How to Analyze Add-Backs: What SBA Lenders Actually Accept

Make sure you're accounting for all true owner replacement costs and verify the revenue isn't artificially inflated.

That's a real buyer quote. Mid-deal, mid-diligence, posted after running the numbers and feeling something was off. They were right to be nervous.

There are two versions of add-back math in every deal. Only one of them matters to the person writing the check.

The broker builds SDE from the seller's financials and adds back everything that can plausibly be called owner-specific or non-recurring. The lender builds adjusted EBITDA from the federal tax return and accepts a shorter list. Those two numbers are rarely the same. On Main Street acquisitions, the gap between broker SDE and lender-calculated adjusted EBITDA runs 15–40% — based on what practitioners see consistently across deals at underwriting.

That's not a rounding error. It's the difference between a deal that clears DSCR and one that doesn't.

Why Add-Backs Exist — and Why Brokers Use Every One of Them

The broker's job is to maximize SDE. The lender's job is to strip it back down.

Owner-operated businesses run personal expenses through the P&L. The owner's salary is often set by what the business can pay, not market rate. One-time legal fees, a building repair, a piece of equipment — these show up as expenses and depress the trailing numbers. Add-backs exist to normalize for all of that. To show what the business actually earns when you strip out the owner's personal footprint and events that won't repeat.

That's legitimate. The problem is that brokers — doing their job — add back everything that can defensibly be called an add-back. Their incentive is to maximize the presented SDE because SDE drives valuation. Your job, and the lender's job, is to figure out which of those add-backs will hold up when someone checks the tax return.

The broker's recast is a starting point. The lender's adjusted EBITDA is the number that determines whether you get the loan.

Most buyers don't understand how different those two calculations are until they're sitting across the table from an underwriter who's asking questions they can't answer.

The Four Add-Backs That Come Out First

These get stripped before the underwriter asks a follow-up question.

If they're on the broker's add-back schedule, plan on losing them.

1. Meals and entertainment

Automatically removed. SBA underwriters treat meals and entertainment as discretionary regardless of what the business does. Even in industries where client entertainment has a legitimate case, the documentation burden to retain this add-back is high enough that most deal teams don't attempt it. It comes out. Model accordingly.

2. Personal auto

Vehicle expenses run through the business are stripped unless the business genuinely requires a company vehicle for core operations — and the operations are obvious from the tax return. A pest control company with a fleet of service trucks: different conversation. An accounting firm whose owner leases a luxury SUV through the business: stripped. The lender's test: would a new owner incur this same expense to run the business? Personal vehicles don't survive that.

3. Personal cell phone

Removed. This one rarely creates debate because the amounts are small — but it's worth noting because brokers include it routinely and lenders strip it routinely. If the add-back schedule shows $2,400/year in owner cell phone, that $2,400 doesn't go into the lender's model.

4. Excess owner compensation — only the portion above market replacement

This is the add-back that generates the most confusion and the most deal math errors. The full owner draw is never the add-back. Only the amount above what it would cost to hire a functional replacement.

Here's how it works: a business owner draws $280K per year. The business needs an operator — a GM or a working owner — and someone in that role in that market earns $110K. The add-back available is $170K, not $280K. SBA underwriters apply market rates for the functional role based on geography and industry. They're asking: what would a new owner actually have to pay someone to do what this owner does?

When brokers present the full owner draw as an add-back, they're not being dishonest. They're being optimistic. The lender will apply the replacement rate. If your model uses the full draw, your DSCR is overstated.

Add-Backs That Survive Underwriting

Not everything gets stripped. These are the ones that hold.

Depreciation and amortization

Always accepted. This is the cleanest add-back in any package because it appears on the tax return and doesn't require explanation. Depreciation is a non-cash charge — it doesn't affect the business's ability to service debt. Every lender adds it back. This is the foundation of adjusted EBITDA.

Interest expense

Added back to get from net income to EBITDA. The existing debt service is replaced by the new acquisition loan's debt service in the DSCR calculation — so existing interest isn't double-counted. Standard treatment; no documentation required beyond the return.

One-time non-recurring expenses — with documentation

This is where deals get separated. A one-time legal fee, a capital expenditure that won't repeat, a loss from a property sale — these can survive underwriting. But the word "one-time" requires proof.

What documentation looks like in practice:

• Legal settlement: the closing agreement, confirmation the matter is resolved, confirmation the new owner isn't inheriting the liability
• Capital expenditure: invoices showing what was purchased, explanation of why the asset won't require replacement soon
• Repair expense: contractor invoices, explanation of the specific event, evidence it was truly non-recurring

Without documentation, one-time items get treated as operating expenses. The lender's position: if you can't prove it was non-recurring, it's recurring. Come prepared or lose the add-back.

Owner benefits — selectively

Health insurance, retirement contributions, and similar owner-specific benefits can be added back — the logic being these expenses go away under new ownership if the new owner doesn't structure the same benefits. SBA generally accepts these when they're identifiable on the tax return and the amounts are reasonable.

Above-market owner compensation

The amount above market replacement rate is accepted — with documentation. If you're presenting $170K as an add-back on a $280K owner draw, you need a clear explanation of what the market replacement looks like. A job description, a salary range for the role and geography. Lenders will set their own replacement rate if you don't set it first. Theirs is often lower than yours.

How Lenders Actually Calculate Adjusted EBITDA

Not from the CIM. From the tax return. In this exact sequence.

Landscaping company, asking $680K. Broker CIM: $248K SDE. Seller draws $175K, runs the usual personal expenses through the business. The Schedule C tells a different story.

Both clear. $47K evaporated between the broker's number and the lender's, and the deal still works. That's a deal with cushion.

Then the lender asks for the invoice on the $22K repair. The seller doesn't have one.

Lender EBITDA drops to $179K. DSCR: 2.17x. Still fine. But now the underwriter looks at the owner comp. The seller manages the crews, does all the client relationships, handles estimating. The replacement rate the lender uses is $110K — not $95K. That call is theirs to make. Add-back drops from $80K to $65K. EBITDA: $164K. DSCR: 1.99x.

Still clears. This deal has cushion even when two things go wrong simultaneously.

The one that doesn't have cushion looks like the scenario at the end of this post — broker SDE $320K, lender strips $135K, DSCR goes from 1.35x to 0.85x. Same mechanism, no margin for error.

Run this math before you're attached to a price. Not to find a reason to walk — to know which deal you're actually buying.

The Add-Back Defensibility Test

Before you submit a financing package, run every add-back on the broker's schedule through this test. Each item needs to pass all three questions to be defensible in underwriting.

A useful heuristic: if you're uncertain whether an add-back will survive, model the deal without it. If the deal works without the questionable add-backs, you have a buffer. If it only works because you're counting every dollar the broker listed, that buffer doesn't exist — and any one item getting stripped breaks the deal math.

When Add-Backs Get Stripped Mid-Underwriting

This is the scenario that costs people six months of work.

The deal looked fine at LOI. Your model showed 1.35x DSCR. You used the broker's SDE, the add-backs seemed reasonable, you submitted the package. Then:

The lender removes $60K in meals, entertainment, and personal auto. They apply a market replacement rate to owner comp and cut $45K from that add-back. They ask for documentation on a $30K "one-time legal expense." That documentation doesn't exist, so the $30K stays as an expense.

Net result: $135K in add-backs removed from the broker's SDE. Adjusted EBITDA drops from, say, $320K to $185K. Proposed debt service on a $900K acquisition loan at current SBA rates runs roughly $100K annually in this example. DSCR goes from 1.35x to 0.85x. The deal doesn't close.

This scenario — a deal that clears LOI and breaks at underwriting — is the most common way self-funded searchers lose six months of work and the earnest money that went with it. The mechanism is almost always the same: add-backs the lender won't accept, applied to a business that doesn't clear DSCR without them.

For a deeper look at exactly how DSCR requirements work and the floors you need to clear, that post walks through the lender-by-lender thresholds and the global DSCR calculation.

Understanding how brokers inflate SDE gives you the seller-side view of why these gaps exist — and how to spot the pattern early, before you're deep in diligence.

The fix: build your model from the tax return, run the defensibility test on every add-back line before LOI, and run the deal math the way the lender will run it. That one step catches most of the problems before they become expensive ones.

FAQ

Disclaimer