Financial modeling — statement construction & verification.

The seller's financial statements arrive in a deal data room as a finished narrative. The buyer's job is to take them apart. Every income-statement line gets verified against source documents. Every add-back gets defended or rejected. Every ratio gets benchmarked against agency-industry norms. The cluster covers the five statement-level work products that produce a defensible model: normalized EBITDA, the buyer's forward pro-forma, balance sheet forensics, cash flow analysis, and the ratio benchmark layer. Run in sequence, they take you from the seller's narrative to a model you can underwrite.

Rebuild the number from source.

EBITDA verification is the foundation. The seller's pitch deck and CIM lead with an "Adjusted EBITDA" figure built from a list of normalizing add-backs — owner compensation above market, personal expenses run through the agency, one-time legal fees, deferred technology spend, family-member payroll, vehicle expenses, club memberships. Each add-back, taken individually, may be legitimate. Taken collectively, they often inflate EBITDA by 25–40% over the unadjusted number. The buyer's job is to test every one.

• Owner compensation normalization. Replace the seller's actual W-2 with a market-rate replacement salary for an operating principal at the agency's revenue tier. Industry benchmarks land in the $150K–$250K range for sub-$5M agencies; higher for larger books. The delta between actual and market is the legitimate add-back — the rest stays inside EBITDA.
• Personal expense unwinds. Cellular phones, country-club memberships, family travel, vehicle leases. These are real cash leaving the business that the buyer won't continue. They are legitimate add-backs only if the buyer can substantiate the line items with cancelled checks, credit-card statements, or vendor invoices. "Trust the seller's spreadsheet" is not substantiation.
• One-time items. Legal fees from a non-recurring lawsuit, severance from a producer separation, AMS conversion costs. Genuine if the source document confirms it was non-recurring. Suspect if the same "one-time" item appears in two consecutive years.
• Deferred maintenance. Technology spend the seller postponed to flatter earnings. The buyer should subtract this from EBITDA, not add it back — and the subtraction is what funds the tech-debt closing adjustment.

A defensible Quality of Earnings report — discussed in the parent financial due diligence cluster — packages all of this. Without one, the buyer's normalized EBITDA is the buyer's word against the seller's. With one, it's a verified number both sides can negotiate against.

Forward, not backward.

The pro-forma is the buyer's forward-looking P&L for the combined entity — the seller's book operated under the buyer's cost structure, with synergies counted and one-time integration costs absorbed. It is not the historical seller statement with an EBITDA adjustment line at the bottom.

A pro-forma built on unverified historical EBITDA is a forecast of fiction. The sequence matters: clean the base year first, then project forward.

The construction sequence:

The pro-forma is the document the buyer's lender will scrutinize line-by-line. It is also the document the buyer's investment committee approves the deal against. A pro-forma that doesn't reconcile to the verified historical base is a credibility problem the buyer carries through every subsequent diligence conversation.

Asset reality, liability completeness.

The income statement gets the attention; the balance sheet hides the deal-killers. Three balance-sheet checks every buyer should run:

• Premium trust account integrity. Agencies that mishandle trust funds — commingling, late carrier remittance, premium fronting — carry contingent regulatory exposure that can exceed the purchase price. The single highest-stakes balance-sheet audit. The check is reconciling the trust account ledger against carrier statements for at least 24 trailing months. Discrepancies are red flags that warrant carrier-level inquiry before LOI.
• Receivables aging and collectability. Premium receivables aged past 90 days are usually uncollectable. Sellers sometimes record them at face value on the balance sheet, inflating book value. The buyer's adjustment: write down aged receivables to realistic collection rates, typically 60–70% for 60–90 day, near-zero past 120 days.
• Producer-book ownership intangibles. If the seller's balance sheet carries goodwill or book-of-business intangibles, the buyer needs to understand what underlies them. Producer-owned books, free-agent risk, and unenforceable non-competes can mean the intangible asset transfers in name but not in economic reality.

The fourth category is hidden liabilities. Pending E&O claims, regulatory examinations, carrier termination notices, deferred capital expenditures, lease tail obligations. Buyers who fail to surface these before close inherit them at full face value post-close. The reps and warranties package can recover some of it, but indemnification baskets and caps mean the buyer absorbs the first layer.

Operating vs. financing, working-capital cycles.

Insurance agencies have unusual cash-flow profiles driven by commission timing. New business commissions flow upfront; renewal commissions arrive at the renewal date; contingency commissions arrive in lump sums months or years after the underlying production. A profitable agency can be cash-poor — and a marginal agency can look cash-rich during a contingency-payment month.

The buyer's cash-flow model needs to separate the three streams clearly:

• Operating cash flow ex-contingencies. The recurring base. This is the cash the business produces in a "normal" month from book renewal commissions and new-business production. The buyer's pro-forma debt service should be sized against this number, not the contingency-inclusive number.
• Contingency cash flow. Lumpy, timing-dependent on carrier loss ratios. Sellers sometimes amortize contingencies smoothly across the year on their reported P&L, hiding the timing risk. The buyer's model should treat contingencies as a separate cash-flow stream with explicit timing — and discount them for variability.
• Working-capital cycle. Premium receivables, carrier payables, producer commission timing. Agencies that float premium to carriers (common but not best practice) carry working-capital exposure that scales with premium volume. A growing book can mean a worsening working-capital cycle even as profitability improves.

Cash flow stress-testing is where lender comfort gets built. A pro-forma that shows debt service coverage of 1.4× on operating-cash-flow-ex-contingencies is a deal a senior lender can approve. The same deal at 1.4× including contingencies is a deal the underwriter will pick apart.

Benchmark against GPS norms.

The fifth layer is ratio benchmarking against the agency-industry GPS dataset — revenue per employee, EBITDA margin, producer productivity, expense ratios, growth rate, retention rate. The ratios that matter are the ones that diverge materially from the GPS norm for the agency's revenue tier. Divergence is the signal that something deserves a deeper look — favorable or unfavorable.

The standard ratio panel for an agency deal:

• Revenue per employee. Sub-$5M agencies typically land at $140K–$200K per FTE. Above-norm signals either operational efficiency (good) or under-staffing that will reverse post-close (bad). Below-norm signals over-staffing the buyer will need to address — and the post-close severance becomes a deal-cost line.
• EBITDA margin. Mid-teens to low-20s percent is typical for personal-lines-heavy books; 25–35% for commercial-lines specialists; higher for niche programs. Margins materially above the tier norm warrant scrutiny — typically the result of under-investment in producers or technology that will surface post-close.
• Producer productivity (book per producer). Industry norm: $400K–$700K commission per producer. Below-norm producers are retention risks (the seller has been carrying them); above-norm producers are key-person risks (their departure breaks the book).
• Retention rate (policy and premium). Policy retention 92–95% is the industry baseline. Premium retention typically runs 1–3 points higher because the book grows through rate. The two numbers should be reported separately; sellers who conflate them are sometimes hiding policy attrition behind rate-driven premium growth — the "leaky bucket" pattern covered in the customer DD cluster.
• Growth rate decomposition. Top-line growth is the sum of new business, rate, and retention. The buyer needs all three numbers — top-line alone hides which one is driving the result. A 7% top-line growth with 5% rate and -3% new business is a fundamentally different agency than a 7% top-line growth with 2% rate and 5% new business.

The five-spoke financial-modeling workflow — normalize, project, validate, stress-test, benchmark — is the foundation everything else in the cluster builds on. Without statement integrity, the buyer's risk assessment (the sibling cluster covering revenue concentration, premium trust audit, quality of earnings, and hard-market valuation) lacks a reliable base. With it, the buyer has a model that the lender, the investment committee, and post-close operating reality will all reconcile against. The Pillar — Financial Due Diligence for Buyers — covers the broader cluster.