The Underwriting Question That Determines Everything
A sponsor submits a Confidential Information Memorandum. The headline reads: $12.4M Adjusted EBITDA. 4.8x leverage. Strong free cash flow conversion. Your credit committee wants an answer by Thursday.
Here is the problem: that $12.4M figure is management’s number. It starts with $8.1M in reported EBITDA and adds back $4.3M across eleven line items — restructuring charges, a one-time legal settlement, above-market owner compensation, COVID-era revenue disruption, and a catch-all labeled “other non-recurring.” The question is not whether adjustments are appropriate. Some clearly are. The question is: which ones survive rigorous lender scrutiny, and what does the business actually earn on a sustainable, arm’s-length basis?
That answer — the normalized EBITDA a lender can defensibly underwrite — is the single most consequential number in your credit analysis. It drives your leverage multiple, your debt sizing, your covenant thresholds, and ultimately your loss-given-default exposure. Get it wrong on the high side, and you have underwritten a loan that was never serviceable at inception.
This is not a theoretical concern. The Financial Stability Board’s May 2026 vulnerability report found that EBITDA add-back practices in private credit may mean that true leverage is materially understated — with actual debt-to-EBITDA ratios potentially closer to 7x once aggressive adjustments are stripped out. This article builds the framework to prevent that from happening in your portfolio.
The Normalization Framework: Four Adjustment Categories
Effective EBITDA normalization is not a single calculation — it is a structured audit of every line item between reported net income and the earnings figure your credit model will rely on. The methodology follows four primary adjustment categories, each with distinct evidence standards.
Category 1: Owner-Specific Compensation & Personal Expenses
In owner-operated middle-market businesses, reported compensation frequently diverges from true market-rate replacement cost. An owner drawing $950,000 annually where a qualified CEO commands $400,000 in the market creates a legitimate $550,000 add-back — but only if it is documented against comparable compensation surveys and validated in the Quality of Earnings review. The inverse is equally important: an owner paying themselves $150,000 when market replacement costs $380,000 requires a downward adjustment that many seller-prepared schedules quietly omit.
Category 2: Non-Recurring Revenue and Expense Items
One-time legal settlements, restructuring charges, M&A transaction costs, natural disaster impacts, and COVID-era disruption are the most commonly presented add-backs — and the most frequently abused. The critical test is not whether an item is labeled “non-recurring” by management, but whether a competent lender can independently verify it will not recur. A company that has “non-recurring” restructuring charges in three consecutive years has a recurring restructuring problem, not three separate non-recurring events.
Category 3: Pro Forma Run-Rate Adjustments
Acquisitions, new customer contracts, completed cost initiatives, and facility openings can justify run-rate adjustments that project a full-period benefit not yet reflected in trailing financials. These are the highest-risk adjustments in any lender’s analysis. Pro forma add-backs for projected synergies or forward cost savings require demonstrated management execution history — not projections alone. As institutional lenders now demand: show the evidence of similar completed initiatives before crediting forward assumptions.
Category 4: Non-Cash Items
Stock-based compensation, goodwill impairments, unrealized gains or losses, and certain reserve movements are structurally appropriate add-backs — but require precise scoping. Stock-based compensation, in particular, is an economic cost that ultimately dilutes equity holders and should be treated with care in cash-flow-based lending where debt service competes directly with operating cash generation.
The Lender’s Standard: Lenders use the most conservative EBITDA definition because it drives debt sizing, covenants, and repayment capacity. Unlike sponsor-prepared schedules, lender-underwritten EBITDA only credits adjustments that can be clearly justified and traced back to the company’s financial statements.
Walking the Bridge: A $10M EBITDA Manufacturing Company
Consider a concrete example. A Midwest industrial manufacturer is being acquired by a private equity sponsor at 5.5x EBITDA. The seller’s CIM presents the following bridge:
| Line Item | Seller Claim | Lender Assessment | Rationale |
|---|---|---|---|
| Reported EBITDA (LTM) | $8,100,000 | $8,100,000 | Starting point — agreed |
| Owner compensation normalization | +$550,000 | +$420,000 | Market comp survey supports $480K replacement; $30K in personal perks excluded |
| One-time legal settlement | +$380,000 | +$380,000 | Fully documented; closed matter; accepted |
| Restructuring charges (Year 2 & 3) | +$620,000 | $0 | Two consecutive years = recurring operational pattern; rejected |
| New customer contract run-rate | +$900,000 | +$450,000 | Contract signed but 6-month revenue ramp not yet demonstrated; 50% credit applied |
| Non-cash stock compensation | +$250,000 | $0 | Real economic cost to equity; excluded from cash-flow underwriting base |
| “Other non-recurring” (undocumented) | +$600,000 | $0 | Cannot be traced to P&L line items; rejected per standard lender policy |
| Underwritten EBITDA | $11,400,000 | $9,350,000 | $2.05M gap — leverage moves from 4.8x to 5.9x |
This is not an edge case. The $2.05M normalization gap is entirely typical in sponsor-backed middle-market transactions. At a 5.5x purchase multiple and $55M enterprise value, that gap changes the lender’s leverage calculation from a manageable 4.8x to a stretched 5.9x — well above most senior lender comfort thresholds of 5.0x–5.5x for manufacturing credits.
Stress Testing the Base Case
A disciplined lender runs at minimum three scenarios against the underwritten EBITDA base:
- Base Case: Underwritten EBITDA of $9.35M, 5.9x leverage. DSCR at current debt service? Calculate as: ($9.35M × [1 minus assumed cash tax and capex rate]) ÷ annual debt service. For a $45M senior term loan at SOFR+500 (~10.5%), annual interest is approximately $4.7M. DSCR ≈ 1.5x — thin but serviceable.
- Downside Case (–20% EBITDA): $7.48M EBITDA. Leverage rises to 6.0x on a stressed basis. DSCR falls below 1.2x — triggering most maintenance covenant thresholds. This is the scenario where covenant design becomes your primary defense.
- Severe Stress (–35% EBITDA): $6.08M EBITDA. DSCR approaches 0.9x. Lender now relies entirely on collateral recovery. At current first-lien recovery rates, loss-given-default exposure becomes material.
What the Credit Data Tells Us in 2025–2026
The stakes of EBITDA normalization discipline have never been higher. The credit cycle has turned, and the costs of permissive add-back practices are now appearing in realized default and recovery statistics.
📊 Credit Performance Data Box — Mid-2026
- Private Credit Default Rate (2025): Moody’s estimates the private credit default rate in 2025 ranged between 1.6% and 4.7%, depending on whether distressed exchanges are included — a meaningfully wide band driven by definitional ambiguity in the asset class. (Source: Moody’s, April 2026)
- Fitch Record High: In May 2026, Fitch Ratings reported that the U.S. private credit default rate hit a record high of 6.0% in April 2026, with private-credit-backed corporate borrowers experiencing a 9.2% default rate across all of 2025. (Source: Forbes / Fitch, May 2026)
- Distressed Exchanges Dominating: Approximately 65% of all 2025 corporate defaults were distressed restructurings — debt exchanges and maturity extensions agreed under duress — not hard defaults. This masks realized loss severity in portfolio reporting. (Source: Moody’s, May 2026)
- First-Lien Recovery Compression: 2024 first-lien term loan recovery rates fell to 49.2%, well below the 1987–2024 long-run average of 76.4%. Second-lien recovery dropped to a historically low 14.5%, versus a long-run average of 46.5%. (Source: Moody’s Annual Default Study, February 2025, as cited by CTA Acquisitions Recovery Report)
- Proskauer Default Tracking: Proskauer’s Private Credit Default Index, tracking 697 loans totaling $189.2 billion, recorded a 2.73% default rate in Q1 2026 — up sharply from 1.84% just two quarters prior. (Source: Forbes, May 2026)
- Leveraged Loan Defaults Elevated: For leveraged loan issuers, the Moody’s-projected default rate reached 7.5% at end-2025 and was forecast to remain near 7.3%–7.9% into Q2 2026 — nearly double the 3.4% historical average. (Source: Moody’s Credit Strategy Outlook, July 2025)
The Structural Warning: True Leverage May Be Understated
The Financial Stability Board’s May 2026 private credit vulnerability report raised a systemic concern directly tied to EBITDA normalization practices: the widespread use of add-backs in private credit documentation may mean that reported leverage ratios significantly understate actual debt burdens. The FSB analysis suggested that after aggressive adjustments are stripped out, true debt-to-EBITDA in private credit portfolios could be materially higher than headline figures suggest.
This concern is echoed at the highest levels of institutional finance. JPMorgan CEO Jamie Dimon, in his 2026 shareholder letter, specifically identified the pattern of “more aggressive and positive assumptions about future performance (called add-backs), weaker covenants, more use of PIK” as the defining underwriting risk in leveraged lending — warning that losses will be “higher than expected.” These are not abstract risks. The late-2025 bankruptcy of First Brands Group surfaced allegations of borrowing against the same assets more than once — a direct consequence of inadequate collateral and EBITDA verification at origination.
The Bond Capital EBITDA Normalization Underwriting Checklist
Use this checklist for every credit submission where adjusted EBITDA deviates meaningfully from reported EBITDA. No adjustment should survive without clearing each applicable gate.
📋 Phase 1: Establish the Reported Baseline
- ☐ Obtain three years of audited (or CPA-reviewed) financial statements — do not accept compilations alone for loans above $5M
- ☐ Reconcile management-prepared EBITDA bridge back to GAAP net income line-by-line
- ☐ Confirm reported EBITDA is consistent across all submitted documents (CIM, tax returns, bank statements)
- ☐ Flag any discrepancy between EBITDA growth and operating cash flow growth — if EBITDA is rising but operating cash flow is flat, ask why
📋 Phase 2: Interrogate Every Add-Back
- ☐ Map every proposed adjustment to a specific P&L line item — reject any add-back that cannot be traced to the general ledger
- ☐ Request bank statements or third-party invoices for all add-backs above $100,000
- ☐ For owner compensation adjustments: obtain a compensation benchmarking survey from an independent source; apply market replacement cost net of any actual expected replacement hire
- ☐ For “non-recurring” items: require documentation that the event is closed, final, and not indicative of an ongoing operational pattern. A legal reserve that recurs in Year 2 and Year 3 is a recurring cost.
- ☐ For run-rate / pro forma items: require a signed contract or completed operational milestone. Apply a haircut (typically 25%–50%) for any run-rate benefit that has not yet been demonstrated in trailing cash flow
- ☐ For synergy add-backs: require evidence of comparable completed initiatives by this management team. Future synergies without execution history receive zero credit in the underwritten base case
- ☐ Apply a cap: consider limiting total add-backs to no more than 25%–30% of reported EBITDA absent extraordinary deal-specific circumstances
📋 Phase 3: Build the Lender’s Earnings Bridge
- ☐ Construct the independent lender EBITDA bridge separate from the seller’s schedule — do not simply edit the sponsor’s version
- ☐ Calculate EBITDA to free cash flow conversion: deduct normalized capex, cash taxes, and working capital requirements. A business with 90% EBITDA-to-FCF conversion is materially different from one at 55%
- ☐ Compute Debt Service Coverage Ratio (DSCR) on the lender-underwritten EBITDA, not the sponsor-presented figure
- ☐ Run base, downside (–20%), and severe stress (–35%) scenarios and confirm covenant headroom in each
📋 Phase 4: Set Covenants to the Underwritten Number
- ☐ Anchor maximum leverage covenant to the lender-underwritten EBITDA, not the sponsor’s adjusted figure
- ☐ Build in an EBITDA definition in the credit agreement that caps add-backs — particularly projection-based or synergy-based items — as a percentage of EBITDA (common market practice: 15%–25% cap on non-cash and projected items)
- ☐ Require quarterly financial reporting with EBITDA covenant certification tied to a defined, add-back-limited formula
- ☐ Set covenant step-downs in the first 12–18 months post-close to ensure the business is performing to its underwritten case — not just floating on post-close momentum
Key Takeaways for Credit Underwriters
- The gap between reported and adjusted EBITDA is a diligence starting point, not a conclusion. Every dollar of add-back represents a claim that must be independently verified against source documents before it earns a place in your underwritten base case.
- Non-recurring is a characterization, not a fact. Restructuring charges appearing in two or more consecutive periods are recurring by definition. Reject this add-back category without documented evidence that the underlying cause has been permanently eliminated.
- Run-rate and synergy adjustments require demonstrated execution history. Forward-looking pro forma EBITDA credits that are not yet reflected in trailing cash flow carry the highest uncertainty. Apply systematic haircuts and refuse to credit forward synergies without comparable completed initiatives.
- Covenant architecture must mirror underwritten EBITDA. If the credit agreement’s EBITDA definition is more permissive than your underwritten figure, you have created a structural gap where the borrower can remain technically compliant while the actual business deteriorates well below serviceable levels.
- Recovery rates are not your safety net in 2025–2026. With first-lien recoveries compressing toward 49% in recent default cycles — versus a historical average above 76% — the underwriting discipline applied at origination is the primary determinant of credit outcome. Collateral alone will not rescue a loan built on an unsupportable EBITDA base.
