The complete story edition: 50 chapters and the conclusion. One idea per frame. Tap anywhere on the right to advance, left to go back.
One tax bargain built this whole system. Nine frames and you'll have it.
Jordan sells stock on January 15, 2027. The gain: $750,000.
The Internal Revenue Service (the IRS) would like its share on his 2027 return.
Congress offers a trade: invest the gain, defer the tax.
Defer means pay later, not now. The 2017 tax law created the offer: move a capital gain into a designated low-income area and the clock pauses.
180 days
The window. Gain into a fund, generally within 180 days of the sale, or the deal is off. The tax code is generous, not patient.
The money can't go straight to a building. It goes through a Qualified Opportunity Fund, a QOF.
A QOF is just a corporation or partnership set up to invest in zone property. Jordan buys a piece of the fund, not the building.
An Opportunity Zone is a census tract, one of the small neighborhood-sized areas the Census Bureau draws. Governors nominated low-income tracts, and Treasury certified 8,764 of them in 2018.
The money moves in a line: investor, fund, business, property.
Most funds own an operating business that owns the property. The law calls it a Qualified Opportunity Zone Business, a QOZB. Yes, the acronyms are stacking up. That was the last big one.
Patience pays three ways.
Pay the deferred tax later. At year five, 10 percent of the deferred gain comes off before the bill is figured. Hold ten years, and the fund investment's own growth can be tax-free. Chapter 35 runs the fine print.
Alex runs the fund. Kim checks his work. Maria keeps the books.
Kim is the CPA, a certified public accountant. Add Devon on property, Marcus on construction, Frank on payroll, and investors Bob, Jordan, Maya, and Dana. Same cast, whole book.
Fund, business, project, property, site. Five words. Not synonyms.
Alex calls Main Street 'one project.' The IRS sees three separate things.
A fund: Main Street Opportunity Fund I, LP, a limited partnership. A business: Main Street Apartments. And a property, sitting at a site, inside a zone.
A project is a plan. An entity is a taxpayer.
Every duty in this book attaches to a legal entity. Project names are for pitch decks.
A fund can hold exactly three kinds of qualifying property.
Shares in a zone business, a partnership stake in one, or property the fund owns itself. That's the whole menu.
Most funds take the indirect route: the fund owns a business, and the business owns the property.
Owning property directly is allowed. It also moves every business-level duty onto the fund itself. Choose with your eyes open.
An LLC is a label from state law. It is not a tax answer.
LLC: limited liability company. For federal tax, that same LLC could be a partnership, a corporation, or ignored completely. The tax code does not care what your lawyer named it.
Every ownership arrow needs a date.
Structures change mid-year. An undated chart can't answer a June 30 question. And June 30 questions are coming in Chapter 15.
Draw two maps: legal and reporting.
One shows who owns what. The other shows who reports what, and when. Date both. Then, and only then, open a form.
Today's filing world: 8996, 8997, 8949. Learn the person, then the number.
Three forms run the current program.
8996 · 8997 · 8949
Each one follows a different person. That's the trick to remembering them.
Am I a real Opportunity Fund, did I pass the 90 percent test, and where exactly is my property?
Down to an 11-digit census tract number. Attached to the fund's own tax return, every year.
Every deferred dollar, tracked year after year.
Opening balance, changes, closing balance. Until the gain finally comes home.
Code Z starts the deferral. Code Y ends it.
One letter in, one letter out, filed with the investor's own return. The entire deal hangs on two letters of the alphabet.
One sale shows up in three places. None of them is a duplicate.
Bob sells 40 percent of his stake. The fund reports the exit, his 8997 balance drops, and his 8949 finally pays tax on the deferred gain.
The business on the ground files none of the three.
It just emails numbers upstream when somebody asks. No form, no deadline, no rule. Hold that thought for exactly one chapter.
The printed forms are not the final map.
Form 8996 was last revised in December 2021. The 2025 law wrote new questions the paper hasn't caught up with yet. Paper is slow.
Hold three lanes: the fund, the investor, the gain.
Every new rule you're about to meet lands in one of those lanes.
Short version: the program got renewed, and the receipts came due.
The 2017 law shipped with a blind spot.
Billions flowed into 8,764 zones. The forms captured almost nothing the public could see. No industries. No housing counts. No jobs.
of fund filings reported potentially inaccurate investment information, per the Treasury Inspector General for Tax Administration (TIGTA), the tax system's internal auditor, in its 2022 program audit. The report's own title said the quiet part: additional actions are needed.
So, was the program working?
Nobody could say. Researchers and watchdogs asked for data on investment, housing, and jobs. The statute never required any. Awkward.
Dec 31, 2026
The pressure valve. Every old deferral comes due that day, and the original zone designations run out at the end of 2028, a year earlier for Puerto Rico. Congress had to renew it, fix it, or watch it lapse.
July 4, 2025
The One Big Beautiful Bill Act, Public Law 119-21, becomes law. Yes, that is its real name. Opportunity Zones become permanent.
Permanence was traded for transparency.
Three new sections of the tax code: 6039K, 6039L, and 6726. The § symbol you'll keep seeing just means 'section.' Plus a mandate to publish, in aggregate, what the zones actually produce.
The deal was renewed. The price was data.
The rest of this book is the price list.
OZ: Opportunity Zones. 2.0: the 2025 rewrite. Eight changes, one frame each, and a deep-dive chapter for every one. Chapter tags from here on: Law is enacted rules, Advice is OZfile practice, Pending means Treasury hasn't ruled yet.
One 2018 map expiring in 2028 becomes a fresh map every ten years.
First new cycle: July 1, 2026, with a stricter low-income test. Chapter 6.
of the 8,764 current zones are entirely rural. New tier, better terms: the year-five reduction runs 30 percent of the deferred gain instead of 10, and the renovation threshold is halved. Chapters 25 and 26.
The one-time 2026 cliff becomes a rolling five-year clock.
For new investments, the deferred tax comes due at year five, after the year-five reduction lands. Ten years still earns tax-free growth, and there's now a thirty-year outer limit. Chapter 35.
Form 8996's short quiz grows into the §6039K questionnaire.
Asset values on both test dates, one block per qualifying investment in a business, industry codes (NAICS, the North American Industry Classification System), tract locations, housing unit counts, and employment. Chapters 19 and 27.
§6039L: the operating business must hand its fund a written statement.
The first federal reporting duty a QOZB has ever had. The email era is over. Chapter 29.
Sell your stake, and the fund must send you a written statement.
Your name, your dates, the amount you disposed of. Think of it as a 1099's new cousin. Chapter 33.
Per day for a late or incomplete fund return, at the level set for returns filed in 2027. Yes, per day. Capped at $10,000, or $51,000 for funds with gross assets over $10.23 million. The statute's base is $500, indexed annually. The written statements carry their own separate penalties. Chapter 7.
Treasury must now publish what the zones produce.
Industry, housing, and jobs data, in aggregate, on the record. The blind spot from Chapter 4 gets glasses. Chapter 28.
Same bargain. New receipts.
Next: the map itself. The zone you picked in 2018 now has an expiration date and a renewal cycle.
The 2018 map is expiring. Here's how the new one gets drawn, and why your street address suddenly has paperwork.
Riverbend Works shows up on Treasury's eligible list for the 2027 cycle.
Eligible is not designated. It means the tract passed the income test and may be nominated. Keep the champagne corked.
Every ten years, a fresh determination of which tracts qualify.
The first decennial date (decennial: every ten years) was July 1, 2026. Governors nominate from that list, and the new zones go live January 1, 2027.
The new low-income test is narrower than 2017's.
Newer census data, tighter thresholds. Some 2018 zones will not requalify under the new math.
A 2018 designation does not roll into the new map. The tract must qualify and be nominated again, or its zone status ends with the old map.
Dec 31, 2028
When the original designations generally expire; Puerto Rico's generally end December 31, 2027. Existing investments keep conditional continued treatment, and property acquired after 2026 in an old zone has only narrow paths in. Chapter 24's bridge is one of them.
The statute says property acquired 'after' the applicable start date.
Each zone generation has its own start date, and a purchase qualifies against the map that covers it. For property landing in a 2018 zone after 2026, three narrow paths remain: a new-map acquisition, Chapter 24's transition bridge, or the ordinary-course replacement rule. The closing date starts the analysis; it no longer ends it.
Every property file gets a designation record.
Which tract, which list, which map generation, verified on which date. Ten years from now nobody will remember. The file will.
The stakes chapter. Read it once now, calmly, instead of once later, frantically.
Alex sees the words 'per day' next to a dollar figure and does what most sponsors will do.
Panic first, read second. Let's reverse the order.
Three penalty tracks, and each runs its own meter.
The fund's annual return, the two written statements, and the 90 percent asset test itself. Step one of any penalty question: name the track.
§6726: $510 per day for a late or incomplete fund return.
The 2027 filing-year figure. Capped at $10,000 per return, or $51,000 for a fund with gross assets over $10.23 million on the last day of its year.
Intentional disregard: $2,550 a day, and the caps jump.
For 2027 filings, up to $51,000, or $255,000 for a large fund. Intentional disregard is a factual and legal determination the IRS has to establish, not a bigger setting on the calculator. Chapter 43 covers responding if a notice ever argues it.
$500 → $510
The statute wrote $500; the IRS's annual inflation update (Rev. Proc. 2025-32) sets $510 for returns filed in 2027. Every figure in this lane indexes yearly, and the Internal Revenue Bulletin publishes the current set. Quote the current figure, not the statute's.
Written statements run on a different meter: per statement, not per day.
Congress filed both new statements under the existing payee-statement penalty rules in §6722: $340 per statement for 2027, reduced when corrected quickly. Different lane, different math, different fix. Never borrow a correction rule across lanes.
Failing the 90 percent test is its own track, priced in Chapter 18.
Different statute, different math: monthly shortfalls, an interest-style rate, its own waiver. A notice can involve one track or several; name each before touching a calculator.
Reasonable cause is real relief, not a magic phrase.
§6724 waives the return and statement penalties when the failure had reasonable cause and no willful neglect. Lane three's waiver lives in its own statute. All of them are argued with records, and the records get built during the year, not after the notice arrives.
A notice says the return was 12 days late. 12 times $510 is $6,120.
Verify the duty, verify the dates, check the cap, then decide whether reasonable cause fits. In that order. Chapter 43 runs the full workflow.
Spoiler: it started before you opened this book.
First filings land in 2027, so Alex plans to start in 2027.
Reasonable. Also wrong. The first calendar-year filings are expected in the 2027 cycle, on forms and dates still pending. The facts are 2026.
The new rules apply to taxable years beginning after July 4, 2025.
That's the enactment date itself. For a calendar-year fund, the first covered year is 2026, the one currently on your wall calendar. A fiscal-year fund is in as soon as its own taxable year begins after that date, which for some meant late 2025.
Reporting year and filing season answer different questions.
The reporting year creates the facts. The filing season packages them. Confuse the two and you'll arrive in 2027 to package facts nobody collected.
June 30, 2026
A calendar-year fund's first measurement date of the new era. It has already happened. The snapshot exists whether or not anyone took it.
Fiscal-year funds run different clocks.
Your dates come from your fund's own taxable year: the end of month six and the end of the year. Borrow another fund's calendar and you will measure the wrong day.
New zones start January 1, 2027. Reporting is already underway.
The law carries several effective dates. The designation clock and the reporting clock are different clocks, and the reporting one is already ticking.
A missing form does not pause a legal duty; for now the statute is the instruction sheet. And if June 30 already slipped past you, breathe. Chapter 36 is the rescue chapter, and reconstruction is allowed.
The book's operating thesis in one chapter. Everything after this is technique.
Alex used to wait for Kim's year-end request email. Then the scramble.
Bank statements, spreadsheets, one frantic week in March. That workflow just retired.
The tax form is the ending, not the story.
A return compresses a year of decisions into boxes and percentages. Tidy, and completely silent about how any number got there.
Every reported fact should answer five small questions.
What it is, which document it came from, who holds that document, which date it speaks to, and what proves it. Miss one and the fact is a rumor with a dollar sign.
The sponsor coordinates the facts. He does not become a tax preparer.
Right structure, known source owners, records that arrive on time. That is the entire sponsor job description, and it is plenty.
The CPA reviews the path, not just the total.
A number that cannot show where it came from does not pass review, however round and reassuring it looks.
A sound process starts before filing season.
Evidence gets gathered while things happen, not excavated afterward. Next up: the pipeline that carries it.
The pipeline: how a June snapshot becomes a March filing without the March scramble.
June 30 does not produce a tax return. It produces a snapshot.
One as-of picture of the fund's assets. The return is months away. The facts freeze today.
A test date freezes facts, not paperwork.
You can assemble the proof in July. You cannot change what was true on June 30. Collect accordingly.
The six months between snapshots are not an intermission.
Capital arrives, leases sign, investors exit. Record events when they happen. December-you will be grateful.
Year-end closes three things at once.
The second asset snapshot, the reporting year itself, and the annual facts: housing units, employment months, the descriptive record.
The business sends its written statement up to the fund.
That is the §6039L moment from Chapter 5. The fund's return cannot be finished without it.
Then the fund adds what only it knows.
Its own asset values, investor contributions and exits, the certification record. The business never had these. The fund always did.
CPA review is its own stage, not a final spell-check.
Tracing values back to sources takes real calendar time. Budget it like a stage, because it is one.
Start from the date a reviewed package must be ready, then work backward.
Statement in hand by then, records collected by then, snapshots taken on the dates the law picked. The filing deadline is the finish line, not the plan.
Before anyone asks anyone for anything: who files, what existed, which dates, whose records.
Last year's Form 8996, two business names, a bank login, and a hunch about who has the records.
That's a real starting point. It is not a setup. Eight frames from now it will be.
Name the reporting taxpayer: the legal entity that will file.
Main Street Opportunity Fund I, LP. Not 'Main Street,' not 'the project.' The exact name that signs the return goes on top of everything.
Draw what the fund owned during the year, with dates on every arrow.
Not the structure you plan for next year. The one that actually existed, month by month. Chapter 2's two maps, now in service.
Put every date on a single page: test dates, statement handoff, review, filing.
A calendar showing only the return's due date is a countdown to a scramble. You've met that workflow. It retired in Chapter 9.
The person legally responsible is rarely the person holding the document.
The fund owes the return. Maria holds the ledger. Write both columns down now, and the year stops running on guesswork.
Some facts are stable. Others must be measured fresh.
Entity names and tract numbers usually carry. Asset values, employment, and ownership percentages never do. Sort them before day one, not in March.
Six questions before the first request goes out.
Who files, what existed when, which dates control, who owns each record, what's needed, what carries forward. Answer all six and Chapter 12 gets easy.
One reply-all email versus seven precise requests. Guess which one this chapter recommends.
Alex opens a new email and adds everyone he can think of.
Subject: 'tax stuff.' Attachment: none. Deadline: vibes. Delete the draft. Let's route instead.
Ask people for records, never for tax answers.
Maria knows the accounting system cold. Deciding what the tax law means by 'qualifying' is not her job, and asking her to do it produces confident, wrong answers.
Tom confirms who the business is and what it does.
Legal name, address, activities. He confirms facts from documents he already has. Nobody types from memory.
Maria supplies the ledgers and the schedules tied to them.
Fixed-asset schedule, lease file, bank statements. Her strongest records are the ones the accounting system already prints.
One property, two custodians.
Marcus holds the construction stage: contracts, pay applications, completion dates. Devon holds the operating stage: sites, units, what's actually renting. Ask each about their half.
Frank sends the payroll report. The formula is not his problem.
Monthly headcounts and hours, straight from the payroll system. Someone else runs the math on them later, and Frank gets to go home on time.
Alex supplies the fund-side records. Kim decides what they mean.
Contributions, investor events, the certification file: Alex. Professional judgment on all of it: Kim. She receives records. She should never have to be their source.
In a small fund, seven roles might be three people.
Fine. Routing doesn't move responsibility, it moves each question to whoever can answer it correctly. Even if two of the hats are on the same head.
Every fact, its owner, its source document, its deadline. The whole year on one table.
Seven roles touch the report. Six originate data. Kim originates nothing.
The CPA reviews everything and creates none of it. That's the design, not a loophole.
Tom: identity, industry codes, and the signature.
He confirms the business's legal name, address, and EIN (Employer Identification Number, the business's federal ID number), picks its NAICS codes, and signs the annual statement. Once at setup, reconfirmed yearly, signed at year-end. The only mandatory role on this list.
Maria: property values, owned and leased, on both test dates.
Source: the last page of the fixed-asset ledger and the lease schedules. A phone photo is fine. This is evidence, not an art contest. Measured twice a year, collected mostly once.
Devon: sites, tracts, and residential unit counts. Marcus: improvement spending, as it happens.
Devon reports annually. Marcus's receipts get captured while the concrete is wet, because reconstructing them later is miserable.
Frank: twelve monthly headcounts, plus the hours of everyone who isn't full-time.
Generated monthly by the payroll system, collected once a year. The division math from Chapter 5 happens downstream of Frank.
Alex: contributions, investor events, certification records, and the fund's own values.
Per event as they happen, plus both test dates. No operating business has these facts. The fund always did.
Every row lands in one of three places.
The fund's annual return, a written statement, or the evidence file behind both. The rest of this book takes each row and turns it into a chapter. First row, next part: the fund proving itself.
No application. No approval letter. Which is convenient, right up until it isn't.
A partnership agreement, an EIN, and 'Opportunity Fund' in the name.
Which of these makes it a QOF? None of them. The name especially is doing zero legal work.
Start with an eligible entity: a corporation or partnership for federal tax purposes.
An LLC works if it's taxed as one of those. A disregarded LLC doesn't; remember Chapter 2, the label is not the answer.
The purpose goes in the governing documents.
The partnership agreement or corporate charter must say it: organized to invest in Qualified Opportunity Zone property. Written down, not implied.
The fund certifies itself, on Form 8996, on its own tax return.
Self-certification means exactly that. No IRS waiting room, no certificate suitable for framing. The form is the certificate.
The form asks for the first month. Choose it deliberately.
That month starts the 90 percent clock, sets the first-year test dates, and shapes the penalty math if things go wrong. It's one box. It controls several calendars.
Checking the box states a position. It doesn't make a single asset qualify. The records have to support what the form claims, every year, forever.
Build the proof while everyone still remembers.
Formation documents, the purpose language, the filed form, the first-month support. One folder, assembled now. Future Alex sends his thanks.
The cadence spine of this entire book. Everything else keeps time to these two days.
June 30 arrived long before Alex expected to think about a tax return.
That's the point. The law measures mid-year on purpose, while there's still time to fix things.
Last day of month six. Last day of the year.
For a calendar-year fund: June 30 and December 31. A test date is an as-of freeze: whatever was true that day is what gets measured.
QOZ property ÷ total assets
The fraction, computed separately on each date. Qualifying Opportunity Zone property on top, everything the fund holds on the bottom.
The average of the two dates must clear it. One soft date doesn't automatically sink the year; the average decides. Which is mercy, with math.
The second date gets its own fresh calculation.
Capital lands in October, a lease signs in November. Carrying June's numbers into December because 'nothing changed' is how funds discover something changed.
The new reporting law reads its values on these exact two dates.
The §6039K return wants total assets and qualifying property, as of each date. One calendar now serves two duties. You're welcome.
Your dates come from your fund's own tax year.
Fiscal-year funds shift both. First-year funds count from the first QOF month, the box Chapter 14 told you to choose deliberately. Never borrow another fund's calendar.
Label every record three ways: entity, date, purpose.
Which company, which snapshot, which test. Do that all year and the fraction assembles itself. Skip it and Chapter 36 becomes autobiography.
The numerator, the denominator, and the two classic ways to get both wrong.
Alex opens the June file and finds three large numbers.
Two partnership stakes and one cash balance. Which go on top, which go on the bottom, and does the cash count against him? Nine frames, full answer.
The 90 percent test measures the fund. Not the project.
Count what the fund itself holds. What the business owns shows up later, through a different test. Keep the levels separate or nothing else works.
The numerator holds exactly three things.
The same three from Chapter 2: shares in a zone business, a partnership stake in one, or qualifying property the fund owns directly. Nothing else gets on top.
The denominator is everything: all assets the fund holds or leases.
Valued under the method the fund selected. Which method? Chapter 17. Yes, that's a choice, and yes, it matters.
Do not count the building twice.
The fund owns the partnership interest. The business owns the building. Put the interest in the fraction and stop there; adding the building too is project thinking, and the test just told you whose test this is.
October 1: a $1.1 million contribution lands, and sits in cash.
Relief exists, with three conditions: the money came in exchange for an interest in the fund, landed within the six months before the test, and sat continuously in cash, cash equivalents, or debt of 18 months or less from the fifth business day after arrival. Meet all three and it sits out the calculation entirely.
Proceeds from selling qualifying property follow a different clock.
Generally twelve months to reinvest, and only while the proceeds sit continuously in cash, cash equivalents, or short-term debt. New money and recycled money are different rules. Label which one you're holding.
Build the fraction in the same order every time.
Numerator items, then the full asset list, then the relief items, then the math. A fraction assembled in a repeatable order is a fraction someone can review. That someone is Kim.
Three columns on a spreadsheet. Only one of them is your friend.
Maria sends the December fixed-asset schedule. It has three columns.
Cost, accumulated depreciation, net book value. Kim needs one number per asset. The spreadsheet offers three opinions.
The rules allow exactly two ways to value assets.
One method per entity, per year, applied to everything. This is a menu with two items and no substitutions.
The financial statement method requires an applicable financial statement.
Meaning audited financials or a few other formally recognized statements. An internal report your bookkeeping software prints does not clear the gate, however official the font. And clearing it is permission, not a mandate: an AFS fund may still choose method two.
The alternative method: unadjusted cost for what you bought or built.
What the entity actually paid, ignoring depreciation entirely, for property purchased or constructed at fair value. Property that arrived any other way is measured at fair market value on the test dates, and leases get Chapter 23's math. For most small funds without audited financials, this is the method.
Net book value is cost minus depreciation. Every year it quietly shrinks.
The building did not get smaller. The spreadsheet says otherwise. Grab that middle column and your reported values sink a little more each year, for no legal reason at all.
The method is an entity-year decision, not a column-by-column preference.
Pick it, write it down, apply it to every asset on both test dates. It governs this fraction and the business's 70 percent test coming in Chapter 20. Leases get their own valuation math in Chapter 23.
One short memorandum: which method, why, signed and dated.
Every value review starts by reading it. Ten seconds of writing now buys back hours of December archaeology.
The original penalty, older than OZ 2.0. Month-by-month math, and a real relief valve.
Kim reaches the last line of the annual calculation, and the average is under 90.
Deep breath. A failed average does not mean a flat fine. It means a calculation begins.
The annual average opens the penalty. The months decide its size.
One weak date can be rescued by a strong one. Only when the annual result on the form comes up short does Part IV open, and the fund walks month by month through the shortfall.
(required − actual) × rate ÷ 12
For each short month: what 90 percent of that month's total assets demanded, minus the qualifying property actually held, times the annual underpayment rate, divided by 12. The rate is annual; the form applies one-twelfth of it per month.
A first-year fund counts fewer months.
Months before the fund's first QOF month stay out of the schedule. That box from Chapter 14 just earned its keep.
A business-level failure may have its own cure path.
When the operating business slips rather than the fund, the rules may allow a cure window before anything flows upstream. Chapter 20 closes with those rules; check them before running the fund's math.
The fund calculates the amount and reports it on Form 8996 itself.
The fund does the math and puts the number on its own form; a partnership fund's amount then flows through to its partners proportionately. The IRS follows with a notice carrying the payment procedure and the reasonable-cause process.
Reasonable cause is real relief, and it runs on records.
This penalty carries its own statutory waiver: none is imposed where the failure is due to reasonable cause. A different provision than Chapter 7's reporting relief, argued the same way, with documents created during the year.
This penalty punishes a failed asset test. Chapter 7's punishes a failed report.
Different statutes, different math, different fixes. When a notice arrives, the first question is always: which duty failed? Then, and only then, the calculator.
The first block of the new return. Four numbers, one flag, and a filing trap hiding in your own workpaper.
Four dollar figures sit at the top of the new reporting checklist.
All four come from the fund's own records. Nobody to email, nobody to chase. This block is entirely Alex's problem.
The §6039K return belongs to the Qualified Opportunity Fund, by name.
The business feeds it and the CPA reviews it, but the legal duty sits on the fund. The taxpayer from Chapter 11, line one.
2 values × 2 dates
Total assets and qualifying property, each measured on both test dates. Same June 30 and December 31 from Chapter 15, now with a second job.
The complete snapshot and the adjusted test math are not the same number.
October's $1.1 million sat out the December test under the six-month rule. It still exists, and the report wants the full picture. Keep two workpapers: the whole snapshot, and the test with its relief applied.
The qualifying total is a fund-level rollup.
Every stake in a zone business plus any property the fund holds directly, valued under Chapter 17's chosen method, added once. No borrowing numbers from the business's books.
Funds that run a business directly report their own employees too.
An employment indicator, for the uncommon fund with no operating company in between. If that's you, Chapter 28's math applies to you personally. Most funds skip this frame.
A reviewer should walk from the return to the books without guessing.
Four values, each tied to a dated record and the method memo. That is the whole fund-level file. Next part: the business's turn, where the numbers get company.
Part VI begins. The fund proved itself; now the business answers six questions, every year, in order.
Four percentages sit on the Market Hall file: 70, 50, 40, 5.
Each one is a separate test with its own records. Memorize the four numbers now; the next six frames give each its question.
Is there a real trade or business here?
Before any percentage: actual operations, actual activity. A shell holding an empty lot with no plan fails here. A development-stage business running under Chapter 24's written plan is different: the startup protections can carry the building years.
At least 70 percent of the business's tangible property, by value, must qualify. What makes property qualify is Chapters 21 and 22. What 'value' means is the method you met in Chapter 17.
Half the gross income must come from active conduct in a zone.
The 50 percent test, with four routes: three safe harbors (where the service hours happen, where the amounts paid for services happen, or where the tangible property and the management or operational functions sit) plus a separate facts-and-circumstances case. One route suffices.
Intangibles: 40 percent used in the active zone business.
Assets with no physical form, like licenses and trademarks. Less famous than the property test, still on the exam.
This one is a ceiling: less than 5 percent of the business's average aggregate unadjusted property bases in nonqualified financial property. Idle stock and bond hoards count; reasonable working capital under a written plan does not, and Chapter 24 is about that pass.
Some businesses are simply off the list.
Golf courses, country clubs, casinos, liquor stores, and a few similar categories. The ban is near-absolute, with two slim regulatory allowances: under 5 percent of gross income from such an activity, and under 5 percent of property leased to one, measured by square feet for real property and value for the rest.
A business-caused failure can get one cure window.
A six-month cure period lets the fund keep counting its interest while the defect gets fixed, and each business gets that window exactly once. If the cure fails, the fund's penalty math runs over every short month, the window included.
Property that once qualified can keep counting, for a while.
By statute, tangible property that ceases to qualify stays in the column for the lesser of five years or the rest of its holding, provided it had a real working life first. A cushion for aging assets, not a cure for a failed business.
The new reporting statement is not a qualification certificate.
Answering §6039L's questions proves the business reported. Passing the six questions proves it qualified. Do not let the first feeling substitute for the second.
Six questions, six answers, six sets of records, in order, every year.
That folder is the business's annual qualification file. The next two chapters fill its biggest section: which property counts, and why.
Two doors into the qualifying column. Every row on the schedule walks through one or neither.
Maria hands Kim a property schedule with five rows.
Building, kitchen equipment, furniture, a lease, a truck. Before anyone adds the column, each row has to earn its place in it.
One property list, two paths in.
Owned property qualifies one way, leased property another. Same destination, different paperwork. Pick the right door per row.
A real purchase, after the zone's start date, from an unrelated party.
Related means related: family, and entities with meaningful common ownership. The zone rules draw that line at 20 percent. Buying the building from yourself was never going to work.
Property the business constructs itself can walk the purchase path too.
The materials were purchased, the building went up in the zone. Construction is not a disqualifier; it's just a purchase in slow motion.
Leased property can qualify without anyone ever buying it.
The lease must start after the zone's start date and sit on market-rate terms. What the market would actually charge, provable, not vibes.
A related-party lease is allowed. It just travels with a chaperone.
Market-rate terms and extra conditions, including limits on prepaying rent. The file needs more care precisely because nobody negotiated against you.
70% use · 90% hold
Both doors share two rules: the property's use sits in the zone at least 70 percent, for at least 90 percent of the time you hold or lease it. A qualifying acquisition still needs a real zone life afterward.
After 2026, 'the date' depends on which map your zone is from.
Original 2018 zones and new-cycle zones carry different applicable start dates, and after 2026 an old-zone acquisition needs one of Chapter 6's three narrow paths. The closing date starts the analysis; it no longer ends it.
The total comes last. First decide which rows belong in it.
One page per row: which door, which date, which documents. A schedule that only shows the sum is an answer with no work attached.
Purchased property faces one more question: is it new to the zone, or did you make it new?
Kim has the purchase agreement. Maria has the asset schedule. Marcus has two years of pay applications.
Together, those three files answer the question this chapter asks. Separately, they answer nothing.
The two paths measure different facts.
Original use reads the property's history. Substantial improvement reads your spending. History or money. Every purchased property needs one of them.
Original use means first use inside the zone starts with you.
The equipment can be secondhand from another state. What matters is that its working life inside this zone begins under your ownership. New to the zone, not new to the world.
A vacant building can restart its own clock.
Continuously vacant for three years, generally, or one year if it was already vacant when the zone was designated, and its next productive use counts as original. The building's age is irrelevant; the vacancy record is everything.
basis × 2 in 30 months
Path two. Pick a 30-month window after purchase. Your additions to basis (basis: the property's tax cost) must exceed what the basis was when the window opened. In plain terms: more than double the building.
Land is excluded from the doubling math. A $1 million purchase that is mostly dirt needs to double a much smaller building number. Run the split before declaring the target impossible, or worse, declaring it met.
Costs pool only where the rules say so.
The equipment budget cannot rescue the building's math just because both live in the same fund. But defined groupings exist: buildings on one deed, or adjoining parcels run as one operation with shared facilities, can be improved as a single property. Group by the rule, never by convenience.
In rural zones, the doubling becomes a halving.
The 100 percent threshold drops to 50. Chapter 25 has the details and the fine print about what counts as rural.
Build the proof while the concrete is wet.
Pay applications, invoices, capitalized cost schedules, dated as they happen. A year-end total proves you spent money. It does not prove when, or on what.
You don't own it. It still needs a dollar figure on two test dates. Here's the math.
Market Hall's premises lease: 60 monthly payments, signed at inception.
No purchase price exists, because nothing was purchased. The value has to be built from the payment schedule itself.
Chapter 21 asked whether the lease qualifies. This chapter assumes yes and asks: at what value?
Classification before valuation, always. Pricing a lease that doesn't belong in the column is very precise wasted work.
Start with the payments as they stood on signing day.
The original schedule, at inception. Routine operations don't rewrite the starting math; a material amendment is a new analysis, reviewed before anyone reuses the old number.
The discount rate comes from the month the lease began.
Specifically the short-term applicable federal rate (the AFR), compounded semiannually, from the month the lease began. Not your bank loan rate, not the rate printed in the lease. This is the alternative method's math; a fund on audited financials reads lease values off its statements instead.
each payment, discounted, summed
Present value, defined: a dollar due in year four is worth less than a dollar due next month. Discount every payment back to signing day, add the results, and that total is the lease's reportable value.
The value is calculated once and usually holds for the life of the lease.
No annual re-runs while the lease stays active. The number was born at inception and it retires with the lease.
Five leases can mean five different rates.
Each lease keeps the AFR from its own start month. One short workpaper per lease, showing schedule, rate, and math, makes the whole thing reproducible. Which is the entire point.
A big pile of cash, a written plan, and 31 months. The most important paperwork in the building phase.
The project account holds a large cash balance, and Chapter 20 set a 5 percent ceiling on idle money.
Construction takes years. The money has to sit somewhere while permits crawl. On paper, patience looks like a violation.
A safe harbor: follow the rule exactly, and you're protected.
This one is the working capital safe harbor, the WCSH, and it earns its acronym. Cash held under a qualifying written plan stops counting against the business.
Three requirements, and they only work together.
A written designation of what the money is for, a schedule for spending it, and actual spending that follows both. Two out of three protects nothing.
Months, at the most. The plan's own written schedule governs, and 31 is its ceiling: a 24-month schedule buys 24 months of protection, not 31. Development, construction, substantial improvement, opening a business: the plan names it, the clock times it.
The 31 months run from the day the business receives the assets.
Receipt. Not the day the plan was adopted, not the day the fund got the money. Each dollar's clock starts when it lands in the business's hands.
Multiple plans can run at once, or back to back.
A second raise gets a second plan with its own designation, schedule, and months. Overlapping or sequential plans can reach 62 months in total, but only when each qualifies on its own, the earlier money is spent as written, and the later infusion is substantial and integral to the original plan. No drafting off the first one.
Some delays don't count against you.
A completed government application that stalls can toll the clock for the delay it causes, and a federally declared disaster can add up to 24 more months, conditions attached. The file needs the application dates and the follow-up trail, because 'the county was slow' is a claim, not a record.
The safe harbor protects the plan, not the whole business.
It answers specific tests during the spend, chiefly the 5 percent ceiling. It does not certify QOZB status, and unspent cash is not itself zone business property. A dated file proves the plan was followed; it never certifies the harbor was won.
Notice 2026-40 built a narrow bridge for original zones.
Property acquired after December 31, 2026 in a 2018 zone can still qualify if the written plan was adopted by that date, the buys follow it substantially, the business had received at least 10 percent of the planned working capital by then, and had spent at least 5 percent. Amounts locked by a pre-2027 binding agreement count toward the 5 percent, and only the 5 percent.
31 spends the money. 30 doubles the building.
The WCSH spend clock and Chapter 22's improvement window are different clocks that often run in the same building. A bridge gets you across the timing gap. Nobody gets to live on it.
Part VII begins. One word on the map, one halved threshold, and a surprising amount of paperwork.
Rural Mill: an old factory outside a small town. Looks rural.
'Looks rural' is scenery. The tax code reads a map, not a landscape. The tract decides, and the tract has a definition.
Rural means the zone is comprised entirely of a rural area.
Entirely. A zone that is 95 percent countryside and 5 percent city edge does not make the list. The map is binary about this.
The definition works by exclusion: not a city or town with more than 50,000 people, and not the urbanized area glued onto one. Everything else in the zone counts as rural.
The payoff. Substantial improvement drops from exceeding 100 percent of basis to exceeding 50 percent. Chapter 22's doubling becomes a halving, and on an old factory that is often the difference between feasible and forget it.
Everything else stays exactly as strict.
The purchase rules, the lease rules, the 70 percent use and 90 percent holding tests, the zone-location requirement. Rural is a discount on one test, not a hall pass, and it cannot rescue property that failed Chapter 21's acquisition rules.
One rural building does not make a rural fund.
The property-level result you're building here is separate from Qualified Rural Opportunity Fund status, the QROF, which has its own fund-level test. That's Chapter 26, next.
Chapter 5's 3,309 number lives in Notice 2025-50.
That notice lists which 2018 zones count as entirely rural. New map generations from Chapter 6 will need their own lists. Cite the list that covers your tract's generation, not the one that's convenient.
The rural claim cannot sit in a footnote.
Tract number, the list it appears on, the date verified, saved to the property file. It changes a 100 into a 50. Facts that halve thresholds get receipts.
The property proved it's rural. Now the fund wants the title. Different question, second calculation.
Tract support, building basis, the improvement math. Rural Mill's property file is done.
Chapter 25 earned the property-level result. QROF status is a fund-level result, and it does not come free with the first one.
A rural fund starts life as an ordinary Qualified Opportunity Fund.
The law did not invent a separate vehicle. A QROF is a QOF that also passes a rural version of the asset test. Same animal, extra stripe.
Classify every fund asset twice: once as qualifying, once as rural.
The ordinary 90 percent column from Chapter 16, plus a rural-only column beside it. Two answers per asset, per date.
Rural assets arrive directly or through a business.
Property the fund holds itself in a rural zone, or stock and partnership stakes in rural zone businesses. The same two doors from Chapter 21, wearing boots.
The rural column is measured on the same two dates, under rules similar to the 90 percent standard.
Month six and year-end, Chapter 15's calendar. No new dates to memorize. The calendar stays; the column doubles.
Rural Growth holds rural and non-rural assets, and still passes.
A calendar-year fund launching January 2027 can carry both, as long as each column's annual result clears its own 90. Mixed is a portfolio description, not a verdict.
The 30 percent basis step-up rides on the fund's status.
Dana invests $1 million in 2027. Her enhanced five-year increase, 30 percent of her deferred gain, depends on the fund holding QROF status across the period the rules require. The asset calculation supports that result. Intent alone does not, and marketing copy is not a calculation.
The rural fund file stands on its own.
Both columns, both dates, every classification, every source. When Dana's preparer asks in 2032, the answer is a folder, not a memory. Chapter 49 runs the whole case.
Part VIII begins. For every qualifying investment the fund holds in a business, the return wants a block. Here's what fills it.
Main Street Apartments, December 31: three dollar figures that refuse to match.
The fund's investment, the business's owned property, the business's leased property. They are not supposed to match. That's the whole lesson.
Every stock or partnership investment gets its own reporting block.
Two investments, two blocks. Ten investments, ten blocks. Property the fund owns directly reports under its own separate section, not a block. The §6039K return does not do group photos.
3 values × 2 dates
Per block: the fund's investment value, the business's owned tangible property, and its leased tangible property, each on both test dates. Six numbers, every one with its own source.
What the fund put in and what the business owns are different facts.
A $2 million stake can sit beside $8 million of buildings, or $500,000 of equipment. Neither number explains the other, and neither should be derived from it.
Apartments, commercial space, equipment, leases: one business, four kinds of property, one block.
The block summarizes the business's whole tangible world on two dates. The row-by-row proof behind it lives in Chapters 21 through 23.
The investment holds flat all year while the property values move.
Perfectly normal. The numbers come from different books and move for different reasons. A reviewer who expects them to travel together will 'fix' something that isn't broken.
The return is the fund's. Half the sources are not.
The fund knows its investment. The business knows its property. Filling a business block from fund-side guesses is how blocks go wrong. The pipeline for getting the real numbers is Chapter 29.
Reconcile every block in two directions.
Investment values back to the fund's own books. Property values back to the business's statement. Two directions, or you've checked half of a block and called it done.
Four fields with no dollar signs. They're the reason Congress renewed the whole program.
Alex figured the new record was just a longer asset schedule.
It's a census. These four fields describe what the money is actually doing, and they exist to be counted.
These fields feed the public reports.
Chapter 4's blind spot, remember? Industry, location, housing, and jobs roll up into the aggregate data Treasury must publish. Your December spreadsheet has an audience now.
Every NAICS code the business actually earns under.
What the business does, not what the pitch deck says. A food hall that also leases commercial space carries codes for both. Tom picks them from a plain-English list, and 'every' means every.
Location follows the property, tract by tract.
Not the mailing address, not headquarters. The tracts where the business's qualifying zone property sits, per 11-digit census tract. Property that doesn't qualify doesn't drive the list, and a business with three qualifying sites reports three answers.
Real property comes with an approximate residential unit count.
Approximate is the statute's own word, and it still needs a source. Devon's rent roll is a source; a shrug is not. The reporting cadence awaits the form, so track changes as they happen.
Employment is a monthly calculation, not a December headcount.
The unit is the full-time equivalent, the FTE: a way of counting the hours of everyone who isn't full-time as fractions of full-time jobs.
FT + (non-FT hours ÷ 120)
Each month: full-time employees, plus the hours of everyone who is not full-time, divided by 120, borrowing the method from the health-coverage rules in §4980H. Twelve monthly answers, averaged, reported in ranges or another indicator the Secretary selects. Frank's report already holds all twelve inputs.
Tie every descriptive answer to a person and a document.
Tom owns industry. Devon owns location and housing. Frank owns the payroll inputs. Numbers with no dollar sign still get receipts.
§6039L, the chapter. The business's first federal reporting duty, and the fund return's critical path.
Maria sent the values. Devon confirmed the sites and units. Frank sent payroll.
Chapters 12 and 13, executed. Everything the statement needs already exists. This chapter is about the envelope.
This handoff used to be an email thread with no rules.
No form, no deadline, no duty. Chapter 3 told you to hold that thought. §6039L is the thought, now with legal weight.
The duty lands on every applicable Qualified Opportunity Zone Business.
Applicable means connected, three ways: a business whose stock a fund counts, one whose partnership interest a fund counts, or a trade or business the fund runs itself. The legal relationship creates the duty, not the org chart's mood.
The statement doesn't have its own little list. It points at the full business record.
The statute aims it at one target: whatever the fund needs to complete its §6039K(b)(5) blocks. Identity, NAICS codes, tracts, owned and leased values, units, employment: Chapters 27 and 28 assemble that record, and Treasury will prescribe the exact contents.
One number travels down before the statement travels up.
The fund's investment amount lives on the fund's books, not the business's. The fund supplies it. Everything else flows upward.
The statement reports facts. It does not certify qualification.
Tom's signature says these are the business's numbers. Whether the business passes Chapter 20's six questions is a separate conclusion, reached separately. Chapter 20's warning label, still attached.
Treasury has not yet prescribed the deadline, format, delivery method, or exact contents. The target is statutory and knowable today: the fund's §6039K(b)(5) record. Build that record now, watch the mechanics.
Assemble the record and the statement writes itself.
A business that ran Chapters 27 and 28 all year is a short step from any format Treasury names. A business that didn't is one very long February away. Choose in June.
The statement's life cycle: checked, handed over, provably handed over, and fixed without a shredder.
Review starts before anyone signs.
A statement with missing sources doesn't go to the signer with fingers crossed. It goes back. Tom signs last, after the checking, not instead of it.
A statement is not stronger because it looks formal.
Letterhead is not evidence. The reviewer walks each number to its document: values to the ledger, units to the rent roll, months to the payroll report. Formatting flatters; sources prove.
Furnish is the legal handoff.
Furnish means actually delivered to the fund, not saved in a folder named FINAL. The duty completes on delivery, and the statute chose that word on purpose.
A delivered statement needs a delivery record.
Date sent, method used, copy retained, receipt if you can get one. When anyone asks 'did the fund get it,' the answer should be a timestamp, not a recollection.
Corrections supersede. They never erase.
Maria finds a lease error in March. Version 2 goes out labeled as a correction, version 1 stays in the file with a note. Supersede means the new one governs and the old one remains visible. Silent replacement is how audits get interesting.
The whole review fits on a single sheet.
Each fact, its source, its checker, its date. Sign-off at the bottom, delivery record attached. If the review can't be shown on one page, it probably wasn't finished.
Final rules may move the mechanics. They won't move the habits.
Sourced numbers, provable delivery, visible corrections. Whatever format Treasury eventually prescribes, it will reward exactly this file. Next part: the investors walk in.
Part IX begins. Jordan's money arrives, the fund's investor ledger opens, and two clocks start ticking.
Jordan sells stock January 15, 2027, and wires $750,000 to the fund on April 1.
Chapter 1's opening scene, now with paperwork. Two questions: did he make it in time, and what must the fund write down?
Two clocks start on different dates.
Clock one: did the money arrive inside the deferral window? Clock two: how long has the investment been held? The first is measured in days, the second in years. Chapter 35 owns the years.
Start with a gain that's actually eligible.
The deferral runs on capital gains. Savings, a bonus, a loan: real money, but no gain, so no election. Having cash available is not the same as having a gain to defer.
The 180 days generally start when the gain would be recognized.
For Jordan's stock sale, that's the sale date, and the sale date counts as day one. January 15 to April 1 is day 77. Comfortably inside the window.
Pass-through investors get three separate possible clocks.
The entity's own 180 days; 180 days from the last day of the entity's tax year; or 180 days from the entity's unextended return due date. Each period stands alone: land inside one and record which. A gap between them is a gap, and the arrival date of the K-1, the annual tax slip a partnership sends each partner, changes nothing.
A timely wire is not enough. An equity interest must issue.
Jordan must actually receive ownership in the fund: units, shares, a percentage. Money sitting in escrow with no interest issued hasn't invested in anything yet.
Each investment gets its own lot record, forever.
A lot is one investment tracked separately. Invest twice, get two lots with two dates and two sets of clocks. A single lifetime balance is where basis math goes to die, and Chapter 32 explains why.
Form 8997 is an every-year form, and skipping a year bites.
File it for each year the investment is held at any point. Miss a year and the regulations presume an inclusion event happened, a presumption you can rebut, with evidence. Correct the gap and keep the proof of what actually occurred.
The fund proves the arrival. The investor proves the gain.
The fund records $750,000, April 1, interest issued. Jordan keeps the sale confirmation and makes the election: code Z on Form 8949 under today's instructions, the balance on Form 8997. Two files, one story, permanently linked.
One wire, five numbers, and the counterintuitive zero at the center of the whole system.
Jordan invests $750,000. His starting tax basis: zero.
Basis, again: the tax cost the system measures gain against. For a simple qualifying cash investment, the statute starts it at zero on purpose. It looks like a typo. It's the design.
Five numbers can describe one investment, and they don't have to match.
Amount invested, deferred gain, tax basis, fair market value, and the interest on the fund's books. Each answers a different question. Forcing them to agree breaks at least three of them.
The zero basis is the IOU.
Jordan hasn't paid tax on the $750,000 yet, so the system gives him no credit for it yet. Basis grows as the deal's promises come due. Zero is the receipt for the deferral.
A partnership stake's basis never sits still.
Zero is only the opening line. Fund-level debt allocations, income and loss, distributions, and noncash contributions all move a partner's number from day one. The clean-zero story is the cash-for-stock version; partnership math belongs to the preparer.
At year five, basis rises by 10 percent of the deferred gain. 30 for rural.
Before the deferred gain comes due, the step-up lands: Jordan's basis climbs $75,000, and Dana's rural million climbs $300,000. Less gain gets taxed. That's the discount, delivered through basis.
Invest more than the gain, and you own two investments in one position.
Jordan defers $750,000 but wires $900,000. The extra $150,000 is a normal investment: normal basis, no deferral, none of the ten-year magic. One ownership stake, two tax lives, tracked separately.
Mixed funds also happen by accident.
A wire that misses the 180 days, a gain that wasn't eligible: the non-qualifying slice still owns part of the position. Nobody plans a mixed fund. The ledger has to handle one anyway.
Whatever else moves, the deferred gain is the anchor.
The interest's value rises and falls, distributions happen, units get renamed. The $750,000 under election is the number the year-five calculation will ask for. Guard it.
Bob's 2019 lot and Jordan's 2027 lot follow different laws.
Legacy lots ride the old timeline. Post-2026 lots ride Chapter 35's clocks. Same fund, two regimes, so every lot wears a regime tag, or someone eventually runs 2019 math on a 2027 investment.
The phone rings in November. What the fund must capture, report, and hand to the person walking out.
November 15: Bob tells Alex he sold 40 percent of his stake.
One sentence on the phone. Several new entries in the fund's records. This chapter is the gap between those two facts.
The fund reports the person and the event.
The current form already asks whether anyone disposed of an interest. The new return wants the full record: name, address, taxpayer ID, acquisition dates, disposition dates, and the amount disposed. First names and vibes no longer file.
The record follows the investment lot, not the investor.
Bob invested once in 2019, so it's simple. An investor with three lots who sells 'some' needs the record to say which lot, which portion, which dates. Chapter 31's lot discipline, cashing its first check.
The amount disposed and the cash received are different numbers.
Bob sold 40 percent of the lot; the buyer paid $260,000; the slice carried $200,000 of deferred gain. All three are true, all three get recorded, and which one the form will call 'the amount disposed' awaits the form. Record everything, label later.
Maya moves 25 percent into a family trust. Zero dollars change hands.
Still an event record: transferor, transferee, date, portion, and the trust document. Whether it counts as a reportable disposition, and whether tax follows, are later professional questions. The fact gets captured either way.
The fund must also furnish the departing investor a written statement.
The same facts the return reports, handed to the person who needs them for their own filing, in the time and manner Treasury prescribes. Chapter 30's habits apply either way: delivered, provably, copy kept.
Capture the event in November, not in March.
The fund's job is the facts: who, which lot, when, how much, what proof. The tax verdict on Bob's sale belongs to Bob's preparer. Chapter 34 walks the messier versions.
Sales, gifts, trusts, redemptions, death. The event happens first. The tax answer comes later, from the right person.
The event changes the record before anyone knows the tax answer.
Write down what happened: who, which lot, what date, what portion, what was received, what document proves it. The verdict can wait. The facts can't.
A 40 percent sale can split every number on the lot.
In the simple case, identical shares in a QOF corporation, the $500,000 of deferred gain splits $200,000 to the sold slice and $300,000 staying behind, step-ups riding along proportionally. Partnership and S corporation interests run a special hypothetical-sale computation instead. A blended balance would have hidden all of it.
The proportional table starts the analysis. It doesn't finish the return.
The split hands Bob's preparer clean inputs: proceeds, gain, basis, holding period. The preparer combines them into an answer. The fund's job was making that possible.
One sale can appear in four places without duplicating.
The fund's return and its attachment, a Form 1099-B where required, Bob's Form 8997 Part III, and code Y on his 8949. Different audiences, same event. The overlap is the system working.
The fund's statement to Bob reports facts, not conclusions.
Which lot, which date, which amount, which document. It should never say a transfer 'qualifies for an exception' unless a professional determined that. Chapter 20's warning label, investor edition.
Bob's 2019 lot and Jordan's 2027 lot raise different timing questions.
A legacy sale interacts with the old cliff. A post-2026 sale asks one thing first: before or after the five-year date? The regime tag from Chapter 32 is the first field the reviewer reads.
Some events should never ride the automatic path.
Gift, death, trust transfer, redemption, partnership distribution, multiple lots, mixed funds. Each carries its own rules and exceptions. Routing them to a professional is the process working, not failing.
The closeout file keeps the whole trail.
Subscription record, event document, proceeds, basis history, statement copy, delivery proof, and the open questions. Two years from now, 'what happened' should be answered by a folder. It usually isn't. Yours will be.
Part IX closes with the long game: three dates in Jordan's file, one of them in 2057.
2032 · 2037 · 2057
Jordan invested April 1, 2027. Five years, ten years, thirty years. Three appointments, one investment, and each does something different.
The deferred tax comes due at the earliest of a sale, another inclusion event, or the five-year date.
An inclusion event is any trigger the rules list as ending deferral early; a sale is the everyday one. For Jordan that's April 1, 2032, unless he exits sooner, and the step-up from Chapter 32 lands first, so the bill arrives pre-discounted.
A partial exit stops the clock for only that slice.
Jordan sells 25 percent in 2030: that quarter's deferred gain comes due then. The other 75 percent keeps its 2032 appointment. Proportional, again. The lot record splits, again.
Ten years opens an election. It doesn't grant one automatically.
Hold to April 1, 2037, and Jordan can elect to step his basis to fair market value on sale, so the investment's own growth goes untaxed. Can elect. Someone has to actually make the election, on an actual return.
Thirty years caps the value the election can use.
The new limit legacy investors never had: the fair market value at year thirty, 2057 for Jordan, is the most the step-up can reach. Growth after that rides outside the shelter. Hold forever if you like; the meter stops.
The zone's calendar and the investor's calendar are different clocks.
The tract's designation can lapse in 2028 while Jordan's clocks run to 2057. Chapter 6's map governs the property. These three dates govern the person. Neither pauses the other.
Legacy lots follow the old calendar entirely.
Bob's 2019 investment hit the December 31, 2026 cliff, under the old step-up schedule and the old ten-year rule, with no thirty-year cap. Same fund, two timelines. The regime tag earns its keep one last time.
The old deal has an expiration and a one-way door.
The legacy ten-year election needs a qualifying disposition before January 1, 2048. And the gain forced in on December 31, 2026 cannot ride into a second deferral; that money's zone journey ends there.
These records must outlive everyone currently holding them.
Staff turn over, preparers retire, and the software will die twice before 2057. Each lot's file carries dates, amounts, elections, and events, built to survive custody changes, because it will have several.
Part IX complete: entry, ledger, exits, clocks.
The investor lane is built. Next part: the cycles that run the actual collection year, starting with the rescue chapter some readers already need.
Part X begins. The rescue chapter. If June 30 slipped by unnoticed, this is where the book told you to come.
Alex opens the accounting system on July 8 and asks it for June 30.
Eight days late to a photograph. The good news: the scene can be rebuilt. The rules for rebuilding it honestly start now.
Confirm the fund's own date before collecting anything.
June 30 belongs to calendar-year funds. A fiscal-year fund that reconstructs June 30 has done flawless work on a question nobody asked. Chapter 15's calendar rules the rescue too.
Rebuild inside the structure that existed on the date.
One snapshot for the fund, one per business, using the ownership map as it stood that day. A passed date is not one big pile. It's several small, dated ones.
Start with records that already speak to the date.
A June 30 bank statement, a dated payroll run, a signed lease. Documents born on or near the date outrank anything assembled afterward. Collect those before touching a spreadsheet.
When the best record is dated later, bridge backward. Never copy.
July 31 balance, minus July's activity, equals June 30. Show that arithmetic. Copying the later number and relabeling it June is fiction with a timestamp.
Main Street reconstructs June 30 without guessing once.
The June bank statement, the contribution log, the two partnership stakes at their supported values. Every figure lands on a document. The date comes back whole.
A reconstruction must look like a reconstruction.
Method, sources used, and the date it was rebuilt, written in the file. A rebuilt number dressed up as a pulled number is a small lie waiting for a big audience.
A passed date lowers nothing.
Same evidence bar, different workflow. Whatever stays uncertain goes to Kim as an open item, not into a confident round number. Next chapter: the routine that retires this one.
The vaccine to Chapter 36's cure. Two cycles a year, two different jobs, zero archaeology.
Alex is never reconstructing June 30 again.
Correct instinct. The fix is not heroic effort in July. It's a small, boring routine that runs twice a year on schedule.
Two cycles, two jobs. Never one giant sweep.
Midyear checks the fund's health while there's still time to fix it. Year-end closes the record for keeps. Same calendar, opposite purposes.
The measurement date should never arrive as a surprise.
Two weeks out, the record owners get a heads-up: the date is coming, here's what you'll owe. Chapter 12's routing map doubles as the invite list.
Capture the light snapshot as soon as the period closes.
Bank balances, positions, anything that describes the date, gathered while it's fresh. An hour in early July beats a week in February. Light means light.
Record events when they happen, not when the year ends.
A contribution lands, a lease signs, Bob calls. Into the log that day. Chapter 10 said the six months between snapshots aren't an intermission. This is the part where you believe it.
Year-end does two jobs at once.
The second snapshot on the date, plus the annual facts: unit counts, the twelve payroll months, the event log, closed. The descriptive record from Chapter 28 gets finished here, not invented in March.
The law fixes the measurement dates. You choose the collection dates.
Set internal targets and label them internal. Dressing a house deadline up as a legal one works exactly once, and then the real deadlines inherit the skepticism.
A cycle isn't done when everyone replies.
It's done when the records are checked, filed, and the open items are written down. This division keeps the work small and the evidence strong. Part XI takes it from here.
Part XI begins: where it all goes. The desk changes. This part belongs to Kim.
Kim opens the year-end package and selects one value: $1,080,000.
Market Hall's owned property. Reviewing a package means being able to do what's about to happen with any number in it. Start with one.
Start where the number is going, not where it came from.
Which return line, which block, which statement. A pile of documents is not a review. A value with a known destination is reviewable; everything else is filing-adjacent clutter.
From the reported value, step back one layer at a time.
Block, workpaper, adjustment, schedule, ledger, source document. Every hop written down. The layer you skip is statistically where the error lives.
One document can feed several outputs.
The fixed-asset ledger supports the 70 percent test, the business block, and the statement. One source, three destinations, all noted, so a future correction can find every place it landed.
Trace the other direction too: everywhere the number went.
Backward proves the value is real. Forward proves you know its blast radius. A changed number with unmapped destinations becomes an incomplete correction, on a schedule of its own choosing.
Every amount keeps its date, its method, and its version.
$1,080,000, as of December 31, alternative method, version 2. A naked number looks precise and answers nothing. Chapter 17's memo and Chapter 15's labels, cashing out.
Before accepting a value, Kim asks the same eight questions.
What is it, which entity, which date, which method, which source, who supplied it, what changed it, where does it go. Eight answers or it waits.
A number is reviewed when a second person can retrace it.
Where it came from, what changed it, where it went, all without interviewing anyone's memory. If the file can't do that, the file isn't finished. The review was the point.
A passing year is a stack of separate conclusions. Check the stack first. Then check the crutches.
There is no single combined pass.
The fund's 90 percent, the business's six questions, each property's path: separate tests, separate conclusions. The fund's answer leans on all of them, which is exactly why they get checked one at a time.
Run the ordinary calculation before touching any exception.
Relief is never the first line of a workpaper. First: did it pass on its own? Only then: which special rule applies, and to precisely what.
Each relief provision changes exactly one thing.
The six-month rule excuses one pile of new cash. The safe harbor protects one planned pile. Reasonable cause forgives one penalty. Nothing fixes everything, and stacking reliefs means naming each one.
Main Street passes, conditionally.
December cleared 90 only because the October $1.1 million sat out under the six-month rule. The pass is real. The condition gets written next to it, because next year the same cash won't be new.
Every Market Hall percentage passes. That's still five conclusions.
70, 50, 40, under 5, and not prohibited. 'It all looked fine' is a mood. Five short documented conclusions are a review.
A cure period is not a free second attempt.
A business-level slip may get Chapter 20's one-time cure window. Using it leaves a record: what failed, when it was cured, which rule allowed it. Cures come with receipts, not amnesia.
Sort the open items by what they actually block.
A missing source blocks the package. An open legal question routes to counsel. A pending government detail goes on the watch list. Three kinds of open, three different treatments, zero promotions between them.
One page holds the whole review: test, result, and what moved the answer.
Which fact or special rule carried each result across the line. A result that can't name its mover isn't a result yet. It's a hope with formatting.
Kim needs to change Maria's number. Both numbers survive. That is the entire trick.
Maria's schedule is right, and the reported value still needs to differ.
The books can be perfect while the tax answer diverges: different method, different classification. An adjustment is a translation between two correct languages, not an accusation.
First decide: source error, or professional judgment?
A source error means Maria corrects the record. A judgment call means Kim adjusts on top of it. Different problems, different owners, and confusing the two insults somebody either way.
Keep three layers, permanently visible.
What Maria sent, what Kim changed, what the package reports. Collapse them into one number and the file develops amnesia about its own history.
An adjustment needs a reason, evidence, a name, and a date.
A negative number sitting beside a total is not a workpaper. Why, based on what, decided by whom, decided when. Four fields, no exceptions, including the small ones.
One adjustment can move several results.
Market Hall's lease fix changes the reported leased-property value and the 70 percent denominator. Chapter 38's forward tracing earns its keep right here: find every landing site before calling anything done.
An adjustment cannot cure a weak source.
If the underlying record can't be trusted, adjusting its number just decorates the problem. Weak sources go back for better records. They do not go forward wearing makeup.
The adjustment lands when it's approved, not when it's spotted.
Kim proposes; the package changes after sign-off. And if a better source record shows up later, it can supersede the adjustment entirely, Chapter 30 rules: new version governs, old version stays visible.
Nothing was erased.
Supplied, adjusted, reported, reasons, approvals: all still readable, in order. That's the sentence the file has to be able to say. Next chapter bolts all of it into the package that goes out the door.
Everything reviewed becomes one assembly with several destinations. Built once, in order, with a lid.
The review is done. Now it becomes a package.
Not a folder of finished pieces. One assembly where every piece knows its destination, its evidence, and its status. The difference is the next eight frames.
The package feeds the return, the statements, and the file.
The fund's annual return, one block per qualifying investment, the investor statements, and the evidence archive behind all of it. Different recipients, one source of truth.
Start with a control sheet: what's inside, what's missing, who signed off.
One page listing every output, its status, and its open items. Yes, a cover page. It's the only page that knows the whole package, which makes it the most important one.
Form 8996 is still the published starting point. Complete it properly.
It already collects the investment values, property values, tract locations, and the method. The revised §6039K vehicle isn't posted, and the draft 2026 partnership return already cites a Form 8996 line 24 that today's form doesn't have: an expansion signal, not a spec. File on what's published.
Map the §6039K items without inventing line numbers.
NAICS, units, employment, dispositions: organized by statute section, labeled as awaiting the form. Chapter 8's rule stands. The statute is the spec, and guessing a layout helps nobody.
One reviewed block per business, one schedule of investor events.
Chapter 27's blocks, Chapter 33's dispositions, each carrying its sources. The fund packages investor facts. It never drafts the investor's return. Still true here.
The package climbs a ladder, and each rung means something.
Ready means reviewed, bound to evidence, open items named. Filed means it actually left. No rung gets skipped, and no rung gets claimed early.
A complete package is the bridge to filing. Not the filing.
It stops at ready. The handoff, the delivery, and the proof are the next chapter's job, and they deserve their own.
Every request says complete. Time to find out how many of those completes are true, send the outputs, and keep the proof.
Every request is marked complete. That is not the same as ready.
Complete means the record arrived. Ready means it was checked, reconciled, and bound to its destination. The gap between those two words is this chapter.
Review in layers, each person at a different height.
Maria checks her records against her ledger. Alex checks the map: every entity, every date covered. Kim checks the path from each value to its source. Three altitudes, one pass each, no duplicated laps.
Look for agreement where agreement should exist. Nowhere else.
The bank balance should tie to the ledger. The investment value should never tie to the property value; Chapter 27 spent a whole frame on that. Forcing unlike numbers to match is how correct files become wrong.
Four outputs, four directions, zero confusion.
The return travels with the fund's filing. The business statement goes up to the fund. Investor statements go out to people. The archive stays home. Each output, addressed.
A document in a folder named FINAL_v2_actually is not a delivery.
Chapter 30's rule, now applied to everything: date sent, method, copy kept, receipt where possible. Furnished means provable, for every output, not just the statement.
Preserve a file a stranger could reproduce.
One annual file for the fund, linked files per business and per affected investor. The test: someone who wasn't there rebuilds any reported number from the folder alone. If they'd need to call you, keep building.
Close the year with the open items written down, not wished away.
Pending guidance, unresolved questions, watch-list items, each with an owner. A year can close with open items. It cannot close with secret ones.
Call it ready when the relationships among the numbers are understood.
Not when the boxes are full. Full boxes are storage. Understood relationships are a filing. The year is delivered, proven, and preserved. Unless something needs fixing, which is next.
Part XI closes. Mistakes get fixed in the open, and the envelope from the government gets a workflow instead of a panic.
Kim has one rule before the year closes: name what changed.
A source record, a professional judgment, or a delivered output. Three different kinds of change, three different fixes. The name comes before the fix, every time.
Timing decides the route.
Caught in the draft: just fix it, note it, move on. Caught after furnishing: supersede. Caught after filing: a different road entirely. Where the error sits on the timeline is the whole decision.
A furnished statement gets a labeled version 2. Never a quiet swap.
Chapter 30's supersede rule at full scale: the correction is announced, the original stays visible, the recipients get the new version. History remains readable.
Correct the package before correcting the return.
The package is the bridge from Chapter 41. If a source moved, the bridge gets a new version first, then whatever crossed it. Fixing the return while the package still says the old number builds a contradiction with a paper trail.
An accepted return is not edited like a draft. Whether and how to amend is a professional decision with its own procedures. Nothing here moves fast, and nothing moves without Kim.
An IRS notice is its own workflow, not another open item.
Read what it actually claims. Verify the duty, the dates, and the math against your own records. Calendar the response deadline. Then answer the way the notice itself directs: a call, an upload, a written reply, an amended or superseding return, a partnership adjustment request, or several of those. The notice names its own road.
Reasonable cause is argued with the file you already built.
The records created during the year, the delivery proofs, the corrections done in the open. Whichever relief rule fits the failure on the table, it rewards exactly that trail. A well-kept file was always the response, written in advance.
A correction closes with a final status, in writing.
What changed, which outputs it touched, who approved it, and where it stands now. Part XI complete: reviewed, packaged, delivered, preserved, and correctable. Next part: the structures that don't fit the picture.
Part XII begins. Three structures the simple picture never covered, and the maps that keep them reportable.
Alex opens a new structure chart and sees three different problems.
A business owned by two funds. A fund holding property with no business in between. A stack of entities three layers deep. None of them breaks the rules. All of them bend the workflow.
Hard structures need three maps, not one.
The legal map: state-law entities and who owns what. The tax map: how federal law classifies each one. The reporting map: which entity owes which output. Chapter 2 drew two. Complexity earns the third.
Harbor Works has two fund owners. Each fund keeps its own record.
One business, two investors of the fund kind. Each fund reports its own investment, its own contribution history, its own block. The business is shared. The records never are.
Ownership percentage is not a reporting percentage.
Neither fund reports 'its share' of Harbor Works' buildings. The business reports its whole property picture; each fund reports its own stake. Slicing the building by ownership feels intuitive and is wrong.
A shared business furnishes a statement to each connected fund.
The identity, property, units, and employment facts are common. The §6039L duty runs separately to every fund counting the interest. Same facts, one statement per relationship the statute names, and, as practice, delivery proof on each.
Direct property skips the business and lands on the fund's own return.
The fund reports the property, the tracts, the units, and, per Chapter 19's rare flag, its own employees. And §6039L's definition reaches a trade or business the fund runs itself, so a statement duty exists here too; how a fund furnishes one to itself is a mechanic Treasury still owes us. All the business's homework, now the fund's.
In a stack of entities, the tax map decides which layers exist.
A disregarded entity (one the tax law ignores, treating its assets as its owner's) melts into the layer above for reporting. But it still holds title, signs leases, runs payroll. Invisible to the return, very visible to the records.
A separate lower-tier tax entity never collapses just because it's family.
Common ownership and one consolidated spreadsheet do not merge two taxpayers. But the return's blocks follow the fund's own investments: a company two layers down keeps its own records without earning its own block. Tax map, reporting map, block map: three answers, drawn separately.
Keep a structure register a stranger could redraw.
Every entity, every ownership arrow, its classification, its dates, its evidence. Chapter 42's archive test, applied to the org chart. If the structure lives in one person's head, it's already failing.
June 30 and December 31 were always shorthand. Here's the real rule, and what happens when the structure moves mid-year.
Alex circles June 30 and December 31 on the calendar. For his fund, correct. As a universal rule, no.
The dates come from each entity's own taxable year: end of month six, end of the year. Chapter 15 said it. This chapter is what it costs when structures get complicated.
A sponsor running three funds may be running three calendars.
One calendar-year fund, one September year-end, one first-year fund counting from its opening month. Copying dates between them is Chapter 36 material waiting to happen. Each fund, its own circled dates.
Silver Line closes its books in September. The values still follow the fund's dates.
The business's fiscal year runs its accounting. The fund's statutory dates run the measurement. Silver Line hands over values as of the fund's June 30 and December 31, even though neither is its own year-end. Inconvenient, and required.
Annual facts need a stated covered period, not a vague 'this year.'
Employment months, unit counts, income tests: each answer names its start and end dates. When fund and business years differ, 'the year' is ambiguous, and ambiguous facts don't reconcile. A fourth calendar stays blank for now: the statement's own period and deadline await Treasury.
An ownership change never rewrites an earlier snapshot.
A stake sold in September doesn't reach back and edit June. June's record shows June's owners at June's values, forever. The change creates a new record with its own date. History is append-only.
A classification change can redraw the tax map in one day.
A disregarded entity takes on a second owner and becomes a partnership overnight: new taxpayer, new return, new reporting path. The legal map barely moved. The tax map got a new country.
A new taxable year means a new calendar, with the old one preserved.
Short years get their own shifted dates, and the records that supported the old calendar stay filed under it. Nothing about a fresh start erases the old one's evidence.
Put time on every ownership arrow.
From when, to when, on what document. Chapter 44's structure register only becomes reliable once every edge carries dates. A chart without time is a portrait. The IRS wants the film.
Part XII closes. The last property sells, and it turns out endings have paperwork too.
Alex sells the last property and opens a folder called Final.
Optimistic name. A wind-down is a reporting phase, not an off switch, and the folder is about to have children.
Decide what is actually ending: the property, the status, or the entity.
A fund can sell everything and still exist, still have a taxable year, still owe a return. Three different endings, three different workflows. Pick the right funeral.
The wind-down map is the ownership chart plus exit dates.
Sale closed here, business relationship ended there, entity dissolved then. Each date changes which duties still apply. Chapter 45's rule, one last time: history is append-only, especially at the end.
Use every measurement date that falls before the relationship ends.
The business's final statement still gets built, reviewed, and furnished under Chapter 30's rules. The last handoff follows the same rules as the first one. No senioritis.
The final return is still a full return.
The entity's classification and short-year rules set the last calendar, and the last snapshots still get taken on it. A fund in wind-down is a fund with test dates until the day it legally isn't.
Final distributions can end deferrals.
A liquidation can be an inclusion event for the people holding lots. Chapter 34's stop signs apply at full strength: capture the facts, route the tax verdicts to professionals, promise nothing in between.
The investor file does not end with names and final checks.
Each lot gets its closing statement, its event record, and a note on how it ended. When the qualifying investment ends, its clocks end with it; the records proving what happened must outlast everyone who made them. Keep the file like it's 2057.
Name who keeps the records before the entity stops existing.
An archive without an owner is an abandoned box. Custodian named, retention matched to the facts, and three handoffs made: the final facts, the final forms and calculations, and the open questions, in writing. Then the folder can say Final and mean it.
Part XIII begins. Everything the book taught, run once, end to end, on the fund you already know.
One fund, one business, one calendar year. Done completely.
Alex began this book calling it one project. Watch what 'one project' actually took.
Taxpayer named, structure dated, calendar on one page.
Main Street Opportunity Fund I, LP at the top. Two maps drawn and dated. Test dates, handoff, review, and filing, all circled before anyone was asked for anything.
Seven precise asks replaced one reply-all email.
Tom got identity, Maria got values, Devon and Marcus split the property, Frank got payroll, Alex kept the fund side. The 'tax stuff' draft stayed deleted.
June 30 was rebuilt, and labeled as rebuilt.
The June bank statement, the contribution log, the supported stakes. Bridged where needed, method written down. The reconstruction looks like one, on purpose.
The full $8.1 million snapshot, with the test math beside it.
October's $1.1 million sat out the test under the six-month rule and stayed in the complete picture. Two workpapers, never blended. Chapter 19's trap, dodged in the wild.
Net book value stayed visible and stayed unused.
The alternative method governed every row, per the memo. Kim traced the values both directions, eight questions each. Nothing round and reassuring got waved through.
By February 1, 2027, the record was complete and Tom signed last.
House rules, all of them, run ahead of Treasury's mechanics: reviewed against sources, furnished with a delivery record. The short-step business, taking the short step.
A lease error superseded in the open. Bob's November sale captured in November.
Version 2 labeled, version 1 preserved. The disposition record, the investor statement, the event schedule: all in the package, which climbed to ready and stopped there.
Main Street began as one project in Alex's mind.
It ends as a reporting system: mapped, collected, rescued once, reviewed, corrected honestly, packaged, and proven. That's the whole book, wearing one address.
The business-side case: a working food hall, five kinds of records, six separate conclusions, one statement furnished with proof.
Maria sends Kim a fixed-asset schedule, a lease file, and a payroll report.
Chapter 29's trio, now with contents. Everything this chapter does traces back to those three documents and the people behind them.
A live food hall and commercial lessor, held 99 percent by the fund.
Real income, real staff, real leases. Chapter 20's six questions finally get a subject that eats, hires, and invoices.
The complete business record gets built first, then judged.
Identity confirmed, NAICS codes picked, tracts listed, values dated, units counted, twelve payroll months captured. Chapters 27 and 28, assembled before a single percentage is computed.
All tangible property on one side. Qualifying tangible property on the other.
Every row lands above or below the line for a stated reason: purchase door, lease door, original use, improvement. Chapters 21 and 22, applied row by row, no row waved in on vibes.
Sixty payments, the inception month's AFR, one present value.
Chapter 23's math on a real schedule, preserved in a one-page workpaper. The number was born at signing and will retire with the lease.
Income, intangibles, the 5 percent ceiling, the prohibited list: one at a time.
Four short, separate, documented conclusions. Chapter 39's discipline: 'it all looked fine' is a mood, and moods don't file.
Frank's twelve months go through the ÷120 math. Frank does not.
Full-time counts plus the hours of everyone who isn't full-time, converted downstream into the annual employment answer. Chapter 12's promise kept: Frank sent a report and went home on time.
Market Hall passes because several separate conclusions agree.
Maria's number and Kim's adjustment both preserved, statement signed and furnished with proof, fund block filled and reconciled both directions. A real business, fully told.
The rural case: one old factory, two columns, and a benefit that has to be earned twice a year.
The property is rural. The fund is a QROF. Dana gets 30 percent.
Three separate conclusions, and none of them implies the others. Chapters 25 and 26 drew those lines. This case walks them in order.
Rural Growth: calendar-year partnership, launched January 2027, deliberately mixed.
Rural and non-rural assets in one portfolio. Mixed is a description, not a problem, as long as each column's annual result clears its own bar.
Rural Mill's tract gets verified against the list, not the landscape.
The illustrative tract sits on Notice 2025-50's entirely-rural list, checked on a stated date, saved to the file. The scenery was never consulted.
An old-map factory, bought after 2026, needs a named path in.
Rural Mill's operating company adopted its written spend plan in 2026, took in a tenth of the planned capital, and spent its first 5 percent before New Year's. That is Chapter 24's transition bridge, condition by condition, and it is why a January 2027 fund can count this factory at all.
The old factory clears 50 percent where 100 was out of reach.
In the controlled math, the capitalized improvements land $150,000 above the halved line. Chapter 22's doubling became Chapter 25's halving, and the project became feasible.
Every fund asset gets classified twice.
Ordinary qualifying in one column, rural in the other, side by side per Chapter 26. Two answers per asset, sourced, on every row.
Both columns get computed on both dates, and both annual results pass.
Month six and year-end feed each column's annual answer: the ordinary result and the rural result, neither inferred from the other. The mixed fund earns its title on the calendar, not at the launch party.
Dana's $1 million lot carries a rural regime tag from day one.
Deferred gain, March 1, 2027, tagged per Chapter 32. Her enhanced five-year step-up depends on the fund's status holding across the period the rules require. The lot record and the fund file are now permanently linked.
The package states the benefit as conditional, because it is.
Rural layers a stranger can retrace, results tied to dates, the 30 percent framed on the fund's actual calculations. The file supports the claim. The file is the claim.
The teaching is done. Ninety days turns this book into a running system.
Alex closes the Rural Mill file and asks Kim: where do we actually start?
One goal, three months: every entity, date, record, owner, reviewer, and open item, known and written down. Here's the schedule.
Make the file visible.
Name the taxpayer, date the structure, set the calendar, assign every record an owner. Chapters 11 and 12, executed. Maps first, with one exception: any test date already passed and any 2026 transition checkpoint gets its evidence captured now, not in phase two.
Collect the dated proof for everything on the map.
The master request list goes out, the responses come back checked. Any passed test date gets rescued now, per Chapter 36, while the records are young. February is not invited to this phase.
Turn checked records into a package at ready.
Trace, test, adjust in the open, assemble, stop at the right rung. Part XI, run in miniature, on schedule instead of under deadline.
Sponsors own momentum.
Blockers visible, assigned, and moving. Not doing everyone's job. Making sure everyone's job has a name and a date on it. That's the whole role, and it's the one that decides whether the plan survives contact with July.
CPAs own review and judgment.
From reported value to source, calculation, and evidence, eight questions at a time. On the calendar as a stage, per Chapter 10, not squeezed into whatever March leaves over.
The open government items go on a named list, not into the anxiety.
Statement mechanics, final forms, filing schemas: watched, owned, and revisited on a schedule. Chapter 8's rule holds to the end: a missing form pauses nothing, and panics nothing either.
Readiness fits on one page you can read in a meeting.
Map done. Records dated. Reviews signed. Opens named, with owners. Four lines, honest answers. If a line can't answer honestly, that line is the plan.
Ninety days from now, the scramble is someone else's story.
One fund, one schedule, this year. Start the clock.
Alex closes the handbook. Six frames on what to keep.
The filing is the last page.
A return shows the answer and almost nothing about the path that produced it. You now know the path is the product. The form is just how it ships.
No one person holds the whole story.
Sponsor, CPA, Officer, bookkeeper, property, payroll, investors. The system crosses desks by design. Routing it well was never overhead. It was the work.
Proof is part of the answer.
A number is not finished because it sits in a spreadsheet. It's finished when a stranger can retrace it: source, date, method, destination. Every chapter was secretly this sentence.
Open guidance is not a reason to wait.
Forms will move fields. Regulations will set mechanics. The record you build now, sourced and dated, survives whatever formatting arrives. The statute is the spec until the spec shows up.
Three tags ran through this book. Here's the ledger.
Law: enacted statute, final regulations, current forms. Advice: OZfile practice that final rules may reshape. Pending: mechanics Treasury still owes. Every chapter wore its mix on the title card.
The enacted floor.
Sections 1400Z-1 and 1400Z-2 as amended, the reporting pair 6039K and 6039L, the penalty set 6722, 6724, and 6726, the §1.1400Z2 final regulations, and current Forms 8996, 8997, and 8949 with their instructions. Plus the published guidance: Notices 2025-50 and 2026-40, Rev. Proc. 2025-32.
What Treasury still owes.
The §6039K return with its due date and attachments; the §6039L statement's time, manner, and exact contents; the official meaning of 'amount disposed'; unit-count cadence; the employment ranges or indicator; QROF measurement mechanics; the next Form 8996. Watched, owned, revisited.
What's ours, until the rules speak.
The review order, the officer sign-off, delivery proof, the February 1 house deadline, labeled supersedes, the master request list, the one-page register. Habits built to survive whatever format arrives, and labeled honestly as habits.
The whole handbook reduces to four questions.
What happened. Who holds the proof. Which date does it speak to. Where does it go. Carry those into every reporting year and the rest of this book comes with them.
Written by OZfile. Checked against the sources below.
Statute, final regulations, current IRS forms and instructions, and published transition guidance. Last reviewed August 1, 2026, and current to that date. The pending mechanics can move; this edition moves with them.
Every Opportunity Fund tells the IRS a story.
Begin with one fund. Ninety days, this year, four questions at a time. Make sure yours is ready to be told.
Sources: IRC §§1400Z-1, 1400Z-2, 6039K, 6039L, 6722, 6724, 6726 · §1.1400Z2 final regulations · §4980H FTE method · Forms 8996, 8997, 8949 and instructions · Notices 2025-50 and 2026-40 · Rev. Proc. 2025-32 penalty indexing · TIGTA 2022 program audit. Educational content, not tax or legal advice.