and value-add bridge loans frequently include burn-off provisions that release partial recourse upon achievement of specified performance milestones. Illustrative release tests (assumed; actual thresholds vary by lender and property type, and are distinct from the ongoing maintenance covenants discussed below):

• DSCR threshold: Trailing 3-month or 6-month DSCR ≥ 1.20x (sometimes 1.15x or 1.25x depending on lender / property type)
• Occupancy threshold: Physical occupancy ≥ 90% (a multifamily example) or trailing 6-month effective gross income ≥ 95% of underwritten EGI
• Debt yield threshold: Stabilized NOI ÷ loan balance ≥ 8.0% (sometimes 8.5% for higher-risk property types)
• No event of default: No EOD has occurred or is continuing

The burn-off cleans up the construction or value-add period without permanent recourse. Sponsors who are capable of hitting the stabilization thresholds end up with a nonrecourse permanent loan; sponsors who miss the thresholds continue with partial recourse until they do hit them or the loan refinances. The math: a 50-bps rate discount for accepting recourse on a $40M loan = $200K per year of lower interest cost, vs the cost of permanent recourse exposure. Many borrowers would accept 12–24 months of stabilization- period recourse for that discount (illustrative); few would accept persistent recourse without a burn-off.

Covenant Breach as Separate EOD

Beyond bad-boy carveouts, bank loans typically include affirmative and negative covenants whose breach constitutes a separate event of default. Common covenants:

• Financial covenants: minimum DSCR, debt yield, LTV, sponsor net worth, liquidity
• Reporting covenants: quarterly financial reporting, annual audited financials, occupancy reports, rent rolls
• Operational covenants: property management standards, capital reserve maintenance, insurance compliance
• Information covenants: notification of material adverse changes, tenant defaults, casualty events

Covenant breaches don't automatically trigger springing recourse, but they constitute an EOD that gives the lender remedies including: acceleration of the loan, additional capital contributions, lender appointment of a property manager, transfer-of-collateral, or (in extreme cases) foreclosure. The lender's discretion in exercising remedies post-EOD is what makes covenant compliance critical to maintain even on a nominally nonrecourse loan.

In a stressed credit environment, lenders often tighten covenant thresholds rather than relying only on higher spreads, and covenant-breach negotiations tend to increase as borrowers approach maturity under stress. This is a general market observation, not a quantified benchmark; maintenance covenant levels are set loan by loan.

The 2025-2026 Regional Bank Pullback

Regional banks are a common source of small- to mid-balance ($5M–$50M) construction and value-add CRE debt. Following the 2023 regional bank stress (Silicon Valley Bank, First Republic, Signature), bank CRE lending slowed. Per Federal Reserve H.8 data (Table 1, commercial real estate loans), CRE loans at U.S. commercial banks grew 1.8% in 2024 and 1.6% in 2025, against 4.3% to 11.3% a year in 2021–2023, and growth was back to about 3% in the first half of 2026.

As a general market observation, the slowdown has shown up in recourse structures. Bank loans that closed in 2018–2021 often featured 24-month burn-off provisions with relatively achievable thresholds. Loans closing in 2024–2026 with the remaining regional bank balance-sheet capacity often feature (illustrative):

• Higher burn-off thresholds (for example, DSCR 1.25x rather than 1.15x; debt yield 8.5% rather than 7.5%)
• Longer burn-off periods or no burn-off at all on construction debt
• Full recourse on bridge/construction debt for less-tenured sponsors
• Tighter covenant levels and more frequent reporting
• Personal guarantees from sponsor principals on small-balance loans

The institutional implication: agency and life co debt (typically nonrecourse) is the usual alternative. Sponsors who can clear agency or life co terms often prefer them even when bank pricing is nominally lower, because the recourse difference is material and the near-term refinance optionality (rate certainty for 7–10 years vs 3–5 years on bank perm) is valuable.

Worked Example: $40M Bank Loan

A $40M institutional multifamily acquisition financed with a bank acquisition / value-add loan. Three structures on the term sheet (rates are assumed for illustration; interest-only on a fully funded $40M, undiscounted):

The pricing differential between Option A and Option C is $2.0M over 5 years, 5% of the loan balance. Sponsors who treat recourse as a discrete yes/no choice without pricing the spread either accept too much liability (taking Option C without compensation) or pay too much for nonrecourse (rejecting Option B, whose 40-bps lower rate saves $800K over 5 years and which burns off to nonrecourse after 24 months).

The institutional framework: price the recourse exposure as if it were an option. Full recourse is the sale of an option to the lender in exchange for a 100-bps lower rate. Partial recourse with burn-off after 24 months is the sale of a 24-month option in exchange for a 40-bps lower rate (assumed in this illustration). Compare the rate savings to the borrower's estimated probability, over the 5-year hold, of triggering recourse. As an illustrative assumption, a 1–5% probability on a $40M exposure implies an expected cost of $0.4M to $2.0M against the $2.0M of interest saved under Option C. Because a recourse claim is a deficiency claim (the shortfall after the collateral is sold), actual exposure is typically below the full balance, so these figures are an upper bound; trigger probability is higher for construction and value-add. Sponsors with low trigger probability prefer the option-sale (Option C savings); sponsors with high trigger probability prefer the nonrecourse premium (Option A).

Five Mistakes Borrowers Make

• Treating "nonrecourse" as the term sheet says it. Every nonrecourse loan has bad-boy carveouts. The carveout list often runs to roughly 8–20 items (illustrative); reading them in detail is the actual underwriting of the recourse exposure. The most common sponsor error is signing nonrecourse documents without diligence on the springing-recourse triggers.

• Underestimating SPE covenant violations. SPE covenants restrict the borrower entity to owning the property only, with no other business, no commingled assets, and limited indebtedness. Sponsors who fund acquisitions out of operating accounts of the borrower entity, who lend money to or from the borrower entity, or who collapse multiple property holdings into a single entity routinely breach SPE covenants, producing springing recourse exposure they didn't intend.

• Missing the burn-off thresholds. Burn-off provisions only release recourse when the thresholds are met. Sponsors who don't hit DSCR or debt yield thresholds at stabilization continue with partial recourse indefinitely. Underwrite to a healthy buffer above the thresholds; don't underwrite to the threshold itself.

• Confusing covenant breach with bad-boy event. Most covenant breaches are EODs that give the lender remedies including acceleration and foreclosure; they don't automatically trigger full springing recourse. But "material" covenant breaches (which the lender's discretion to call) can be tied to springing recourse in some loan documents. Read the specific cross-reference between covenant defaults and bad-boy springing recourse in the loan agreement.

• Underpricing the recourse trade-off. Sponsors who agree to full recourse "to get the lowest rate" without pricing the trade-off are systematically accepting too much exposure for too little savings. The 75–150 bps illustrative discount for full recourse is real money; the trigger probability is real risk. Run the math both ways.

Do It in Apers

Related Articles

• Bridge Loans: Floating-Rate Risk and Exit Assumptions: bridge debt typically has partial recourse during the value-add period.
• Construction Loans: Draws and Interest Reserves: construction debt often carries partial or full recourse in the current bank market.
• CMBS Conduit vs SASB: CMBS is nonrecourse standard with similar carveout taxonomy.
• IRR Calculator and Formula for Real Estate: the return metric that the recourse discount feeds into.

FAQ