The Little Used Federal Backstop for Manufactured Home Loans
HUD’s monthly tables reported just three Title I manufactured housing endorsements in 2025. Understanding that small number means looking beyond the house to the land, the lender and the loan.
Begin the story
The number is three.
Add the manufactured housing endorsement counts in the Department of Housing and Urban Development’s twelve monthly Title I tables for 2025, and that is the total. The corresponding total for 2024 is zero. For January through August 2026, the latest period listed on HUD’s production index when checked on October 19, it is also zero. These are the tables’ reported new insurance endorsements, not a count of loan applications or approvals. HUD monthly production reports, Table 3
That is a remarkably small new flow for a federal program intended to help finance manufactured homes. But it is not evidence that only three Americans financed one, or that federal support for the entire market has disappeared. The same August 2026 table records an existing Title I manufactured housing portfolio of 4,396 insured loans. Other government-backed and private lending routes operate outside this narrow program line.
The revealing contrast is between a financing tool that exists in law and the new activity visible in its published reports. A program description, a higher loan limit and an available loan are three different things.
Manufactured housing makes those differences unusually important. The building can be a permanent place to live while its legal classification, the land beneath it and the security offered to a lender follow different paths. A lower purchase price does not make that financing architecture disappear.
Here, manufactured housing means homes built under the federal manufactured housing standards. It is not a name for every factory-built building. Modular homes follow applicable state and local building codes, a distinction also set out in the 2026 housing law. Public Law 119-101, section 303, definitions
Three endorsements in a full year
Reported Title I manufactured housing endorsements
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2024 | ||||||||||||
| 2025 | ||||||||||||
| 2026 | Not covered | Not covered | Not covered | Not covered |
All monthly values and original sources
| Month | Reported endorsements | Original source |
|---|---|---|
| Jan 2024 | 0 | HUD Table 3 |
| Feb 2024 | 0 | HUD Table 3 |
| Mar 2024 | 0 | HUD Table 3 |
| Apr 2024 | 0 | HUD Table 3 |
| May 2024 | 0 | HUD Table 3 |
| Jun 2024 | 0 | HUD Table 3 |
| Jul 2024 | 0 | HUD Table 3 |
| Aug 2024 | 0 | HUD Table 3 |
| Sep 2024 | 0 | HUD Table 3 |
| Oct 2024 | 0 | HUD Table 3 |
| Nov 2024 | 0 | HUD Table 3 |
| Dec 2024 | 0 | HUD Table 3 |
| Jan 2025 | 1 | HUD Table 3 |
| Feb 2025 | 0 | HUD Table 3 |
| Mar 2025 | 0 | HUD Table 3 |
| Apr 2025 | 0 | HUD Table 3 |
| May 2025 | 0 | HUD Table 3 |
| Jun 2025 | 0 | HUD Table 3 |
| Jul 2025 | 0 | HUD Table 3 |
| Aug 2025 | 0 | HUD Table 3 |
| Sep 2025 | 1 | HUD Table 3 |
| Oct 2025 | 0 | HUD Table 3 |
| Nov 2025 | 0 | HUD Table 3 |
| Dec 2025 | 1 | HUD Table 3 |
| Jan 2026 | 0 | HUD Table 3 |
| Feb 2026 | 0 | HUD Table 3 |
| Mar 2026 | 0 | HUD Table 3 |
| Apr 2026 | 0 | HUD Table 3 |
| May 2026 | 0 | HUD Table 3 |
| Jun 2026 | 0 | HUD Table 3 |
| Jul 2026 | 0 | HUD Table 3 |
| Aug 2026 | 0 | HUD Table 3 |
What the federal backstop actually does
Title I is insurance for loans made by private lenders. It can cover a manufactured home, a lot, or a home and lot together. The borrower still needs a lender willing to make the loan and must satisfy the applicable requirements. HUD is not handing every buyer a direct loan or a grant. HUD program description
Insurance changes who bears part of the loss if a covered loan goes wrong. It does not remove the need to assess the borrower, process the application, fund the loan and service it afterward. The statute limits an individual claim payment to 90 percent of the loss, not 90 percent of the original loan as a gift to the buyer. 12 USC 1703, version current through October 9, 2026
Those obligations help explain why writing a program into existence is only the beginning. Public support may make a product more attractive, but a lender must still decide it is practical to offer.
The tiny reported endorsement flow does not tell us which obstacle is decisive today. It cannot separate limited eligible demand from lender economics, administration, risk management or borrowers using other products. Nor do the tables identify families who wanted a loan but could not get one. They establish a much narrower fact: very little new manufactured housing insurance activity appears on this particular endorsement line.
That limitation does not make the number meaningless. It makes the next question more precise. What would need to connect the federal backstop to a loan a buyer can actually obtain?
The house and the land are different financial objects
Three questions sit beneath the phrase “manufactured home finance.”
Who owns the land? How is the home classified under the relevant law? And what property is pledged as security for the loan?
Those questions are related, but their answers are not interchangeable. An owner can hold title to the land without pledging it for a particular loan. A home can be classified as real property even when a loan is secured only by the home. Federal mortgage disclosure instructions explicitly recognize that second possibility. Regulation C, section 1003.4 and its official interpretations, December 31, 2025 version
In Oftu’s selected 2025 public mortgage-disclosure cohort, 12,494 of 56,209 home-only origination records, about 22 percent, report direct ownership of the underlying land. That does not establish that those borrowers qualified for a different mortgage. It shows why “home-only” cannot simply be translated into “does not own land.”
The same care is needed with the word chattel, meaning personal property. A personal-property loan and a mortgage secured by real property can have different requirements and protections. But the home-only collateral field in the public data is not a reliable substitute for a complete legal-title record. It would be wrong to label every home-only observation a chattel loan.
Nor is a personal-property loan outside all federal consumer protection. Regulation Z defines a dwelling to include a mobile home or trailer used as a residence, whether or not attached to real property. Other rules have different coverage boundaries. The specific transaction matters. Regulation Z, sections 1026.2, 1026.19 and 1026.43, Regulation X, section 1024.2
There can also be legitimate reasons not to pledge land. GAO’s 2023 review identified borrower preferences, tax considerations and difficulties obtaining small mortgages among possible explanations. A route that offers access to a broader mortgage market can also require additional paperwork, approvals or security over property the borrower would prefer to keep outside the loan. GAO, September 26, 2023, especially pages 17–18
The practical lesson is not that one label always identifies the right loan. It is that ownership, legal title and loan security each deserve their own line in the explanation.
Three separate questions
Owning land is not the same as pledging it.
Land ownership. The homeowner, a cooperative, a landlord or someone else may own the land. Direct land ownership does not establish that the land secures this particular loan.
All three definitions
- Land ownership
- The homeowner, a cooperative, a landlord or someone else may own the land. Direct land ownership does not establish that the land secures this particular loan.
- Legal title
- A home may be real property or personal property under the applicable law. HMDA’s home-only collateral field is not a complete legal-title record.
- Loan security
- The loan may be secured by the home only or by the home and land. Even a home classified as real property can be reported as home-only collateral.
Refinancing looks very different across the two markets
The financing distinction is visible in the mix of loans reported under the Home Mortgage Disclosure Act, or HMDA.
The refinance mix changes with the lender group
Refinance or cash-out refinance records as a share of selected 2025 originations.
Loan composition, not an individual approval chance. Both bars use the same 0–100% scale. The unfilled portion is purchase loans.
Both populations in a source table
| Population | Collateral | Refinance / cash-out | All selected originations | Share |
|---|---|---|---|---|
| Full reported cohort | Home only | 1,029 | 56,209 | 1.8% |
| Full reported cohort | Home and land | 29,126 | 97,747 | 29.8% |
| Seven reporting IDs | Home only | 376 | 46,517 | 0.8% |
| Seven reporting IDs | Home and land | 143 | 11,026 | 1.3% |
Among the selected 2025 home-only origination records for purchases, refinancing and cash-out refinancing, 1,029 of 56,209 were refinances or cash-out refinances. That is about 1.8 percent. Among the corresponding home-and-land records, 29,126 of 97,747 were, about 29.8 percent. These are Oftu calculations from the 2025 public HMDA snapshot, using the filters described below.
The gap is striking, but its meaning needs care. This is a comparison of the composition of newly originated records. It is not the percentage of existing homeowners who successfully refinance, and it is not an applicant’s probability of approval. There is no register of all outstanding homeowners in the denominator.
The broad distinction is not a new discovery. The Consumer Financial Protection Bureau documented a similar refinance-composition difference using 2019 data. Pew, GAO and FinRegLab have examined manufactured housing finance and its barriers in detail. The useful contribution here is an updated public-data check, together with tests that make a simple explanation less comfortable. CFPB, May 2021
One such test changes the lender population. Restrict the 2025 comparison to the seven reporting lender identities with at least 100 originations in each collateral category, and the refinance shares become about 0.8 percent for home-only and 1.3 percent for home-and-land.
That is not an adjusted estimate of what the same borrower would receive. The selected group contains about 83 percent of the full home-only cohort but only 11 percent of the home-and-land cohort. Changing the group changes the market being described. It demonstrates how much the headline comparison can depend on which lending channels and customers appear in it.
Public HMDA does not provide actual credit scores or a matched account of otherwise identical borrowers, homes and offers. A separate attempt to make a rate comparison at an exact 30-year maturity also lacked enough observations on both sides of the selected purchase-loan cell. A broad rate difference should not be turned into a precise surcharge caused by collateral alone.
The evidence supports a financing-system question, not a universal verdict about the price or availability of a loan for an individual family.
A concentrated market is not an absent market
The reporting landscape is different too. In the selected 2025 cohort, the five largest reporting lender identities account for about 82 percent of home-only originations, compared with about 30 percent of home-and-land originations. Reporting identities are not necessarily independent corporate groups, and national shares do not measure the choices available in a particular town.
Concentration alone does not establish unlawful steering, collusion or why a particular loan costs what it does. But it is another reason to look beyond the physical home when describing the market.
It would also be a mistake to say home-only loans have no secondary market. In the selected public records, 19,413 of 56,209 home-only originations in 2025 have a reported same-year purchaser, about 34.5 percent. In the 2024 snapshot, the corresponding share is about 25.9 percent. The 2025 records include private-securitizer purchasers.
These fields do not observe every eventual sale. A loan transferred in a later year is not captured as a same-year sale, and the reporting rules have other conventions. Still, the records show activity, not an empty market. 2024 public HMDA snapshot, 2025 snapshot, official purchaser reporting instructions
Private refinancing exists as well. Triad, for example, publishes a rate-and-term refinance product for primary-residence chattel loans, subject to its requirements. Its own page warns that refinancing can increase total finance charges over the life of the loan. A product page is not an approval or a promise of savings. Triad product page, checked October 19, 2026
That warning points to a trade-off worth making visible.
A smaller payment can mean a larger interest bill
For the same principal and interest rate, spreading repayment over more years lowers the monthly principal-and-interest payment. It also leaves the balance outstanding for longer, increasing total interest if the loan runs to its scheduled end.
Compare 20 and 30 years
Illustrative $100,000 fixed-rate loan at 8%. Held to maturity.
20 years
/ month
interest
30 years
/ month
interest
Total repayment (principal and interest)
About $103 less each month.
About $63,410 more scheduled interest.
Amount: $1,000–$500,000. Rate: 0–20%. Both terms use the same amount and rate.
Consider an illustrative $100,000, fully amortizing fixed-rate loan at 8 percent. Over 20 years, the monthly payment is $836.44 and scheduled interest totals about $100,746. Over 30 years, the payment is $733.76, but scheduled interest rises to about $164,155.
The longer term frees roughly $103 a month while adding roughly $63,410 of scheduled interest. Neither choice is automatically irrational. Monthly cash flow matters, and actual borrowers may sell, refinance or repay early. But “more affordable this month” and “less expensive over the whole term” are different claims.
This calculation is not a current loan offer or a comparison of actual Title I and private products. It excludes fees, mortgage insurance, taxes, land rent, maintenance and other costs. Changing those inputs, or changing the interest rate, can change the comparison.
A manufactured home’s purchase price is only one part of that household budget. Financing a home does not eliminate the cost of occupying and maintaining its site.
Real alternatives can bring real costs
Resident ownership of a community addresses a different part of the problem: control over the land beneath the homes.
NHPR’s Jackie Harris reported in July 2026 that Albany Acres residents in New Hampshire had bought their community with support from the New Hampshire Community Loan Fund. The reporting also described a planned lot-rent increase from $455 to $600 to support improvements, alongside applications for grants to keep costs down. Those were reported plans, not a verification of October bills or grants received. NHPR, July 13, 2026
The example matters because greater control and a lower immediate payment are not the same benefit. Residents can gain a say over their community while still facing the cost of acquisition, maintenance and infrastructure. Collective ownership does not make those bills vanish, and a community-acquisition loan is not the same as a mortgage on an individual home.
Nonprofit and private financing routes can therefore be important without being interchangeable with a federal insurance program. Their usefulness depends on the borrower, the property, the place and the support around the loan.
The rules are changing but the results need their own measure
The public record does not support a claim that nothing has been done.
HUD increased Title I manufactured housing loan limits in 2024. In January 2025 it adopted the industry-standard Uniform Residential Loan Application, with mandatory use for relevant case numbers 120 days later. The letter described that change as a way to reduce the burden of separate program forms. HUD Title I Letter 488, March 18, 2024, Title I Letter 489, January 8, 2025
The July 11, 2026 housing law changed statutory loan parameters and allowed HUD to determine maturities up to a ceiling of 30 years. It gave HUD a year to develop or choose the new indexing method. That deadline is in July 2027, not a deadline already missed. A statutory ceiling is not an offer of a 30-year loan to every buyer. Public Law 119-101, section 303
State and conventional-mortgage changes matter too. Washington’s 2025 title law added qualifying cooperative interests to its land-ownership definition. The current application form asks for cooperative documents, an actual implementation step beyond an announcement. Consent and filing requirements remain. Freddie Mac’s September 2, 2026 bulletin also expanded eligibility for some previously relocated manufactured homes, with inspection and other conditions. Washington enacted law, Washington application instructions, Freddie Mac Bulletin 2026-12
Not every recent change expands access. HUD’s March 2025 residency-policy revision removed the non-permanent-resident eligibility category while retaining specified eligible categories, including lawful permanent residents under the applicable requirements. It did not make all noncitizens ineligible. The records examined here do not quantify its effect on Title I demand. Title I Letter 490, March 26, 2025
Industry expectations are not yet outcomes either. In UMH’s company-linked August 2026 earnings transcript, CEO Samuel Landy described expected benefits from expanded financing for home sales and demand for community lots, while acknowledging uncertainty about when the expected changes would take effect. That is a business expectation, not a count of completed insured loans. UMH Q2 transcript, August 6, 2026, page 10
The right test is neither to dismiss reform nor to count each announcement as a household helped. It is to follow the sequence: a rule changes, a usable process exists, lenders offer a product, and borrowers obtain loans on terms that serve them.
The latest HUD construction roadmap, published in October, is later than the August endpoint of the production series. These numbers cannot judge its effects. HUD action plan, October 7, 2026
A changed rule is only the beginning
- March 2024HUD raises Title I manufactured housing loan limits
- May 2025Industry-standard application becomes mandatory
- October 2025Washington’s qualifying cooperative-interest change takes effect
- July 2026Congress changes statutory loan parameters
- September 2026Freddie Mac permits some relocated homes, with conditions
- Rule
- Process
- Lender offer
- Completed loan
Measure the route to the loan
Manufactured housing finance is not one market with one blocked door. It is a set of routes through land arrangements, legal classifications, lenders, insurance and investors.
The low Title I endorsement line makes one part of that system worth watching closely. The larger public loan records show why it should not stand in for the whole system. There are private loans, refinancing, transfers and policy changes, alongside sharp differences in the markets being served.
For buyers, the useful question is not simply whether a program exists. It is whether a lender offers a suitable product for this home, on this site, with these legal arrangements, at a cost the household can carry.
For policymakers, the useful measure is not the number of promising announcements. It is the path from a changed rule to an available product and a completed loan, with the costs and risks visible along the way.
Sources and method
Research cutoff: October 19, 2026. Human contribution is assigned at 64 percent, with 36 percent AI. This is an illustrative editorial allocation, not a measured share of words or a detector score.
The HUD totals sum the manufactured housing endorsement loan-count line in Table 3 of each monthly FHA production report. Calendar 2024 and 2025 each have twelve reports; 2026 covers January through August only. An endorsement records insurance acceptance and need not coincide with the month of origination. The tables also include separate net adjustments for late-posted prior-period activity. Those adjustments do not reveal the gross transactions behind them and are not added to the endorsement count. The endorsement total is not an independently reconciled census of every underlying insurance event. The existing insured portfolio is a separate stock measure.
HMDA calculations use the public 2024 and 2025 snapshots already acquired by Oftu on October 13, 2026. The main comparison selects owner-occupied manufactured housing, one to four units, first liens, non-preapproval originations explicitly coded as forward, closed-end and noncommercial. It includes purchase, refinance and cash-out refinance purposes and compares collateral codes for home-and-land and home-only loans. Exempt or unavailable collateral categories are kept separate, not assigned to either group. HMDA covers reporting institutions, not cash purchases or every loan.
The reported cohort includes property-state codes for the 50 states, DC, Puerto Rico and unavailable locations. Restricting the 2025 comparison to the 50 states and DC gives 1,024 refinance or cash-out records among 56,155 home-only originations, versus 29,116 among 97,686 home-and-land originations. The descriptive pattern is almost unchanged. Neither version measures current households or individual refinance chances.
The common-lender sensitivity selects seven reporting identities with at least 100 originations in each collateral category in the full three-purpose 2025 cohort. Its refinance counts are 376 of 46,517 home-only records and 143 of 11,026 home-and-land records. This changes the populations; it is not a matched-borrower or causal adjustment. Reporting identities need not be separate corporate groups.
Rates are reported interest rates, not APRs or current quotes. Coarse product, amount and maturity groupings leave important differences unresolved. An additional exact-360-month test in the selected conventional-purchase $150,000-to-under-$250,000 band did not meet the predeclared minimum of 100 numeric-rate observations on both sides. No paired rate result from that cell is presented.
The payment illustration uses the standard level-payment formula, a monthly rate equal to the annual nominal rate divided by twelve, scheduled monthly payments and no extra payments. Results are rounded for display. It assumes the loan runs to maturity and is not a forecast of any borrower’s behavior.
The underlying subject has a substantial research history. Alongside the primary records linked above, useful prior work includes FinRegLab’s December 2024 qualitative study and Pew’s February 2026 titling analysis, updated April 15. Their evidence, periods and comparisons are not identical to Oftu’s selected 2025 cohorts. Regional experiences are credited to the original reporters. No new interviews or information requests were used for this article.