The Underwriting Question Every Credit Officer Must Answer First
A sponsor sends you a CIM. The business — a $10M EBITDA specialty manufacturer — looks clean on the surface. But buried in the footnotes, the management team is proposing $2.8M in add-backs: a one-time consulting project, two years of “above-market” owner compensation, a plant relocation charge, and projected cost savings from a software migration that hasn’t yet been implemented. The headline Debt/EBITDA looks like 4.2x. Your lender-adjusted number? Try 5.7x.
This gap — between what a borrower presents and what a disciplined lender underwrites — is where credit losses are born. EBITDA normalization is not a compliance exercise. It is the single most consequential analytical judgment in leveraged credit underwriting. The number you accept as the earnings base will determine leverage, covenant headroom, debt service capacity, and ultimately whether your loan gets repaid.
This article teaches Bond Capital’s analytical framework for building a lender-adjusted EBITDA — step by step, with worked calculations, red-flag identification, and grounding in credit performance data that shows exactly what happens when underwriters get this wrong.
The Framework: From Reported EBITDA to Lender-Adjusted EBITDA
What Normalized EBITDA Actually Means
The terms are often used interchangeably, but the distinction matters in underwriting. Reported EBITDA is the raw accounting figure. Management-adjusted EBITDA is the seller’s proposed bridge — built to maximize enterprise value. Normalized EBITDA is the more defensible estimate of sustainable earnings after rigorous lender scrutiny. As one practitioner framework puts it, normalized EBITDA represents the earnings a buyer can sustainably underwrite after considering the full operating reality, including steady-state staffing, market compensation, and necessary reinvestment.
The formula is deceptively simple:
Lender-Adjusted EBITDA = Reported EBITDA + Approved Add-Backs − Required Deductions
Every item in that middle column must survive a structured test before it belongs in your underwriting model.
The CARVE Test: A Repeatable Add-Back Decision Framework
Bond Capital applies the CARVE test — a five-factor screen — to every proposed add-back before it is accepted, partially credited, or rejected in the credit file:
- C — Cash: Did the item consume cash, and will similar cash outflows continue after closing? A non-cash accounting charge may pass. A recurring cash consulting fee likely will not.
- A — Auditable: Is there objective evidence — invoices, board minutes, contracts, payroll records — or only a borrower narrative in a broker memo?
- R — Recurring: Review at least 24 to 36 months of financials. An item that appears in multiple periods is by definition not non-recurring.
- V — Verifiable: Can the adjustment be independently confirmed against third-party data — BLS compensation benchmarks, market lease comparables, supplier invoices?
- E — Excluded from operations: Is the item genuinely outside the normal course of operating the business going forward?
The cleanest credit memo shows both the borrower-adjusted EBITDA and the lender-adjusted EBITDA, with a documented bridge between them — noting which items were approved, haircutted, or rejected and why.
Core Add-Back Categories Lenders Commonly Accept
The following categories represent the most defensible add-backs in credit underwriting, subject to documentation:
- Excess owner or executive compensation: The delta between actual founder pay and a benchmarked market replacement salary. Requires third-party compensation data; the add-back is limited strictly to the excess.
- Documented one-time expenses: Legal settlements, M&A transaction costs, site relocation charges — provided they are non-recurring and evidenced by closing statements or invoices.
- Non-cash charges: Depreciation and amortization (already in the base EBITDA calculation), non-cash stock-based compensation, and unrealized loss provisions — subject to disclosure and materiality assessment.
- Related-party normalization: Adjustments to bring above- or below-market related-party costs (rent, services, supply agreements) to arm’s-length market rates. Requires contracts and market benchmarks to support both directions of adjustment.
The Analysis: Walking Through a $10M EBITDA Manufacturing Company
Step 1: Spread the Financials — Three Full Years Plus LTM
Start with three fiscal years of audited financials plus trailing-twelve-month (LTM) management accounts. Build a side-by-side spread with consistent line items. Look for revenue and margin trends — acceleration in the LTM period relative to the three-year average deserves particular scrutiny.
Step 2: Reconstruct the Add-Back Bridge
For our $10M EBITDA manufacturer, the sponsor’s proposed adjustments look like this:
| Add-Back Item | Sponsor Amount | CARVE Verdict | Lender Credit | Rationale |
|---|---|---|---|---|
| Owner compensation excess | $800K | ✓ Approved | $650K | Haircutted to BLS-benchmarked replacement cost; $150K benefits excluded |
| One-time consulting project | $600K | ✓ Approved | $600K | Contract evidenced; vendor closed; no recurrence in 36-month review |
| Plant relocation charge | $900K | ½ Partial | $400K | Moving costs approved; ongoing facility overlap costs rejected (recurring for 18 months) |
| Projected software savings (not yet realized) | $500K | ✗ Rejected | $0 | Implementation incomplete; synergies belong to acquirer, not current earnings base |
| Total Add-Backs | $2,800K | — | $1,650K | $1,150K rejected or haircutted |
Step 3: Calculate the Credit Impact
The delta between sponsor-adjusted and lender-adjusted EBITDA has a direct, material effect on every key credit metric:
| Metric | Sponsor View (EBITDA = $12.8M) | Lender View (EBITDA = $11.65M) | Difference |
|---|---|---|---|
| Senior Debt ($48M) | 3.75x Debt/EBITDA | 4.12x Debt/EBITDA | +0.37x leverage |
| Total Debt ($54M, incl. mezz) | 4.22x Total Leverage | 4.63x Total Leverage | +0.41x leverage |
| Interest Coverage (at 9% all-in) | 2.6x | 2.4x | Marginal headroom erosion |
| DSCR (incl. amortization) | 1.35x | 1.22x | Approaching 1.10x stress threshold |
A 0.4-turn difference in leverage may look modest in isolation. Under a modest stress scenario — a 10% revenue decline, a 150bps rate increase on floating-rate debt — that same borrower at 4.63x can breach a 5.0x maintenance covenant with very limited cushion. The scenario that looked “investment-grade private credit” on the sponsor’s EBITDA is a watch-list credit on the lender’s.
Industry Standards: What the Data Says About EBITDA Discipline and Credit Outcomes
The S&P Warning on Expanding Definitions
S&P Global has been direct on this risk. The agency views the “ever-expanding” definition of management-adjusted EBITDA as an artificial deflation of leverage and an inflation of profitability that contributes to understated valuation multiples. The practical consequence: deals underwritten to aggressive add-backs enter the credit cycle with hidden leverage that surfaces at the worst possible time.
The Covenant-Recovery Connection
The structural consequences of weak EBITDA definitions extend directly to loss severity in default. S&P data from companies emerging from bankruptcy found that lenders with maintenance covenants in loan contracts recovered an average of 82.2%, well above the 71.6% recovery rate for lenders without such contractual protections. The divergence was even wider at the median: 84.1% for loans with maintenance covenants versus 63.5% for cov-lite structures. When EBITDA definitions are loose, covenant headroom is artificially inflated — and lenders lose the early-warning mechanism that drives higher recoveries.
Credit Performance Data Box
📊 Private Credit & Leveraged Loan Benchmarks (2025–2026)
- Private credit default rate (2025, excl. distressed exchanges): approximately 1.6% — but rises to 4.7% when distressed restructurings are included, per Moody’s. The gap reflects how EBITDA-inflated structures avoid hard default through maturity extensions and amendments rather than outright payment failure.
- U.S. private credit default rate (April 2026): Fitch Ratings reported a 6.0% rate, reflecting continued deterioration in heavily leveraged, loosely documented structures.
- Distressed restructuring share: Moody’s estimates distressed exchanges accounted for roughly 65% of all 2025 private credit defaults — the dominant credit event type, not hard payment failure.
- First-lien senior secured recovery (historical): 70–80% ultimate recovery on average, per long-term Moody’s data. Recent expected recoveries have drifted toward 60–70% as capital structures carry more first-lien debt and fewer junior cushions, and as distressed exchanges deliver weaker outcomes than traditional bankruptcy proceedings.
- Cov-lite recovery penalty: Moody’s models a mean recovery of approximately 50% for first-lien-only cov-lite structures versus 65% when maintenance covenants are present — a 15-percentage-point recovery premium for covenant discipline.
- High-yield bond default rate (S&P, trailing 12 months through Sept. 2025): above 4%, near long-term averages of approximately 4.5%, with a forecast of 4.25% by June 2026.
The Most Dangerous Red Flags in Borrower EBITDA
- Recurring “one-time” restructuring charges: A company with non-recurring charges in five consecutive years does not have non-recurring costs — it has a permanently impaired cost structure. Challenge any add-back that appears in multiple periods.
- Revenue-side add-backs: Adjustments that increase the revenue line — projected price increases, expected new contracts, anticipated customer wins — belong in the investment thesis, not the underwriting EBITDA base.
- Unrealized synergies: Synergies belong to the acquirer. Including pre-close synergy savings in standalone normalized EBITDA is a structurally flawed adjustment that inflates the earnings base before value has been created.
- Stock-based compensation exclusion without disclosure: For technology-oriented businesses with high SBC, excluding it without flagging its magnitude can inflate normalized EBITDA by 10–20%. Always assess SBC relative to total EBITDA and model it as a real economic cost.
- Large add-backs without source documents: If an adjustment cannot be tied to an invoice, a board resolution, a contract amendment, or a payroll record, it has no place in the credit memo.
Underwriting Checklist: EBITDA Normalization Due Diligence
Documentation Requirements
- ☑ Three full fiscal years of audited or reviewed financial statements
- ☑ LTM management accounts with a clear revenue and margin bridge to the most recent audit
- ☑ General ledger detail for any proposed add-back category exceeding $100K
- ☑ Payroll records and third-party compensation benchmarks for owner/executive adjustments
- ☑ Invoices, contracts, or closing statements for all one-time expense add-backs
- ☑ Related-party lease or service agreements with market-rate comparables
- ☑ Quality of Earnings (QoE) report from an independent accounting firm for transactions above $5M in senior debt
Analytical Steps
- Build the EBITDA bridge: Start from reported GAAP net income; reconstruct EBITDA line by line with documented adjustments.
- Apply the CARVE test to each proposed add-back; document the decision (approve / haircut / reject) and the rationale in the credit file.
- Construct the lender-adjusted EBITDA schedule as a formal reconciliation separate from the borrower’s management adjustments.
- Recalculate all key credit metrics — Debt/EBITDA, Interest Coverage Ratio, DSCR, and Fixed Charge Coverage Ratio — using only lender-adjusted EBITDA.
- Stress-test the EBITDA base under a −10% and −20% revenue scenario; model the covenant headroom at each stress level.
- Confirm covenant definitions in the draft credit agreement use lender-adjusted EBITDA; flag any sponsor-favorable carve-outs that would allow further add-backs post-close.
Key Questions to Ask in Diligence
- Has any “one-time” add-back category appeared in more than one fiscal year?
- What is the company’s cash conversion ratio (unlevered free cash flow / EBITDA)? A low ratio signals working capital intensity or capex requirements that EBITDA obscures.
- How does the LTM EBITDA margin compare to the 3-year average? Accelerating margins near a transaction close warrant deep scrutiny.
- Are any revenue-side adjustments included in the sponsor’s add-back schedule?
- What portion of EBITDA is attributable to the top three customers? Concentration risk can invalidate normalized earnings assumptions quickly.
Key Takeaways for Credit Practitioners
- 🔴 The earnings base is the leverage base. Every dollar of accepted add-back reduces apparent leverage and expands covenant headroom. Lenders who accept the borrower’s number without independent verification are underwriting to a fiction.
- 🔴 CARVE before you commit. Apply Cash, Auditable, Recurring, Verifiable, and Excluded-from-operations screens to every proposed adjustment. Document the decision. The credit file discipline you build at underwriting becomes your legal and analytical record if the loan moves to workout.
- 🔴 Cov-lite structures amplify EBITDA risk. When covenant definitions are loose and maintenance tests are absent, there is no mechanism to surface credit deterioration early. The data is unambiguous: loans with maintenance covenants have produced materially higher recovery rates than cov-lite structures in default.
- 🔴 Distressed exchanges are the dominant default form in private credit. With roughly 65% of 2025 private credit defaults taking the form of debt exchanges and maturity extensions rather than hard payment failures, the risk of EBITDA inflation is not just one of upfront mispricing — it is one of delayed recognition and invisible deterioration across the portfolio.
- 🔴 Build two EBITDA lines — always. The cleanest credit memo presents both borrower-adjusted EBITDA and lender-adjusted EBITDA in a formal bridge table. The gap between the two numbers is itself a credit signal: the wider it is, the more aggressive the sponsor’s assumptions and the more disciplined the underwriter must be in response.
