How to Answer “How do you normalize EBITDA for one-time or non-recurring items in valuation?” in Investment Banking Interviews
In investment banking interview prep, this is a common prompt because it blends accounting judgement with valuation mechanics: “How do you normalize EBITDA for one-time or non-recurring items in valuation?” Interviewers want to see that you can move from reported results to a sustainable earnings base that makes multiples comparable.
A strong answer explains what gets adjusted, why it’s adjusted, and how you support the adjustment using disclosures and a repeatable workflow—without hand-waving.
What Interviewers Look For When You Adjust EBITDA
Interviewers are checking whether you understand that valuation is based on maintainable operating performance. Normalising EBITDA is effectively a mini “quality of earnings” exercise: separating core operations from noise so that the EV/EBITDA multiple you apply (or derive) is apples-to-apples.
They’re also testing practical judgement around EBITDA adjustments—what qualifies as genuinely non-recurring versus recurring “one-offs,” whether an item belongs above or below EBITDA, and how to avoid double counting (e.g., adjusting EBITDA but not the cash impact in working capital/capex, or vice versa).
Finally, they’re assessing communication: can you explain your logic clearly, tie it to valuation techniques (trading comps, precedent transactions, LBO), and show you’d document assumptions like an analyst—using notes, management guidance, and footnotes rather than intuition alone.
Normalize EBITDA for Valuation: Step-by-Step Framework
- 1
Step 1: Define the goal and the EBITDA starting point (reported vs run-rate)
Start by stating that to normalize EBITDA for valuation you need a sustainable run-rate measure that reflects ongoing operations and is comparable across companies and time periods. Specify your starting point (LTM reported EBITDA, last fiscal year, or consensus/NTM) and confirm consistency with the multiple you’ll use (LTM multiple with LTM EBITDA; NTM multiple with NTM EBITDA).
Then clarify what you mean by EBITDA in this context: operating profit before D&A, excluding non-operating income/expense, and typically pre-exceptionals depending on the reporting framework. The key is to be consistent: if comps are “EBITDA before exceptional items,” your target’s EBITDA should be adjusted on the same basis.
Close the step by framing the output: a bridge from reported EBITDA to adjusted/normalized EBITDA, supported by specific line items from the income statement, MD&A, footnotes, and (in deals) the management presentation/QoE workstream.
- 2
Step 2: Identify candidate one-time items and classify them (add-backs vs deductions)
Walk through how you find non-recurring items: start with “exceptional,” “non-recurring,” “restructuring,” “impairment,” legal settlements, gains/losses on asset sales, integration costs, and unusual revenue reversals. Use disclosures to understand what happened and whether it’s operational, financing, or non-cash.
Then classify each item into one of three buckets:
- Add-backs to EBITDA (increase normalized EBITDA): truly one-off costs that depress current EBITDA but won’t persist (e.g., a discrete plant closure cost).
- Deductions from EBITDA (decrease normalized EBITDA): one-off gains or temporary benefits (e.g., a large insurance recovery or gain on property sale) that inflate EBITDA.
- Reclassifications: items that should be moved below EBITDA (or removed from operating results) for comparability.
Emphasise judgement: frequent “one-time” adjustments every year are a red flag—those may be recurring and should stay in run-rate, or be normalised to a steady-state level rather than fully excluded.
- 3
Step 3: Quantify and normalise properly (tax/ownership, run-rate, and comparability)
Explain how you quantify each adjustment and avoid mechanical mistakes. Use gross amounts consistent with EBITDA (pre-tax, pre-interest), and ensure you’re not mixing cash and non-cash concepts incorrectly: impairments are non-cash but still distort operating profit and should typically be removed for a run-rate EBITDA.
If an item is partially recurring, normalise to a steady-state run-rate instead of removing it entirely—e.g., if litigation costs are likely to continue at a lower level, adjust to the expected ongoing level. For cost savings, be conservative: only include synergies or savings in EBITDA if they’re credible, timed, and consistent with how the market would underwrite them.
Finally, ensure comparability for the valuation technique you’re using: for trading comps, align with peers’ “adjusted EBITDA” definitions; for precedent deals, be mindful that transaction EBITDA may reflect sponsor-friendly add-backs; and for LBOs, sanity-check that adjusted EBITDA converts to cash (maintenance capex and working capital still matter).
- 4
Step 4: Sanity-check the result and document the bridge for the valuation output
Close with checks and how you’d present it. Sanity-check normalized EBITDA by comparing adjusted margin versus history and peers, and by testing sensitivity: what happens to implied EV and equity value if you exclude/retain the top 1–2 adjustments?
Confirm you haven’t double counted with other parts of the model: if you adjust EBITDA up for removing a one-time expense, you shouldn’t also add back the same cash outflow elsewhere. If you remove a gain from EBITDA, make sure the cash proceeds (if any) are treated appropriately (often as non-operating and not part of run-rate).
Document the outcome as a clear bridge (Reported EBITDA → adjustments by category → Normalized EBITDA) with sources (10-K/annual report footnotes, investor deck, management guidance). This is what interviewers expect from analyst-level work: transparent assumptions that can be audited and challenged.
Sample Answer (Analyst-Level) for Normalized EBITDA
To normalize EBITDA for valuation, I start with reported LTM EBITDA and build a bridge to a maintainable run-rate figure that’s consistent with the multiple I’m applying. The goal is to remove items that distort operating performance so EV/EBITDA is comparable across companies.
Then I identify potential one-time or non-recurring items from disclosures—things like restructuring and integration costs, legal settlements, impairments, gains or losses on asset sales, or unusual insurance recoveries. I classify each item as an add-back if it’s a truly discrete cost that won’t repeat, or a deduction if it’s a one-off gain that inflated EBITDA. If something is labelled “non-recurring” every year, I treat it as recurring or normalise it to a steady-state level rather than removing it entirely.
When I quantify the adjustments, I keep it consistent with EBITDA—pre-interest and pre-tax—and I’m careful not to double count cash impacts elsewhere in the model. I also make sure the definition matches the market data I’m using: if comps are based on “adjusted EBITDA,” my target’s EBITDA needs to be adjusted on the same basis.
Finally, I sanity-check the result by looking at margins versus history and peers and running sensitivities on the largest adjustments to see how much they move implied value. I’d present the work as a clean EBITDA bridge with sources so the assumptions are transparent and defensible.
- Lead with the objective: maintainable run-rate EBITDA for comparable multiples.
- Name both directions of adjustment (remove one-off costs and one-off gains).
- Flag the “recurring one-time” issue to show judgement.
- Tie the definition back to the valuation technique (trading comps / precedents) and data consistency.
- End with documentation + sanity checks (bridge, margins, sensitivity).
Common EBITDA Adjustments Mistakes in IB Interviews
- Treating management’s adjusted EBITDA as automatically correct, without checking frequency and underlying drivers.
- Only adding back costs and forgetting to deduct one-off gains that temporarily boost EBITDA.
- Double counting: adjusting EBITDA for an item but also adjusting working capital/capex or other cash lines for the same item.
- Using an LTM multiple with an NTM EBITDA (or vice versa), which creates a mismatch in the valuation output.
- Stripping out costs that are clearly part of normal operations (e.g., ongoing stock-based compensation or regular “restructuring”).
- Failing to explain *source support* (footnotes/MD&A) and presenting adjustments as intuition rather than analyst work.
Follow-Ups in Investment Banking Valuation Interview Questions
What are typical non-recurring items you’d adjust for in an EBITDA bridge?
Common ones are restructuring/integration costs, litigation settlements, impairments, gains/losses on asset sales, and unusual one-time income like insurance recoveries—always validated in footnotes.
Would you add back stock-based compensation when normalizing EBITDA?
It depends on the market convention for that sector and the comps dataset, but I’d be cautious because it’s often recurring; if excluded, I’d highlight the rationale and impact.
How does normalized EBITDA affect an EV/EBITDA valuation?
Because EV is typically the multiple times EBITDA, any adjustment to EBITDA directly scales implied enterprise value, so the biggest adjustments should be sensitivity-tested.
How do you handle run-rate cost savings or synergies in normalization?
I’d only include them if they’re credible and clearly timed; otherwise I’d keep EBITDA conservative and show value impact separately via sensitivities or deal-specific cases.
How is this different for precedent transactions versus trading comps?
Precedent deals often use more aggressive “transaction EBITDA” add-backs, so I’d normalise carefully and reconcile to what the market actually underwrote, not just headline numbers.
Practice Drills for Valuation Techniques and Run-Rate EBITDA
- Build a 60–90 second “bridge” script: start point (LTM EBITDA) → categories of adjustments → sanity checks and consistency with the multiple.
- Practise with two mini-cases: one where EBITDA is inflated by a one-off gain, and one where it’s depressed by restructuring; explain both directions confidently.
- Prepare 5–7 common adjustments and one sentence on why each is (or isn’t) truly non-recurring—this comes up in ib technical interview questions.
- When you practise on AceTheRound, ask for feedback on (1) whether your adjustments are defensible and (2) whether you tied them to valuation techniques and comparability.
Ready to practice with AceTheRound?
Create an account to unlock AI mock interviews, feedback, and the full prep library.