I thought my own labor counted toward the improvement number. It doesn't.

The sweat equity thing catches every construction person who comes at this from our side. I did the same arithmetic in my head about my own crew's time on a self-performed job and it took a hard conversation to understand that the cost has to actually leave the entity.

What I don't understand yet is whether paying yourself properly through a company changes it. solder did anyone explain that part?

On girder's question, the mechanism to understand is that the substantial improvement test measures additions to the basis of the building. Basis grows from amounts the owning entity actually incurs and capitalizes. Unpaid labor by the owner creates no expenditure, so there is nothing to capitalize. Whether wages or contract payments genuinely paid by the entity to a related party get counted, and on what terms, depends on how the related party rules and the capitalization rules apply to the specific arrangement, and that is a question for a CPA reviewing the actual entity documents and payroll records.

The more expensive item in solder's post is the ordering. The incentive works by reinvesting eligible gain into a qualified opportunity fund within the applicable window, and the fund then holds or invests in the qualifying property. Buying property personally first and hoping to contribute it later runs into the requirement that the property be acquired by the qualifying entity by purchase after the relevant date. That sequencing is not something to improvise, and confirming it in writing with a tax professional before closing costs a fraction of what discovering it afterward costs.

One timing note that matters for anyone reading this now rather than in 2024. The current set of designated zones sunsets at the end of 2026 and a newly designated set takes effect January 1, 2027, with the enhanced benefits under the One Big Beautiful Bill Act attaching to investments in the new zones. That gap changes when a project is worth starting, separately from whether it pencils.

The order orchard gave was gain, then window, then entity, then buy. The word "then" is doing a lot of work in that sentence and I would absolutely have gotten it backwards.

orchard's framing is right and I'd add the document you'd have needed. The fund's organizational documents and its self-certification are what establish it as a qualifying entity, and those exist as of a date. There's no retroactive version of a date. That's why the ordering isn't a formality.

solder, when your accountant looked at it in February, did anyone raise whether a contribution of the property into a new entity was worth exploring at all, or was the window issue simply fatal?

1,500 for an hour with a CPA versus 18,400 in avoidable cost. I'm starting from zero and this is the cheapest lesson I've read all month.

I'll say the unpopular version. Even if you'd nailed the structure perfectly, 96k of gain into a beat up two story you were going to self-perform over a year was always going to be a hard deal to make the numbers work on, because your labor is real cost even when it's uncompensated. You just weren't going to see it on the spreadsheet. The tax mistake made this visible sooner than the operational one would have.

keystone that's a bit harsh but it's the right question. solder, how many hours did you actually put in and what would you have paid a crew for the same scope? That number tells you whether the deal worked at all.

I've been slow to pull the trigger, and threads like this are the reason I've stayed that way. My checklist now has one item ahead of everything else, which is a conversation with a CPA about ordering before I make an offer in any designated tract. Everything else can be figured out at my own pace. That can't.