Planning a sale, move, or exit before 2028? Book a 20-min planning call →
Holding QSBS? Get a fixed-fee Section 1202 issue-spotting review →
Washington’s 9.9% income tax is now law. Get the Tax Planning Guide →
Section 1202

The $2M Cash Box: How a Section 1045 Rollover Dies Quietly

By Joe Wallin,

Published on Jul 22, 2026   —   16 min read

Tax Planning

Summary

A §1045 rollover relocates your tax problem onto the replacement company's balance sheet. One $2M rollover, five trajectories — and the maintenance system that decides which one you get.

A Section 1045 rollover moves your QSBS gain into a new company. It does not move your tax problem — it relocates it onto the new company's balance sheet, where most founders never look. The pattern that kills these rollovers is always the same: $2 million lands in a newly formed corporation, the founder gets to work, the cash sits, and two years later a set of rules almost nobody reads has quietly decided whether the founder owes nothing or owes six figures with interest. This post walks one $2 million rollover through five trajectories, from clean to catastrophic, and ends with the maintenance system that keeps you on the right one.

For the mechanics of the rollover itself — the 60-day window, the six-month minimum holding period, basis adjustments, and holding period tacking — see Section 1045 Rollovers: How to Defer QSBS Gains. This post assumes the rollover was executed correctly and asks the question that determines whether it was worth executing at all: what does the replacement company have to be for the next two years?

The Rule Stack: The Replacement Company's Balance Sheet Controls

Three layers of rules apply to the company receiving your rollover money.

The 80% test. Section 1202(e)(1) requires that at least 80% by value of the corporation's assets be used in the active conduct of one or more qualified trades or businesses. This is a balance-sheet composition test, measured continuously — not a spending requirement. Money spent on payroll leaves the balance sheet; the test looks at what remains and asks what fraction of it is working. Pre-revenue activity counts: Section 1202(e)(2) treats start-up activities and research and experimental activities under Section 174 as active conduct, so a solo founder writing code full time is engaged in a qualified trade or business from day one.

The six-month test for the deferral. Under Section 1045(b)(4)(B), the replacement corporation must satisfy the active business requirement during substantially all of the first six months of your holding period in the replacement stock. Fail that window and the consequence is not a lost exclusion at exit — it is an invalid rollover from inception, meaning the deferred gain was recognized in the year of the original sale, with interest running from that year's return due date.

The full-period test for the exclusion. If you later claim the Section 1202 exclusion on the replacement stock, the active business requirement must be satisfied during substantially all of your holding period — which, because of tacking under Section 1223(13), includes the years you held the original stock. "Substantially all" is undefined in the statute and regulations. Practitioners commonly assume something in the range of 80–90% of the period, and whether tacked years from the original issuer count toward the replacement stock's fraction is genuinely unsettled. Neither point is one to plan around; both are points to stay away from.

The Working Capital Rules: Section 1202(e)(6)

Cash is not an active business asset. Section 1202(e)(6) provides the two exceptions that every rollover into a young company depends on. Cash counts as used in the active conduct of a qualified trade or business if it is:

(A) held as part of the reasonably required working capital needs of the business — no time limit, but bounded by what operations actually require; or

(B) held for investment and reasonably expected to be used within 2 years to finance research and experimentation or increases in working capital.

Then the cliff: once the corporation has been in existence for at least 2 years, no more than 50% of its assets can qualify as working capital under these rules — both prongs combined.

Run the algebra on a company whose only assets are cash and active-use assets, and the 50% cap produces a clean planning threshold: after the company's second birthday, cash and investment assets cannot exceed 70% of total asset value. At exactly 70%, the 50% allowance plus the 20% of slack in the 80% test gets you to a pass with zero margin. The comfortable target is cash at or below 50% of value, where the composition test passes regardless. A $2 million cash position therefore needs roughly $860,000 of genuine active-use asset value beside it to sit at the cliff edge — and about $2 million to sit somewhere safe.

Two adjacent limits can independently kill the active business test and belong in the same annual check: portfolio stock in non-subsidiary corporations exceeding 10% of net asset value (Section 1202(e)(5)(B)), and real property not used in the active business exceeding 10% of total asset value (Section 1202(e)(7)).

The Case That Proves It: Owen v. Commissioner

This is not a theoretical trap. In Owen v. Commissioner, T.C. Memo. 2012-21, a taxpayer sold his interest in an insurance-marketing business and attempted to defer roughly $1.87 million of gain under Section 1045 by forming a retail jewelry corporation, J&L Gems, and depositing about $1.9 million of proceeds into its bank account. Over the period that followed, the company purchased sixteen pieces of jewelry for roughly $147,000 — about 8% of its assets. The other 92% stayed in cash. The Tax Court held that J&L Gems never satisfied the active business requirement: 92% cash fails the 80% test outright, and the working capital rules could not rescue it because even the 50% allowance leaves 42 points of cash uncovered. The Section 1045 deferral failed entirely, and the gain was taxable in the year of the original sale.

The numbers in Owen are almost exactly the numbers in every modern founder's version of this fact pattern. What follows is that fact pattern, run five ways.

One Rollover, Five Trajectories

The common setup: you sell QSBS after a 3-year hold and roll $2 million of gain into a newly formed C corporation under Section 1045. Your holding periods tack, so you need 2 more years for the 100% exclusion tier. What happens next depends entirely on what the new company becomes.

Scenario 1: The Executed Plan

At formation, the board adopts a written deployment budget — engineering hires, infrastructure, sales, a reserve — and the company executes it in reasonable approximation. The cash qualifies under prong (B) during the first 2 years because the expectation was real and documented; by the second birthday, spending and asset-building have brought cash below the thresholds. The six-month test is met, the full-period test is met, and at year 5 the exclusion applies. This is the base case the statute was written for, and it is boring by design.

Scenario 2: The Honest Plan That Slips

The founder genuinely intended to deploy the money but never wrote the plan down, and eighteen months in, the cash is largely untouched. The deferral is likely safe: the six-month test looks at the first six months, when the expectation was real, however loosely specified. The exclusion is the problem — the (e)(6)(B) expectation is not tested once at formation; it decays as the facts accumulate. At month 3, an undisturbed $2 million is consistent with a ramp-up. At month 18, the position that deployment within the original 2-year window is still "reasonably expected" is nearly gone, and each non-qualifying month erodes the "substantially all" fraction across the full tacked holding period.

The distinguishing feature of this scenario is that it is fixable in real time. Before the expectation goes stale, the founder can deploy against a real hiring or contracting plan; revise the plan in writing when circumstances have changed — a revised expectation is still a reasonable expectation; or shrink the problem by distributing excess cash out. The one thing that does not work is doing nothing and hoping the two-year mark passes unnoticed. It doesn't.

Scenario 3: The Cash Box Redeemed by Value

The founder spends almost nothing — cloud fees and little else — but builds a product, and 2 years later the company is worth $10 million or more. The 80% test runs on value, not basis, and self-created IP has zero basis but full value. At a $10 million company value with $2 million of cash, cash is 20% of assets and the test passes on raw composition, with no reliance on the working capital rules at all. The early months remain the soft spot — the company was mostly cash by value in year one — but the hindsight evidence now runs in the founder's favor: a $10 million product built through continuous development is strong proof the company was engaged in qualifying research the entire time, and Section 1202(e)(2) treats that as active conduct even pre-revenue.

One warning at exit: if the original stock was issued on or before July 4, 2025, the anti-reset rules keep the exclusion cap at $10 million (or 10x basis) — a rollover does not convert pre-OBBBA stock into the $15 million post-OBBBA regime. Gain above the cap is taxed at the 28% Section 1202 rate plus the 3.8% net investment income tax, plus Washington's capital gains excise for a Washington seller. If the exit will exceed the cap, the time for stacking and gifting strategies is before the sale process starts.

Scenario 4: The Cash Box at Liquidation

Same founder, same minimal spend — but no value gets built, and after 2 years she liquidates for roughly her cash back. This is Owen with a software wrapper. The liquidation price is itself the evidence: if a buyer would pay nothing beyond the cash, the IP was worth approximately nothing at every testing date, the company was a cash box throughout, and the retrospective case that the (e)(6)(B) expectation was ever real collapses. The active business test fails for essentially the whole replacement period — and because it fails in the first six months too, the deferral itself is at risk, not just the exclusion.

The cost, for a Washington founder with a $2 million recognized gain and no exclusion: roughly $476,000 federal (20% long-term capital gains plus 3.8% net investment income tax), plus Washington's capital gains excise — 7% on the gain above the inflation-adjusted standard deduction ($278,000 for tax year 2025) up to $1 million, and 9.9% on the portion above $1 million — roughly another $150,000. Call it $626,000, or about 31% of the gain, on money that a working plan and a functioning balance sheet would have sheltered entirely. Section 1202-excluded gain sits outside the Washington capital gains base; disqualified gain sits squarely inside it.

Scenario 5: The Rollover That Never Should Have Happened

Now assume the honest version of the worst facts: at the outset, the founder did not actually expect to deploy the $2 million within 2 years. She parked it. Prong (B) fails at formation because the expectation it requires never existed, and later success cannot cure it — a test measured at the outset is not retroactively satisfied by good things happening afterward. Prong (A) covers only what operations reasonably require, which for a pre-revenue company burning cloud fees is a small fraction of $2 million. The company fails the 80% test in the first six months, the Section 1045 election was invalid from inception, and the gain belongs on an amended return for the year of the original sale — with interest from that year and accuracy-related penalty exposure.

The fix is structural, and it is available to everyone: Section 1045 permits partial rollovers and multiple replacement issuers within the 60-day window. Gain is recognized only to the extent proceeds exceed the cost of replacement stock. If the new company can credibly absorb $500,000 in 2 years, roll $500,000. Roll the balance into other qualifying companies with genuine capital needs, or recognize it and pay tax at whatever exclusion tier the original stock earned. A rollover sized to the proceeds instead of the plan is not tax planning — it is an invalid deferral with a delayed fuse.

How Much Can You Actually Roll?

The sizing question has a method, not just a principle. Work it in three steps before the wire goes out.

Step 1: Build the capacity number from the bottom up. The amount a replacement company can safely absorb is roughly the sum of three components: the credible 2-year deployment budget — hires, infrastructure, sales, development spend the board would actually approve (this is the prong (B) money); a defensible operating reserve, bounded by projected burn rather than by round numbers (the prong (A) money — 12 to 24 months of projected spend is defensible for a company executing a scaling plan; $2 million against $30,000 of annual burn is not); and the non-cash asset value the company will realistically build over the period, because value in the denominator is what relaxes the cash constraint as the 2-year mark approaches.

A worked example: a founder with $2 million of proceeds and a company whose honest 2-year budget is $1.2 million of hiring and infrastructure, with roughly $300,000 of reserve defensible against projected burn, and real product development underway. Something in the range of $1.5 to $1.7 million rolls comfortably. The remaining $300,000 to $500,000 gets recognized at whatever exclusion tier the original stock earned, or rolled into a different qualifying company with genuine capital needs of its own.

Step 2: Run the math on proceeds, not gain. Section 1045(a) measures the deferral against proceeds reinvested: gain is recognized to the extent the amount realized exceeds the cost of replacement stock. For a founder with near-zero basis, proceeds and gain are nearly the same number and the distinction is invisible. For a seller with meaningful basis it changes the arithmetic — to defer the full gain you must reinvest the full sale proceeds, not just the gain portion. A seller with $2.5 million of proceeds and $2 million of gain who reinvests only $2 million recognizes $500,000 of gain, whatever the intent was.

Step 3: Stress-test the second birthday. Before wiring, project the replacement company's balance sheet at its 2-year mark under the slow case — hiring delayed, milestones slipped, half the budget executed. If projected cash and investment assets would still exceed 70% of total asset value even then, the rollover is oversized regardless of how genuine the plan is, and the fix is cheaper now than in year two: roll less.

Or skip the arithmetic — the calculator below runs all three steps. Enter your proceeds and the replacement company’s honest numbers, and it sizes the rollover, computes any gain recognized now, and stress-tests the balance sheet at the company’s second birthday against the 50% and 70% lines.

Section 1045 Rollover Sizing Calculator

Estimate how much of your QSBS sale proceeds a replacement company can safely absorb under the Section 1202(e)(6) working capital rules, and stress-test the balance sheet at the company's 2-year mark. This tool provides estimates only and is not legal or tax advice.

The Maintenance System

Every scenario above was decided by two things: what the replacement company's balance sheet showed, and what the contemporaneous paper said. Both are controllable. The system has four parts.

Adopt the deployment plan at the outset, in writing, board-approved and dated — at or before the rollover closing. Categories and rough timelines tied to milestones are enough; the date matters more than the precision, because both the six-month test and the (e)(6)(B) expectation are measured against what was expected then. Contingent, milestone-dependent plans are fine. A plan the founder does not actually hold is not — papering an intention that doesn't exist creates a document that misstates intent, which is worse than no document.

Size the rollover to the plan, not the proceeds. This is the single highest-leverage decision in the sequence, because it is the only one that protects the deferral itself rather than just the exclusion.

Refresh the plan annually, in writing. When hiring slows, the product pivots, or a milestone slips, revise the plan and minute it. A revised expectation is still a reasonable expectation; a stale one silently expires somewhere in the back half of year one, and nobody notices until diligence.

Test the balance sheet annually against the plan. Cash as a percentage of total value, with the 50% cap flagged as the company's second birthday approaches; the 10% portfolio-stock and 10% real-property limits alongside. The plan says what was expected; the tested schedule says what happened; together they are the entire evidentiary record. This is the same annual cadence as a QSBS attestation letter, and the two belong in the same engagement — for rollover investors especially, who have no visibility into a replacement company's balance-sheet drift and the most to lose from it.

Rolled gain into a company you don't control?

An annual attestation practice like QSBS Sentinel™ covers replacement stock from day one — the deployment-plan check, the working-capital math, and the letter that documents both. Book a 20-minute call to talk through your rollover before the 2-year mark decides it for you.

Frequently Asked Questions

Does the replacement company have to spend the rollover money within 2 years?

No — and this is the most common misreading of the rules. Section 1202(e) imposes a balance-sheet composition test, not a spending requirement. During the company's first 2 years of existence, cash can qualify in full if it is reasonably expected to fund research or working capital increases within 2 years. After the company's second birthday, no more than 50% of assets can qualify as working capital, which as a practical matter caps cash and investment assets at 70% of total value — and a comfortable position keeps them at or below 50%.

What is the six-month rule in a Section 1045 rollover?

Under Section 1045(b)(4)(B), the replacement corporation must satisfy the active business requirement during substantially all of the first six months of your holding period in the replacement stock. Failing it invalidates the deferral itself: the gain is recognized in the year of the original sale, not the year the failure is discovered.

Can I roll into more than one company?

Yes. Section 1045 permits partial rollovers and multiple replacement issuers within the 60-day window, and gain is recognized only to the extent sale proceeds exceed the total cost of replacement stock. Sizing each rollover to each company's genuine deployment capacity is the cleanest protection against the working-capital trap.

What happens if the replacement company fails the active business test?

It depends on when. A failure within the first six months invalidates the Section 1045 deferral from inception — amended return for the original sale year, interest, and potential penalties. A failure later in the holding period leaves the deferral intact but can destroy the Section 1202 exclusion at exit, and because Section 1045 applies only to sales of stock that is QSBS, a disqualified position cannot be rolled again. The failure is terminal, not deferrable.

Has the IRS actually litigated this?

Yes. In Owen v. Commissioner, T.C. Memo. 2012-21, the Tax Court denied a Section 1045 deferral of roughly $1.87 million where the replacement corporation held 92% of its assets in cash and purchased only about $147,000 of inventory. The working capital rules could not cover the gap, the active business requirement failed, and the gain was taxable in the year of the original sale.

This post is for educational purposes only and is not legal or tax advice. Consult a qualified attorney about your specific situation.

Share on Facebook Share on Linkedin Share on Twitter Send by email

Subscribe to the newsletter

Subscribe to the newsletter for the latest news and work updates straight to your inbox, every week.

Subscribe
}) } })