House Flipping 2026: Flip Loans by State and the Dealer Tax Rules

What a flip looks like in federal loan data

The top Google results for this subject explain the steps of a flip and quote a rule of thumb for the purchase price. None of them shows what flippers' lenders actually reported. Lenders above the HMDA reporting threshold file every decided application with the loan purpose, occupancy, term, amount, property value and interest rate. A loan for “a commercial or business purpose” is excluded from HMDA unless it is a home improvement loan, a home purchase loan or a refinancing (12 CFR 1003.3(c)(10)), so a business-purpose loan to buy a house, which is what a flip loan is, is reported.

We took the nationwide 2025 file of originated and denied applications that we streamed from the FFIEC Data Browser and kept first-lien, closed-end, conventional home purchase loans on one-to-four-unit investment properties made for a business purpose with a term of 24 months or less. That is our proxy for flip financing; HMDA has no field called “flip”. A long hold is not in it, and neither is a flipper who paid cash, borrowed from a private lender below the reporting threshold, or used a line of credit.

Short-term investor purchase loans, 2025 (originated)

Shares and structure (our arithmetic): 91.0% were on a single-unit property, 6.0% on two units, 1.8% on three and 1.2% on four. By property value, 7.0% were at $150,000 or less, 20.2% above $150,000 to $250,000, 26.8% above $250,000 to $400,000, 17.3% above $400,000 to $600,000, 14.6% above $600,000 to $1 million and 14.2% above $1 million, among the 46,304 loans with a value. Terms were 12 months on 79.5% (38,579 loans), 6 months on 4.0%, 15 months on 3.6%, 18 months on 3.5%, 24 months on 3.3% and 9 months on 2.6%. 95.3% were interest-only and 94.3% had a balloon payment. 86.0% of the borrowers were reported as not a natural person (a company), which is how Regulation C has a lender report the applicant's age when the borrower is an entity (12 CFR 1003.4(a)(10)(ii) and the Regulation C commentary on it).

The loan-to-value ratio here is the loan divided by the property value the lender relied on, both rounded in the public file (our arithmetic). On a flip that value may be the as-is value or, for some lenders, a value after repairs, and the loan may include a repair holdback, so the ratio is a market-level reading and not a rule: do not read 70.7% as the leverage you will be offered.

What the median flip costs to carry

The one cost HMDA does measure is interest. On the median loan ($265,000) at the median rate (10.2%), interest-only, a year of interest is $27,030 and a month is $2,252 (our arithmetic). That is before points, the origination fee, draw and extension fees, insurance, taxes, utilities, the repair budget and selling costs, none of which is in the federal file: the origination-charge, discount-point, total-loan-cost and prepayment-penalty fields are empty on all 48,550 loans. The consequence is simple: a flip that takes 12 months instead of 6 costs about $13,500 more in interest on the median loan (our arithmetic), and 79.5% of the loans had a 12-month term. For a calculator that does the rest, see our fix and flip calculator; for what lenders charged, see hard money loan rates.

Where flips are financed: the 15 largest states

California alone accounted for 14.5% of the loans, and its median property was $925,000, against $205,000 in Ohio and Michigan. The right-hand column is the useful one for comparing places: short-term loans as a share of every first-lien investor purchase loan reported in the state.

Short-term investor purchase loans by state, 2025 (HMDA, our arithmetic)

Read the table with the lender concentration in mind. The 547 lenders with a short-term origination in the file are very uneven: the largest, Kiavi Funding, made 18,194 of the 48,550 loans (37.5%) and the ten largest made 72.1%. A state where one lender is active will look busy, and a state's median rate reflects that lender's pricing: Missouri's 8.15% median is far below the 10.2% national median; we did not test why, and with a few lenders setting the pricing it should not be read as cheaper flip money for every Missouri borrower (our reading). The state columns sum to 48,455 loans because 95 loans carry no state code in the extract.

The largest metro areas (metropolitan divisions where the Census Bureau defines them), same measures:

Another 7,489 loans (15.4%) are outside any metropolitan area code in the file. Among areas with at least 300 loans, short-term loans were the largest share of investor purchases in Camden (45.5%), San Jose (35.7%), Oakland (33.4%), Virginia Beach (33.2%) and Kansas City (32.5%), and the smallest in Orlando (7.8%) and Houston (8.8%). A high share means flips are a large part of investor buying that reached a reporting lender, or that a flip-focused lender is local; it does not mean flips are more profitable there.

What this data cannot tell you

• It is financing, not flips. A short-term loan may be a bridge for a buyer who then keeps the house, and a flip paid in cash does not appear. We did not find an official count of flips from a government source.
• It has no sale price. HMDA records the purchase, not the resale, so it cannot give a profit or a holding time. Nobody can compute a flip margin from these records, and any such margin you read elsewhere comes from a private data set.
• It has no fees. Points, origination fees and prepayment terms are empty (see above), so the real cost of money is higher than the note rate.
• Reporting is partial. Lenders under the HMDA thresholds, and a loan that a lender reports as exempt under the partial exemptions, are not in the file. We counted only the 48,550 originated loans that met every filter.

When a flipper is a dealer: what the statutes say

The federal tax treatment of a house you buy to resell turns on one phrase. The Internal Revenue Code defines a “capital asset” to exclude “property held by the taxpayer primarily for sale to customers in the ordinary course of his trade or business” (26 U.S.C. 1221(a)(1)). The Supreme Court held in 1966 that, as used in that provision, “primarily” means “of first importance” or “principally” (Malat v. Riddell, 383 U.S. 569). It rejected the argument that a substantial purpose to sell is enough. Whether a given flipper meets the test is a question of facts, and the Court sent the case back for fact-finding on that basis (our reading). What follows if the answer is yes:

Two things follow, both our reading of the text. First, the one-year holding period that decides long-term gain on an investment house does not rescue a flipper who is a dealer; the question is why the house was held, and the Court's answer is the purpose “of first importance”. Second, a flipper cannot rely on a 1031 exchange to defer the gain on a house bought to resell; 1031 can still apply to a rental held for investment, which is a different property in the same portfolio. See our 1031 exchange rules and capital gains tax on real estate pages for the investor side. We have not quoted Tax Court decisions on flips: we could not retrieve a court-hosted text we could save, and the decisions turn on facts of individual taxpayers. A tax adviser who has read the cases is the right person for the line between investor and dealer.

The buyer's side: two federal rules that bite on quick resales

A flipper's exit is usually a retail buyer with a mortgage, and two federal rules are written for resales that happen soon after the flipper's purchase.

• FHA. The regulation says: “If the re-sale date is 90 days or less following the date of acquisition by the seller, the property is not eligible for a mortgage to be insured by FHA” (24 CFR 203.37a(b)(2)). Between 91 and 180 days it is “generally eligible”, but “HUD will require that the mortgagee obtain additional documentation if the re-sale price is 100 percent over the purchase price”, including an appraisal from another appraiser (24 CFR 203.37a(b)(3)). Sales by HUD and by financial institutions, and some others, are excepted (24 CFR 203.37a(c)).
• Higher-priced mortgages. A creditor “shall not extend a higher-priced mortgage loan to a consumer to finance the acquisition of the consumer's principal dwelling” without two written appraisals if the seller acquired the property 90 or fewer days earlier and the price is more than 10 percent above the seller's acquisition price, or acquired it 91 to 180 days earlier and the price is more than 20 percent above (12 CFR 1026.35(c)(4)(i)). The rule applies only to higher-priced mortgage loans; most prime loans are outside it (our reading).

The practical reading for a flipper is that a resale inside 90 days cannot go to an FHA buyer at all, and a large jump inside 180 days brings a second appraisal that can delay or reprice the sale (our reading). Our down payment data page shows what investors who keep the house put down.

What a reader can do with this

1. Price the clock before the house. At the median, each month costs about $2,252 in interest alone. Put the number in the deal sheet with a 12-month term and a 6-month stress case, then add the fees you are quoted.
2. Read your own market's row, not the national median. The state and metro tables show loans, property values and rates; a market with a $205,000 median property and a $145,000 median loan is a different deal from one at $925,000 and $685,000.
3. Decide the tax question with your adviser before the first sale, not after the fifth. The statutes above make the answer depend on why and how you hold houses; the consequences (ordinary rates, no 1031, self-employment tax) are large enough to price in.
4. Check the buyer pool. If your exit is an FHA buyer, a resale inside 90 days is closed to them, and a jump in price inside 180 days draws a second appraisal.
5. Compare lenders on the full cost. Our best fix-and-flip lender page compares named lenders; our fix and flip loans page covers denials, loan structure and delinquency.

These links go to the lenders' own sites. This page does not say that any of the three lends on a flip in your market or on terms near the medians above; the HMDA table names Kiavi as the largest short-term lender in 2025, and the other two are included because they are the lenders our other borrower guides compare.

FAQ

Loan-level data: FFIEC/CFPB HMDA Data Browser, nationwide 2025 file of originated and denied applications (actions 1 and 3), streamed on October 10, 2026 with our script stream_extract.py (rows for investment properties and two-to-four-unit properties kept: 906,337) and analyzed on October 13, 2026 with hmda_flip.py (first-lien, closed-end, conventional home purchase loans on one-to-four-unit investment properties, business purpose, term of 24 months or less); the scripts, the extract log, the output and the tables are in sources/. Field definitions from the FFIEC public LAR data fields; lender names from the FFIEC 2025 filer list; metro names from the Census Bureau's 2023 metropolitan delineation file. Rules: 26 U.S.C. 1221, 1222, 1231, 1031, 1401 and 1402 (U.S. Code, 2024 edition, govinfo.gov); Malat v. Riddell, 383 U.S. 569 (1966) (United States Reports, Library of Congress); 12 CFR 1003.2, 1003.3, 1003.4 and Regulation C comments, 12 CFR 1026.35 and 24 CFR 203.37a (Electronic Code of Federal Regulations, current as of October 1, 2026, read October 13, 2026). All percentiles, shares, medians and dollar figures from HMDA are our arithmetic. This is analysis of public documents, not investment, legal or tax advice.