The Underwriting Challenge

A sponsor presents a manufacturing business with $6.5 million of reported EBITDA. The confidential information memorandum, however, shows $10.2 million of adjusted EBITDA — a 57% premium over the audited number. The proposed senior debt is $45 million, which the sponsor frames as a comfortable 4.4x leverage on adjusted earnings. Run the same debt against reported EBITDA, however, and leverage is 6.9x. Which number do you underwrite?

That question sits at the core of every private credit transaction. EBITDA normalization is the process of converting a borrower's reported or management-prepared earnings into a figure that reflects the sustainable, recurring cash-generating capacity of the business under new or continuing ownership. Done rigorously, it is one of the most protective steps in the underwriting process. Done carelessly — or under competitive pressure to deploy capital — it is one of the most dangerous sources of credit loss.

The stakes are measurable. Research published by the Financial Stability Board in May 2026 found that, with EBITDA adjustments stripped out, true leverage across private credit borrowers could be approximately one full turn higher than reported — translating directly into unpriced default risk embedded in headline leverage multiples. Getting normalization right is not a detail. It is the foundation on which every other underwriting ratio rests.

The Framework: A Three-Layer Normalization Taxonomy

Experienced credit practitioners organize the EBITDA bridge into three structurally distinct layers. The distinction between layers determines the level of scrutiny, documentation standard, and haircut policy applied to each line item.

Layer 1 — Definitional Adjustments

These are the mechanical components of the EBITDA formula: adding back interest expense, income taxes, depreciation, and amortization to net income. No analytical judgment is required. Every credit memo begins here, and every leverage ratio is anchored to this floor.

Layer 2 — Quality-of-Earnings Adjustments (One-Time Items)

These adjustments cover items the borrower asserts are genuinely non-recurring: settled litigation costs, M&A advisory fees, restructuring charges from a facility closure, or severance from a completed workforce reduction. The single evaluative test applied to every item in this layer is whether the expense will genuinely not recur. Most diligence effort concentrates here, and most disputes between borrower-adjusted and lender-adjusted EBITDA originate in this layer.

Layer 3 — Normalization Adjustments (Structural Corrections)

These adjustments correct for recurring items that are mispriced on the borrower's books relative to arm's-length market rates. Classic examples include owner compensation set materially above the market rate for the role, related-party rent paid to an LLC owned by the selling founder, family members on payroll without substantive documented roles, or discretionary personal expenses run through the business. Unlike Layer 2 items, these are not asserted as non-recurring; they are asserted as owner-specific costs that a third-party-managed business would not bear.

The Lender's Evidentiary Standard

A defensible add-back in any layer requires four things: source documentation that ties to the number, a narrative explaining why the item is non-recurring or owner-specific, a benchmark supporting any market-rate claim (e.g., a compensation survey for owner salary normalization, a CoStar comparable for related-party rent), and categorization that is consistent with how comparable transactions have treated the same item type. Without all four, the adjustment should be rejected or haircut at the lender's discretion.

The cleanest credit analysis shows both borrower-adjusted EBITDA and lender-adjusted EBITDA, with a clear, line-by-line bridge between them. That bridge is not a formality — it is the central analytical deliverable of the underwriting process.

The Analysis: Walking a $10M EBITDA Manufacturing Company

The following example illustrates how borrower-presented adjustments translate into lender-accepted EBITDA after underwriting scrutiny. The company is a regional industrial manufacturer with $38 million of trailing twelve-month revenue.

Leverage Impact

On $45 million of proposed senior debt, the leverage outcome changes materially depending on which EBITDA is used:

• Reported EBITDA ($6.5M): 6.9x leverage
• Borrower-adjusted EBITDA ($8.2M): 5.5x leverage
• Lender-adjusted EBITDA ($7.2M): 6.3x leverage

The difference between the borrower's framing and the lender's underwriting EBITDA is 1.0x of leverage — a full turn of unpriced risk. At 6.3x, the transaction exceeds the 5.0x to 5.5x threshold that most mid-market direct lenders apply to manufacturing credits of this size and sector profile. The structuring question becomes whether covenant headroom, collateral coverage, and equity cushion adequately compensate for that incremental leverage — not whether the borrower's presentation was internally consistent.

Industry Standards and Credit Performance Context

Understanding why EBITDA normalization discipline matters requires grounding the analysis in live credit performance data. The market context in late 2026 reinforces the cost of underwriting to inflated earnings.

The Add-Back Cap Problem

Because expansive add-backs can materially affect leverage outcomes, institutional lenders often impose caps — including shared caps expressed as a percentage of EBITDA — on projection-based or non-recurring adjustments to better reflect actual operating cash flows. According to a Proskauer survey cited in market analysis, the most common maximum for aggregate add-backs in private credit agreements was 20% to 24.9% of EBITDA, and only 4% of private credit lenders permitted uncapped add-backs. Yet aggregate leverage has continued to rise: MSCI data reported that median net debt among private credit borrowers rose 9.7% year-over-year in Q1 2026, slightly faster than EBITDA growth, pushing median net leverage to 5.3x.

Sponsors increasingly negotiate add-backs for projected synergies, cost savings, and one-time charges — with some market commentary suggesting that such adjustments can inflate EBITDA by 20% to 30%, making leverage ratios appear artificially conservative. The Financial Stability Board's May 2026 vulnerability report noted that, stripping out adjustments, private credit borrowers' true leverage could be closer to 7.0x against a reported 5.0x to 6.0x — a gap created by contractual definitions rather than underlying business performance.

Recovery Rates Are Sensitive to Covenant Quality

The structural consequence of weak EBITDA discipline shows up most clearly in recovery outcomes. According to Moody's Ratings research published in September 2025, lenders recovered an average of 57% on first-lien covenant-lite loans between 2023 and mid-2025, compared with 66% on the shrinking number of loans that retained financial maintenance covenants. That 9-percentage-point gap in recovery rates — directly attributable to the presence or absence of early-warning covenant mechanisms — is the downstream cost of accepting inflated EBITDA definitions at origination.

Default Rate Backdrop

The credit environment in which this normalization discipline operates is not benign. The private credit default rate in 2025 was estimated by Moody's to range between 1.6% and 4.7% depending on whether distressed exchanges are counted — with distressed restructurings accounting for approximately 65% of all defaults in the first half of 2025. S&P reported the trailing 12-month speculative-grade high-yield default rate above 4% through late 2025, and while S&P forecasts that rate declining toward 4.0% by September 2026, the environment rewards lenders who built real EBITDA buffers into their underwriting rather than those who accepted management projections at face value.

EBITDA Normalization Underwriting Checklist

The following checklist reflects the diligence sequence a credit officer should execute on every direct lending or private credit transaction before accepting any adjusted EBITDA figure in the underwriting model.

1. Anchor to audited or reviewed financials — not management-prepared statements.
   Start with three years of historical GAAP financials. Tie reported EBITDA to the trial balance and bank statements before opening the adjustment schedule.
2. Build your own EBITDA bridge independently before reviewing the borrower's.
   Review the general ledger, payroll registers, and bank statements line by line before comparing to the sponsor's QoE. Pre-formed views are harder to manipulate.
3. Categorize every add-back into Layer 1, 2, or 3 before evaluating it.
   Apply definitional, quality-of-earnings, and normalization categories explicitly. Each layer carries a different documentation and evidence standard.
4. Require four-point documentation for every non-definitional add-back:
   (a) Source documents tying to the exact dollar amount; (b) narrative explaining non-recurrence or owner-specificity; (c) market-rate benchmark where applicable; (d) precedent for treatment in comparable transactions.
5. Test revenue, not just expenses.
   Disaggregate revenue by customer, product line, and period. Identify any non-recurring revenue sources (one-time contracts, government stimulus, pandemic-era windfalls) and subtract them. Borrowers rarely volunteer these downward adjustments.
6. Reject forward synergies without a hard realization roadmap.
   Projected cost savings, unachieved operational improvements, and run-rate revenue from customers not yet under contract should be rejected or applied only with a significant haircut and a clear milestone-based schedule.
7. Apply a cap to your total add-back acceptance.
   Institutional practice limits aggregate accepted add-backs to 20%–25% of reported EBITDA. If the borrower's adjustments exceed this threshold, require an independent Quality of Earnings report from a credentialed third-party accounting firm.
8. Stress-test lender-adjusted EBITDA at −15% and −25% before calculating leverage.
   Covenant thresholds and debt-service coverage ratios should be evaluated against downside scenarios, not just base-case normalized earnings. Confirm the business can service debt at stress EBITDA.
9. Document the delta between borrower-adjusted and lender-adjusted EBITDA in the credit memo.
   The bridge is a core credit deliverable. Every line-item rejection or haircut should be explicitly noted with rationale. This record is critical in any subsequent amendment, restructuring, or workout discussion.
10. Confirm EBITDA-to-cash conversion with a working capital and capex review.
    Normalized EBITDA is not free cash flow. Deduct maintenance capex, normalized working capital changes, and cash taxes to arrive at a debt-service coverage ratio that actually represents liquidity available to lenders.

Key Takeaways for Credit Professionals

• The borrower writes the arithmetic, but the lender owns the outcome. Every leverage covenant is anchored to the EBITDA definition negotiated at origination. Accepting inflated add-backs at closing embeds risk that no subsequent covenant monitoring can fully correct.
• There are three taxonomically distinct types of EBITDA adjustments — definitional, quality-of-earnings, and normalization — and each requires a different evidence standard. Treating all add-backs as equivalent is an underwriting error.
• Add-backs ease reported leverage but do not generate cash. Interest and amortization require liquidity, not adjusted earnings. Supplement EBITDA-based covenant analysis with independent cash conversion and liquidity tests.
• Recovery rates are structurally higher when maintenance covenants are present. Moody's data shows a 9-percentage-point gap in first-lien recovery rates between covenanted and cov-lite structures — a direct consequence of early-warning mechanisms that inflated EBITDA definitions tend to defeat.
• In the current credit environment, conservatism pays. With private credit leverage at 5.3x and true leverage potentially near 7.0x once adjustments are stripped, lenders who underwrite to independently verified, haircut-adjusted EBITDA are structurally better positioned to withstand the distressed restructuring cycle already evident in 2025–2026 default data.