The unit that pays best is the one the lender will not touch

The post lands on something real, and it is worth spelling out the mechanics for anyone reading this who has not seen a self-sufficiency test before.

The self-sufficiency test is a lender requirement on certain conventional loans for two-to-four-unit properties. The basic version works like this: take 75% of the gross scheduled rent (75 cents of every dollar the units are supposed to collect), and that number has to cover the full PITI, meaning principal, interest, property taxes, and insurance. If it does not, the loan does not move forward regardless of what the property actually earns.

So on a fourplex collecting $4,200 a month gross: 75% of that is $3,150. If PITI on the loan is $3,400 a month, the test fails by $250. The property is cash-flowing fine. The loan says no.

The unpermitted unit problem makes this worse in a specific way. The lender's appraiser is required to value and count only legally permitted space. If one unit is a converted garage with no permit, the appraiser may exclude it entirely from the rent roll they submit. Now your gross rent drops from $4,200 to maybe $3,000, your 75% figure drops to $2,250, and the test fails by a wider margin. On top of that, some lenders will decline the loan outright if unpermitted space exists, because it creates a title and insurance risk they do not want to hold.

The strategy guide for small multifamily covers the financing-first structure clearly: the residential loan advantage is one of the main reasons this sector makes sense, and that advantage is only real when the property qualifies for it.

A permit status check and a lender pre-screen on the specific property are worth running before you spend time on the cap rate math.

What loan product are you working with? The self-sufficiency test applies differently depending on whether you are in a conventional owner-occupied scenario or a straight investment loan, and that changes which numbers matter first.

The unpermitted unit problem cost me eight weeks on a Cincinnati fourplex in 2021. Appraiser flagged a converted basement unit, lender pulled the income, and suddenly the DSCR on the remaining three units was 0.94 instead of 1.18.

What I run first now is the rent roll against the certificate of occupancy, unit by unit, before I ever build a proforma. If the CO and the rent roll don't match, I price the deal as if the non-conforming unit doesn't exist and see if it still pencils.

What does the seller's broker say when you ask them directly which units are on the CO?

The person who told you to filter by cap rate first is giving you advice that works fine on stabilized commercial where the financing is more flexible. On a 2-4 unit conventional loan it is genuinely backwards and I think it creates a false sense that you have done analysis when you have mostly just done arithmetic.

What I run first, before I look at a single rent figure, is the permitted use column on the tax record and the certificate of occupancy status. I have been going through Ocean and Burlington County auction packages for six months and the number of fourplexes where unit four is a converted garage or a basement egress that someone hung a door on is not small. The rent might be 1,100 a month and the seller built the whole asking price around it. The lender sees a three-unit with a storage problem, the appraiser comps it against three-unit inventory, and suddenly you are 80k upside down on the valuation before you even touch the self-sufficiency math. Cap rate told you it passed. Everything downstream said it did not exist. So, what does your spreadsheet actually use as step one, because I am curious whether you run title and CO status before you model rents or after?

The self-sufficiency test is one version of this problem, but the DSCR product version is where it gets structurally interesting. A DSCR loan at 1.20 on a 580k fourplex at 7.5 percent over 30 years means you need roughly 4,350 a month in qualifying rent just to clear the coverage floor, and the lender is not using your actual lease numbers, they are using the appraiser's market rent schedule from the 1007 addendum. If the appraiser pulls comps from a submarket that underperforms the subject property's actual leases, the DSCR fails on the appraisal, not the rent roll. You can have signed leases at 4,200 and a 1007 that returns 3,800, and the deal that penciled on your spreadsheet is now a denial letter.

The piece that changes the picture when you bring in the permitted-unit problem is what happens at the appraisal itself. A sales-comparison appraiser on a conforming loan is already constrained, but a non-conforming unit that the county shows as storage does not just lose its rent from the income approach, it can contaminate the sales-comparison approach by making it genuinely difficult to find true comparable sales that share the property's configuration. The appraiser then either brackets poorly or makes large downward adjustments, both of which compress the as-is value and tighten the LTV before the rate or the DSCR conversation even starts. So the unpermitted unit is not only an income problem, it is a collateral problem, and those two failures compound rather than stack. The question I would want answered before anything else on a deal like this is whether a permit pull is legally possible and what the cost and timeline look like, because that

The wrinkle that makes this worse is what happens when you try to fix the unpermitted unit before closing. Permit-and-legalize sounds like the obvious path, but the moment you pull that permit, the county assessor in most jurisdictions reassesses the square footage, which moves the property tax line in your operating statement, which changes your DSCR on the loan you are trying to close. I have seen the correction run $4,000 to $9,000 annually on a fourplex in a high-cost county, and that is not a rounding error when your DSCR was already sitting at 1.08.

The related strategy that changes the picture here is the seller-carry piece, because it is the only financing structure where the income definition is negotiated rather than dictated. If a seller carries 30 or 40 percent of the purchase price at a fixed note, the conventional first mortgage drops to a size where the self-sufficiency test clears on the conforming units alone, and the unpermitted rent becomes upside rather than a disqualifying line item. The tradeoff is that the seller now has a security interest in a property with a title complexity, so their attorney will find the permit issue anyway, but you are negotiating that as a price adjustment rather than a loan denial. The question I would run before anything else on a deal like this is: what is the permit history on that fourth unit, specifically whether any prior owner pulled and abandoned a permit, because an abandoned permit is sometimes harder to close out than starting fresh.

The cap rate versus financing order problem has a clean resolution when you separate two deal types: a deal you intend to hold with permanent financing versus a deal you intend to refinance into later. If you are underwriting to a conventional investment loan from day one, then the self-sufficiency test is the first filter by definition, because a deal that fails it does not exist in that form regardless of how the income pencils. If you are buying with a DSCR product or a portfolio lender who will underwrite to actual leases rather than agency guidelines, the constraint shifts but does not disappear, it just becomes the lender's coverage floor and their opinion of what rents are creditworthy. A non-permitted unit fails both paths the same way: the agency lender excludes the income entirely, and the portfolio lender will discount it or condition the loan on cure, because the appraiser's comparable analysis cannot support a rent that has no legal basis to exist.

The more expensive version of this is when the permitting gap is not obvious at origination and surfaces at appraisal, because by then you have rate locks, earnest money, and timeline pressure working against your negotiating position. The file to request before any of that is the certificate of occupancy matched against the current unit count, then the rent roll matched against that CO, then the county assessor's record to see what use is listed for each structure on the parcel. Those three documents usually disagree with each other on a problem property, and the disagreement is where the actual underwriting risk lives. A seller's proforma will almost never flag it. What does the rent roll you are looking at actually show for unit descriptions, and has the listing broker disclosed the CO status on the converted or accessory unit?

The assumption doing the most work in that 6.5 percent figure is that all four units are legal conforming residential space, because the moment one is not, you are not underwriting a fourplex, you are underwriting a triplex with a storage problem attached. On a conventional investment loan the appraiser's unit count controls the comparable selection, and if the comp pool shifts from four-unit to three-unit properties in that submarket, the appraised value can drop enough to change the LTV before you even get to the self-sufficiency calculation. So the financing constraint you described can hit twice: once on income and once on value.

What I would run before cap rate is a permit pull and a zoning check, because those two things determine which loan products will even consider the asset. A DSCR product and a conventional investment loan treat the same property differently, and a portfolio lender may count income differently still, but none of them will paper over an appraiser noting unpermitted space. The order that survives contact with an actual lender is legal unit count first, then which loan product accepts that count, then what that product requires for income coverage, and only then does the rate and the cap rate become a useful comparison.

On the fourplex you described at 580k and 4,200 gross, the self-sufficiency math at 75 percent of gross is 3,150 per month that has to cover full PITI, so the question is what interest rate and term produce a payment at or below that number, and at current conventional investment loan pricing on a 20 to 25 percent down payment that is a narrow window. What does the deal look like if you underwrite it as a triplex, dropping the non-conforming unit entirely?

What loan product are you actually in when you're underwriting that fourplex?

Arcadia deal, 2022. Four units on paper, five units on the ground. That fifth unit was a converted detached garage, permitted as an ADU in the city system but the county had it flagged as non-habitable storage from a 2009 code sweep that never got resolved. Appraiser caught it on inspection, lender killed the count, and suddenly the debt service math was underwater because we had underwritten that unit at $1,450 a month. We were 60 days into escrow. The spread I thought I had bought evaporated the week title opened, which is exactly why I keep saying title opens and the deal changes.

What I run first now is the lender's unit count, not the seller's. I call the county, pull the permit history, confirm every unit has a certificate of occupancy or at minimum a legal non-conforming status the lender will accept. Cap rate comes after that because cap rate built on an uncountable unit is just a story someone is telling you. On the Fresno sub-to I did last year the existing financing bypassed most of this, which is part of why subject-to appeals to me on properties where the conventional loan would choke on the income documentation. You are not originating, so the self-sufficiency test is not the hurdle it would be on a new loan. What type of financing were you planning to use on that fourplex?

My duplex came to me already financed at zero, which means I've never had to run these filters in sequence and I keep wondering if that's why I can't pull the trigger on anything else. No muscle memory for what breaks a deal at the loan level.

The unpermitted unit detail is the part that sticks with me, because "the rent is real, the lender won't count it" describes exactly the kind of thing that would not show up in a cap rate screen and would absolutely blow up the financing. Are sellers pricing these as if the unpermitted income is fully countable, or do they at least haircut it?

Seller financing sidesteps every lender filter and puts the unpermitted unit's rent back into the deal's arithmetic.

Confirm the legal unit count with the county assessor before the appraisal, not after.

On a $580k fourplex pulling $4,200 gross, that cap rate is already thin before you run any test. Around Cary I'm watching similar-sized properties sit at 5.8 to 6.1 realized after real vacancy and maintenance, not the 6.5 someone penciled on a napkin.

The question I run first is whether the appraiser will even count all four units as residential income, because if one gets flagged as unpermitted my file dies before the self-sufficiency math matters. Saw that exact sequence on a triplex in Durham two years ago, seller swore the third unit had been rented for a decade, appraiser marked it as storage, lender walked.

What's the permit history on the fourth unit you're looking at?

Lender order of operations matters more than cap rate sequence, and the place I start is the rent roll against what the assessor and the building department will confirm. On that fourplex at 580k, if one of the four units is a converted garage pulling 900 a month, you are building your cap rate on 4,200 gross but the lender's underwriter is working from 3,300, and the self-sufficiency test at 75 percent of that lower number has to cover a PITI that was sized against a 580k purchase price. The math breaks before the rate conversation starts. The more expensive version is when the appraiser flags the unpermitted unit as a liability rather than just ignoring its income, because now you have a property with a disclosed improvement the lender may require remediated as a loan condition, which can run more than the deal's margin.

The filter I actually run first is legal unit count confirmed against permit records, then what a lender will accept as countable gross rents, then whether that countable number passes the applicable coverage test at a rate I can actually get quoted today. Cap rate comes after, because a cap rate built on income a lender treats as zero is just a story about the property's potential.

What loan product are you targeting on deals like this, conventional investment or something in the DSCR space, because the coverage floor and the way the lender handles partial occupancy differ enough to change which deals survive the screen?

Permit status on every unit is the first thing I check, before any income number goes into a model, because one flagged structure collapses the income figure and sometimes the loan simultaneously. After that I want to know whether the lender's acceptable rent schedule matches the actual leases, since some portfolio lenders will credit month-to-month rents at a haircut that conventional lenders would not count at all. The third filter is which loan product is even available on this asset class in this county, because DSCR and conventional self-sufficiency thresholds are different math and the deal can clear one and fail the other.

What does the rent roll look like on the unit that is performing best, specifically whether it is on a signed lease the lender will accept or running month-to-month?

The risk nobody in this thread has named is what happens at the appraisal when a non-conforming unit is discovered mid-process rather than at the front end. The appraiser flags the unpermitted structure, the lender conditions the loan on its removal or legalization, and by that point the buyer has paid for the appraisal, the inspection, and possibly a rate lock that is now expiring. The permit legalization process in most jurisdictions runs 60 to 120 days at minimum, which blows past any standard closing timeline, and the seller often will not wait. That is not a financing problem at that point, it is a deal death that the buyer funded out of pocket to discover.

The practical fix is running the title and permit history pull before the offer goes in, or at minimum making the offer contingent on confirmation that all units are permitted for residential occupancy and that the rent roll the lender will accept matches the rent roll you used to price the deal. Those are two separate confirmations and both need to happen before the appraisal is ordered.

What does the permit pull on this specific fourplex actually show for the fourth unit?

What's the actual vacancy baked into that 4,200?

Financing constraint first is the right order, but the specific constraint changes by loan type, and that matters before you run any number. A conventional investment loan runs the self-sufficiency test; a small-balance commercial loan runs DSCR against actual collections with no self-sufficiency floor; a house-hack owner-occupied loan counts projected rents differently again. The fourplex at 580k might fail conventional and clear a portfolio DSCR product at the same purchase price, same income, different filter.

The unpermitted unit is where this gets expensive regardless of product, because the appraiser flags it first and the loan product question becomes moot until the permit status is resolved or the unit is removed from the income analysis entirely.

Which loan structure were you in when the self-sufficiency test killed it?