Appendix B — Loan Program Comparison Matrix

Six programs, side by side: conventional conforming, FHA, VA, USDA, jumbo, and non-QM. This is the page to open when a borrower describes a situation and you need to know, in ninety seconds, which doors are even worth walking to. Chapter references point to where each program is taught — programs in general in Chapter 5, FHA in Chapter 16, VA and USDA in Chapter 17, jumbo and non-QM in Chapter 34.

⚠️ READ THIS BEFORE YOU QUOTE ANYTHING ON THIS PAGE. Every dollar figure, percentage, score minimum, and ratio benchmark in this appendix is perishable. Loan limits change annually. Mortgage insurance factors change by bulletin. Funding fees change by statute. Minimum scores are partly agency and mostly lender. What is durable here is the structure — which program charges what, in what form, and whether it ever stops. Learn the structure; look up the number. An appendix that reads as current fact is the most dangerous page in a book like this, so every specific value below is marked [illustrative] and paired with the office that publishes the real one.

B.1 Where every number on this page actually comes from

Two habits that keep you out of trouble. First, when a borrower asks what the limit is, say "let me pull the current figure" and pull it — the number is a two-minute lookup and a memorized one is wrong within a year. Second, when a guideline surprises you, find it in the source above before you repeat it. Chapter 14 is built on that discipline.

Durable versus perishable, on this page:

B.2 The one-screen matrix

[all values illustrative — verify at the source in §B.1]

Read that grid as a set of doors, not a ranking. No program is "better." Each is a different answer to a different constraint, and the loan officer's job is to identify which constraint is actually binding on this borrower — cash, credit, income, geography, loan size, or documentation.

B.3 Mortgage insurance — the most important rows in this book

This is the table to know cold. Everything else in this appendix you can look up in front of the borrower without embarrassment. This one you should be able to say from memory, because borrowers ask about it constantly and the wrong answer costs them tens of thousands of dollars (§5.8).

⚠️ Never call FHA's MIP "PMI." They are different charges under different authorities with different termination rules, and the word you choose tells a borrower — and a regulator reading your file — whether you know the difference. PMI is private mortgage insurance on a conventional loan and it ends. MIP is the FHA's own insurance premium and, above 90% LTV, it does not.

B.3.1 The FHA duration trap

FHA's MIP duration is decided once, at origination, by the LTV, and is never revisited. A borrower who puts 3.5% down is above 90% LTV and carries MIP for the life of the loan — not until the balance falls below 90%, not until an appraisal shows equity, not ever. Paying the loan down does nothing. Values doubling does nothing. The only exits are refinancing out of FHA or paying the loan off.

That means a borrower who can reach 10% down is buying something entirely different from a borrower at 3.5%: at or below 90% LTV the MIP runs eleven years and then stops. The cliff between those two files is not smooth, and it is not a matter of a few dollars. If a borrower is close to 10% down, say so out loud before they spend the difference on furniture.

B.3.2 Conventional PMI has forms, and one of them never cancels

LPMI is the one to be careful with. It is not free insurance; it is insurance paid with a permanently higher interest rate. It quotes beautifully against BPMI on a payment-comparison sheet and it loses badly if the borrower keeps the loan long enough for BPMI to have terminated. Show both.

B.3.3 Conventional MI is priced; FHA's annual MIP essentially is not

Conventional PMI factors move with LTV and representative score. FHA's annual MIP moves with LTV and loan size in coarse bands and is indifferent to credit. That single asymmetry explains most program selection in the field.

The book's own anchor figures make the point [illustrative factors — verify current MI rate cards and HUD Mortgagee Letters]:

Five points of LTV cut the conventional factor nearly in half. The two FHA files carry the identical factor at wildly different loan sizes and a 65-point score spread. A strong-credit borrower is overpaying on FHA and a weak-credit borrower is often underpaying — which is exactly what the program was designed to do.

⚠️ A convention that trips people: the FHA annual MIP factor is applied to the total loan amount, base loan plus financed UFMIP, while program LTV is computed on the base loan. Both Linden Street's \$173.26 and Harlow Street's \$96.76 resolve only under that convention.

B.3.4 VA's funding fee is not insurance, and the exemptions are real money

The VA funding fee is a one-time charge that funds the guaranty program. It is financeable, it varies with the type of service, whether this is a first or subsequent use, and the down payment, and it is waived entirely for borrowers with a service-connected disability rating and for certain surviving spouses [exemption categories are statutory and are periodically amended — verify in the VA Lender's Handbook].

Ask about the exemption on every VA file. A borrower who is entitled to the waiver and pays the fee anyway has been charged thousands of dollars for nothing, and the loan officer who did not ask is the reason. On the book's VA counterfactual, the illustrative 2.15% first-use fee on a \$385,000 purchase is \$8,277.50 — financed into the loan, which is why that option shows a larger loan amount than the conventional file and a smaller payment [illustrative — verify the current fee table].

B.3.5 USDA charges twice and never stops

USDA loans carry an upfront guarantee fee, financeable into the loan, and a smaller annual fee collected monthly for the life of the loan [both percentages are set by USDA notice and are revised — verify]. The annual fee has historically been the smallest of the three government monthly charges, which is why USDA payments often beat FHA payments outright at the same price — but "smallest and permanent" is still permanent. Say it plainly: on USDA, the monthly fee never goes away.

B.4 Maximum loan amount, and who decides it

⚠️ "Jumbo" is defined by reference to a number that moves. A file can be jumbo in December and conforming in January without one fact about the borrower changing. When a loan sits within a few thousand dollars of the limit, check the limit for the year the loan will close, and remember that a slightly larger down payment can move a loan across the line into far better pricing. That conversation is worth having before the borrower writes the offer, not after.

Entitlement, in one paragraph, because it is the VA question loan officers get wrong. A veteran's entitlement is a guaranty amount, not a loan amount. With full entitlement restored, there is no VA loan limit — the constraints are the appraised value, the lender's own maximum, and the borrower's ability to repay. Entitlement is reduced while a prior VA loan is outstanding, and is restored on payoff or, in an assumption, only if a qualified veteran substitutes entitlement. See §B.7.

B.5 Ratios, reserves, and residual income

[all benchmarks illustrative — verify at the source in §B.1]

There is no single DTI cap (§4.5). The CFPB removed its own 43% General QM limit, and the number that actually binds a given file is the one in the finding or the investor matrix in front of you. Never tell a borrower they were declined "because of a number" without knowing whose number it was.

Residual income is VA's signature idea and it deserves more attention than it gets. VA asks how many dollars are left over each month after the mortgage, the other debts, taxes, and a maintenance and utilities allowance — measured against a table that varies by region and family size. A high-income borrower with a 44% back-end ratio may have thousands of dollars of residual income and be a perfectly sound file; a low-income borrower at 38% may fail. Ratios measure proportion; residual income measures survival. Chapter 17 works this in full.

Reserves, always stated in months of PITI (§A.9):

On the Linden Street file, reserves after closing were \$12,623.66 = 4.16 months of PITI — comfortable. After the day-46 furniture payoff they fell to \$7,423.66 = 2.45 months, which was still approvable but noticeably thinner. Nothing about the approval changed; the cushion did. On a jumbo file, that same payoff would have failed a 6-month reserve requirement outright.

B.6 Occupancy and property type

The government programs are owner-occupancy programs. That is not a technicality to be worked around; occupancy is certified in writing at closing, and misrepresenting it is loan fraud (Chapter 3). A borrower who intends to rent the property is a conventional or non-QM borrower, and telling them so early is a service, not a rejection.

Eligible property types [program condition standards and approval processes change — verify]:

⚠️ Property condition is a program question, not just an appraisal question. FHA and VA appraisals carry condition standards that a conventional appraisal does not, and a property that will not pass them is not an FHA or VA property no matter how well the borrower qualifies. On a fixer-upper, that fact often decides the program before any borrower fact does.

B.7 Assumability — the row that matters more every time rates rise

Why this row is worth memorizing. When market rates are well above the rate on an existing government loan, an assumable low-rate mortgage is a genuine asset attached to the house. Listing agents rarely know it. A loan officer who spots an assumable 3% VA loan on a listing has just found their client a payment no lender in the country can quote — and has also found the seller a problem, because that seller's entitlement will not come back until the loan is paid off or substituted. Both halves of that sentence are part of the advice.

B.8 Who is a candidate — and the one question that decides

On VA, one more sentence, because it is the most valuable question in this appendix. Ask about military service on every file, out loud, and ask it broadly: active duty, veterans, National Guard and Reserve members, and surviving spouses of service members. Borrowers routinely fail to mention six years in the Guard because they do not think it counts. On the Linden Street file, the loan officer asked on day 1 and the answer was no — which is why that file is conventional and why the VA option appears in the book only as a counterfactual. Asking cost ten seconds. Not asking, on the next file, costs a borrower their entire down payment.

B.9 The worked comparison — conventional 95% versus FHA 96.5%

The book's own decision, from the Linden Street file: \$385,000 purchase, primary residence, representative score 706, income \$10,500.00/month**, debts **\$1,446.00/month (§13.10, §5.8). Factors and rates are [illustrative — verify current pricing and HUD factors].

FHA is \$17.88/month cheaper, needs \$5,775.00 less down, and costs \$38,155.54 more in mortgage insurance.

⚠️ THAT IS THE MORTGAGE-INSURANCE DIFFERENCE, NOT THE TOTAL-COST DIFFERENCE. This is the easiest number in the book to repeat as the wrong quantity. FHA's lower rate returns \$17,446.29 of interest over the term, and its financed UFMIP **adds** \$4,672.94, so the two programs' totals do not differ by \$38,155.54:

Component	FHA vs. conventional
Mortgage insurance	+\$38,155.54
Interest — FHA's rate is 0.375% lower	−\$17,446.29
Upfront — UFMIP financed, less down payment	+\$4,672.94
NET TOTAL COST OF CREDIT	+\$25,382.19

Total cost of credit: \$503,396.01** conventional versus **\$528,778.20 FHA. Say "\$38,155.54 more in mortgage insurance"** or **"\$25,382.19 more in total cost." Never say "\$38,155.54 more" with nothing after it.

Why the file chose conventional. FHA was cheaper every month and easier on cash today, and it was still the wrong answer — because these borrowers had the \$19,250.00 and still kept 4.16 months of reserves. For them, FHA's relief was convenience, not feasibility, and paying \$25,382.19 for convenience is a bad trade. Change one fact — take away \$6,000 of savings — and FHA becomes the correct recommendation on the same file. That is what program selection is: not a ranking, a fit.

⚠️ A note on the MI totals. \$24,218.86 is 137 × \$176.78. \$62,374.40 is a constant-factor approximation across 360 payments; FHA's annual MIP actually recalculates on the declining balance, which is why it is not 360 × \$173.26. State the convention rather than silently multiplying.

B.10 Six questions, in order, and you have the program

Ask these in this sequence on every first call. It takes about four minutes and it eliminates almost every wrong recommendation a loan officer can make.

1. Has anyone in this household ever served in the military? Ask it out loud and ask it broadly — active duty, veterans, National Guard, Reserve, and surviving spouses. Borrowers do not volunteer this. If the answer is yes, VA is on the table before anything else, and it usually wins: no down payment, no monthly mortgage insurance, a funding fee that may be waived entirely. Never assume; never skip it because the borrower "doesn't seem like" a veteran.

2. How large is the loan compared with the local conforming limit? Below it: conventional, FHA, VA, and USDA are all live. Above it: you are in jumbo or high-cost high-balance territory, the pricing structure changes, and reserves become the binding condition. Look the current county limit up — do not recall it — and if the loan is near the line, discuss the down payment that crosses it.

3. Where is the property, and what is total household income? USDA is two binary tests: an eligible address and household income under the area limit. Both are lookups. Note that USDA counts the income of household members who are not on the loan — which surprises borrowers and is the most common reason a USDA pre-approval falls apart later. Check the map and the limit before you promise anything.

4. What is the representative credit score, and how much cash is there? Price conventional AND FHA. Never guess. Conventional MI is priced by score and LTV; FHA's annual MIP essentially is not. There is a crossover band where the answer flips, it moves with MI rate cards and market pricing, and no rule of thumb survives contact with an actual rate sheet. Run both, show both, and show what happens at 10% down if it is within reach — that is the line where FHA's MIP becomes an eleven-year charge instead of a permanent one.

5. How will they occupy it, and what kind of property is it? Second home or rental: FHA, VA, and USDA are out — conventional, jumbo, or non-QM. Condominium: is the project approved for the program you are recommending? Manufactured, 3–4 units, or a property in rough condition: the property may decide the program before the borrower does.

6. How is the income documented? W-2 with paystubs is an agency file. Self-employed with two years of returns that support the income is an agency file (Chapter 32). Self-employed with strong deposits and returns written down to nothing, an investor qualifying on rents, a borrower inside an agency waiting period after a credit event, or a borrower with assets but no income — that is where non-QM earns its price (Chapter 34). Ask this last, because it is the question that turns an apparent decline into a closed loan.

The discipline underneath all six. You are not selecting a program; you are finding the one constraint that actually binds. Cash, credit, income, geography, loan size, documentation — one of them is the real limit and the rest are noise. Find it, name it to the borrower, and the program chooses itself.