Chapter 13: Choosing the Right Loan: Program Selection, Rate vs. Points, ARM vs. Fixed, and Structuring the Deal

"Anyone can quote a rate. The job is to know which of six different loans this particular household should be signing, and to be able to show them why in numbers they can check." — constructed; the working premise of this chapter

Overview

You have spent five chapters gathering facts. Credit came back at a 706 representative score. Income documented to \$10,500.00 a month. Debts of \$1,446.00. Assets verified at \$38,000.00 — \$28,000 of savings and a \$10,000 gift. There is a contract on 4412 Linden Street at \$385,000.

None of that is a loan yet.

A loan is what happens when you take those facts and choose: which program, how much down, what term, where on the rate sheet, what kind of mortgage insurance, and whether anybody is buying the rate down. Six decisions. Change any one of them and the payment moves, the cash to close moves, the ratios move, and — this is the part nobody tells a new originator — the total the household pays for this money over the years they actually own the house can move by more than the down payment.

That set of six decisions is the loan structure, and choosing it is the most consequential thing a loan officer does. It is also the least supervised. An underwriter will check your arithmetic; nobody checks your judgment. If you put a borrower into a structure that qualifies, closes, and costs them twenty-five thousand dollars more than the alternative you did not run, no condition will ever fire, no audit will catch it, and the file will look perfect forever.

So this chapter does the thing the rest of Part II has been building toward. We are going to run the Linden Street file three ways, read an actual rate sheet, price the rate/point ladder in both directions, put an adjustable-rate mortgage next to a fixed one and show the worst case before the best case, cost out a builder's 2-1 buydown to the dollar, and then — the part that matters most — hand the decision back to the two people whose money it is.

Because that is the line. You do the analysis. They decide. Chapter 26 will explain the rules that exist because the incentive to decide for them is real and structural. This chapter is about the craft that makes those rules easy to follow: if your comparison is honest and your arithmetic is visible, handing back the decision costs you nothing, and it is the single most persuasive thing you will ever do at a kitchen table.

In this chapter, you will learn to:

• Name the six decisions that make up a loan structure, and say which of them the borrower owns
• Use four questions — cash, credit, property, horizon — to narrow forty programs to two
• Distinguish a down-payment option that is unavailable from one that is merely expensive
• Read a rate sheet's price grid and convert a price into dollars on a real loan amount
• Compute break-even on points and on lender credits, and know why they are not symmetric
• Compare an ARM to a fixed honestly, worst case first
• Cost a temporary buydown from its payment differences, and price the alternatives to it
• Present options, quantify the trade-off, recommend, and let the borrower choose

Learning Paths

🎓 Exam — §13.5, §13.6, and §13.7. Points arithmetic and the Ability-to-Repay qualifying rate for an ARM are reliably tested. §13.9's anti-steering material shows up in the ethics questions. 🏠 New LO — §13.3, §13.6, and §13.10. If you learn nothing else this chapter, learn to ask the horizon question before you quote a buydown, and learn the script. 🤝 Partner — §13.8 and §13.9. Agents and builders negotiate concessions constantly without knowing what the three uses of the same dollar are worth. This section is your listing-presentation material. 📊 Operations — §13.4 and §13.5. The price convention on a rate sheet, and what a "credit" actually obligates the file to deliver, are where lock desks and closing departments fight.

13.1 Structure is the job

Ask a room of first-year loan officers what they do, and most of them will describe a sales job with a lot of paperwork attached. Ask a fifteen-year originator and you will get something narrower: I figure out which loan these people should have, and then I make it survive contact with an underwriter.

The second half of that sentence is Parts III and IV of this book. The first half is this chapter.

Loan structure is the specific combination of choices that turns a purchase price and a household into one particular note. Six decisions, in roughly this order:

1. Program. Conventional, FHA, VA, USDA, or something outside agency lending. Chapter 5 mapped them; this chapter picks one.
2. Down payment. How much of their own money goes in, which sets the loan amount, the loan-to-value ratio, the mortgage insurance, and how much cash is left afterward.
3. Term. Thirty years, twenty, fifteen. Shorter costs less overall and qualifies fewer people.
4. Rate and point position. Where on the rate sheet you land — paying points to buy the rate down, sitting at par, or taking a lender credit and accepting a higher rate.
5. Mortgage insurance structure. Borrower-paid monthly, single premium, lender-paid, or the government's version — each with a different cancellation story.
6. Concessions and buydowns. Whether a seller or builder is contributing, and what that contribution is spent on.

Notice something about that list. Decisions 1, 2, and 3 mostly determine whether the loan is possible. Decisions 4, 5, and 6 mostly determine what it costs. New originators spend almost all their energy on the first group and almost none on the second, which is backwards, because the first group has few options and the second has many.

Notice something else. The borrower owns every one of these decisions. You own the analysis. That distinction sounds like a technicality until you have watched a loan officer describe a structure to a borrower using the word "we" for forty minutes and then act surprised when the borrower calls three days before closing to say their brother-in-law thinks they should have gone FHA. They never chose anything. They were told.

📞 On the Phone

Day 12. Credit is back, income is documented, assets are verified. The borrowers call.

Borrower: "So which loan is better — the FHA one or the regular one?"

The wrong answer: "Conventional. You'll save money." True on this file, and useless. They cannot check it, so they cannot believe it, so they will keep asking other people until somebody gives them a number.

The other wrong answer: "Well, it depends on a lot of factors." Also true. Also the answer of someone who has not run it.

What actually works: "Better at what? Give me until tomorrow afternoon and I'll send you both of them on one page — same house, same closing date, same everything except the program. Payment, cash you have to bring, what's left in your account afterward, and what each one costs you if you keep it. Then there's one question only you two can answer, and I'll ask it when we're on the phone together."

That last clause is not a stall. It is the horizon question in §13.6, and it genuinely decides this file. Asking it after they have seen both columns is deliberate: a borrower who has already read the comparison answers the horizon question honestly. A borrower who has not will answer it the way they think gets them the lower payment.

The rest of this chapter is what goes on that one page, and how you build it without deciding for them.

13.2 The four questions that pick the program

There are a great many mortgage programs. There are, for any given borrower, usually two that make sense and occasionally one. Getting from the first number to the second is a matter of four questions, plus a gate that comes before all of them.

Question Zero: is there an entitlement or an eligibility? Has either borrower served? Is the property in a USDA-eligible area and is household income under the limit? Chapter 5 covered both programs. The reason this is Question Zero rather than Question One is that a VA-eligible borrower buying a primary residence is usually finished deciding before the other four questions are asked — no down payment requirement and no monthly mortgage insurance is a combination the other programs cannot answer. Ask about military service on every single application, of every applicant, every time. Borrowers routinely do not volunteer four years in the Reserves because nobody asked and they do not think it counts.

Neither Linden Street borrower has served. The property is an inner-ring suburb, not a USDA-eligible area. Question Zero closes, and we proceed.

Question 1 — How much cash can they actually put in, and does the difference between programs buy convenience or feasibility?

This is the question this chapter exists to answer, and it is the one that most often gets answered lazily. FHA requires 3.5% down where conventional requires 5%. On a \$385,000 purchase that is a \$5,775.00 difference. The lazy version of the analysis is: they have less than twenty percent, so FHA. The honest version asks what the \$5,775.00 actually does.

There are only two possible answers, and the whole program decision turns on which one you are looking at:

• Feasibility. Without the relief, the borrower cannot close. There is no version of the transaction that works. The program that lowers the requirement is not a preference; it is the transaction.
• Convenience. The borrower can close either way. The relief changes their bank balance the day after closing and nothing else structural. It is worth something — sometimes a great deal — but it is now one number in a comparison rather than a gate.

The Linden Street borrowers have \$38,000.00 verified. At 5% down the total cash to close is \$25,376.34, which leaves **\$12,623.66 — 4.16 months of the full \$3,033.72 payment — sitting in the account the day after closing. They can close conventional. FHA's \$5,775.00 is convenience.

Hold that word. We will come back to it in §13.3 and it will decide the file.

Question 2 — What does the credit look like?

Not just the representative score. The score sets the pricing and the eligibility floor; the derogatory history sets the waiting periods. A 700 score three years after a foreclosure is a different file from a 700 score with no public records, and Chapters 14 and 16 take the seasoning requirements apart.

The rough shape, and it is only a shape — verify current requirements against the Fannie Mae Selling Guide, HUD Handbook 4000.1, and your own lender's overlays, because all three change:

That last row is the one that decides program selection more often than any other, and it is worth saying plainly: conventional mortgage insurance is priced off the borrower's credit; FHA's is not. As a score falls, conventional MI gets more expensive quickly while FHA's mortgage insurance premium does not move at all. Somewhere on that curve the two programs cross. Above the crossing point conventional wins; below it FHA wins, sometimes by a lot.

A 706 is comfortably above the crossing point in most pricing environments. A 641 is not.

Question 3 — What is the property, and how will they occupy it?

Single-family, condominium, two-to-four units, manufactured. Primary residence, second home, investment. These are eligibility gates before they are pricing questions: some programs will not finance some properties at all, condominium projects have to be approved, and multi-unit properties carry their own down payment and reserve requirements. Chapters 5 and 14 hold the detail.

Linden Street is a 1994 single-family detached home, primary residence. This question is a formality here, which is exactly why you should be suspicious of how easy it feels — a condominium in a project that has not been approved is a two-week delay that surfaces on day thirty if nobody asked on day one.

Question 4 — How long will they keep this loan?

The question nobody asks. It gets its own section (§13.6) because it decides more of the structure than the other three combined and because there is no way to get it from a document. You have to ask a human being, and you have to ask in a way that gets a real answer rather than the answer they think you want.

PROGRAM SELECTION — the decision tree, in order        [constructed teaching example]

   QUESTION ZERO ─── VA eligible? ──── yes ──► VA is almost certainly the answer.
                          │                     Stop. Verify the COE.
                          │ no
                          ▼
                   USDA-eligible area
                   AND income under cap? ─ yes ─► Run USDA against conventional.
                          │ no
                          ▼
   Q1  CASH ──── Is the down-payment difference FEASIBILITY or CONVENIENCE?
                          │
              ┌───────────┴────────────┐
        FEASIBILITY                CONVENIENCE
              │                         │
              ▼                         ▼
      The lower-down program      Both programs are live.
      IS the transaction.         Move to Q2 and price them.
              │                         │
              ▼                         ▼
   Q2  CREDIT ─── score above the MI crossing point? ── no ──► FHA likely
                          │ yes
                          ▼
   Q3  PROPERTY ── eligible for both? ── no ──► whichever accepts it
                          │ yes
                          ▼
   Q4  HORIZON ── short (under ~5 years)? ── yes ──► weight the monthly payment
                          │                          and the cash at closing
                          │ long
                          ▼
              Weight the TOTAL COST and the MI cancellation terms.

Read that tree from the top and notice what it does not do: it never picks a program from the down payment alone. That is the single most common structural error in origination, and the Harlow Street file is the reason it survives — because sometimes the lazy answer is also the right one, and the originator who got it right by accident never learns the difference.

The Harlow Street file: when the relief is the transaction

A single borrower. One income of \$4,150.00** a month, **\$395.00 in monthly debts, a 641 representative score, first-time buyer, buying a townhome at \$215,000** with a **\$10,000 forgivable county down-payment assistance second.

Run the four questions.

Q1 — cash. The DPA covers the 3.5% FHA down payment of \$7,525.00 and leaves \$2,475.00 toward closing costs. Without it, and without FHA's lower requirement, there is no down payment at all. This is not a bank balance that closes either way. This is feasibility.

Q2 — credit. A 641 sits below the score floor most conventional lenders apply once overlays are layered on, and well below the point where conventional mortgage insurance prices competitively. FHA's mortgage insurance does not care about the 641 at all.

Q3 — property. A townhome, primary residence, eligible.

Q4 — horizon. A first-time buyer at the beginning of an earning curve, in a starter property. The honest answer is probably not thirty years.

Every question points the same direction. For this borrower FHA is unambiguously right, and the loan closes at a base loan amount of \$207,475.00 plus \$3,630.81 of financed upfront mortgage insurance premium — a total loan of \$211,105.81 at 6.250%, P&I of \$1,299.81, annual MIP of \$96.76 a month, taxes of \$215.00 and insurance of \$110.00, for a total housing payment of \$1,721.57 and ratios of 41.48% front / 51.00% back. (All figures constructed for this file; verify current MIP factors and county DPA terms at their sources.)

Now hold the two files next to each other. Same program on one side of the desk and not the other. Same country, same year, same rate sheet. The programs did not change. The borrowers did. Program selection is a statement about a household, never about a product, and any originator who has a favorite program has stopped doing the analysis.

13.3 Down payment: the 5/10/20 analysis

Every purchase borrower who has ever read a personal-finance article arrives believing they should put twenty percent down. Every purchase borrower who has talked to a friend arrives believing they should put as little down as possible. Both beliefs are conclusions without arithmetic. The arithmetic is this: run the same house at three down payments and look at what each one buys and what each one costs.

On Linden Street, holding the rate at 6.625% for comparability:

THE 5/10/20 LADDER — same house, same rate, three down payments
                                                       [the Linden Street file]
                        5% DOWN         10% DOWN        20% DOWN
  ─────────────────────────────────────────────────────────────────
  Down payment          $19,250.00      $38,500.00      $77,000.00
  Loan amount          $365,750.00     $346,500.00     $308,000.00
  LTV                       95.00%          90.00%          80.00%
  ─────────────────────────────────────────────────────────────────
  P&I                    $2,341.94       $2,218.68       $1,972.16
  Taxes                    $385.00         $385.00         $385.00
  Insurance                $130.00         $130.00         $130.00
  MI factor                  0.58%           0.32%            none
  MI                       $176.78          $92.40           $0.00
  ─────────────────────────────────────────────────────────────────
  TOTAL PAYMENT          $3,033.72       $2,826.08       $2,487.16
  Housing ratio             28.89%          26.92%          23.69%
  Back-end DTI              42.66%          40.69%          37.46%
  ─────────────────────────────────────────────────────────────────
  Saving vs. 5%                  —         $207.64         $546.56
  Extra cash required            —      $19,250.00      $57,750.00
  Months to pay back             —            92.7           105.7
  ─────────────────────────────────────────────────────────────────
  Verified funds available: $38,000.00
  MI factors illustrative; verify current rate cards at the source.

Three things fall out of that ladder, and only one of them is the one borrowers expect.

First, the monthly saving is real and the payback is long. Moving from 5% to 10% saves \$207.64 a month and costs \$19,250.00 more at the closing table. That is a 92.7-month payback — seven years and nine months before the extra cash has come back as payment savings. Moving to 20% saves \$546.56 a month for \$57,750.00 more, a 105.7-month payback. Neither of those is a bad deal. Neither is an obvious one, either, and a borrower who is going to need that money for a roof in year four should hear the number.

Second, most of the 10% saving is not what borrowers think it is. Look at where the \$207.64 comes from:

• P&I falls \$123.26, from \$2,341.94 to \$2,218.68 — that is the smaller loan.
• MI falls \$84.38, from \$176.78 to \$92.40 — that is the factor dropping from 0.58% to 0.32%.

\$84.38 of \$207.64 is 40.6% of the saving, and it has nothing to do with borrowing less money. It is the mortgage insurance company charging less because a 90% loan is a materially different risk from a 95% loan. This matters because borrowers reason about down payments as "paying down the loan," and that intuition undercounts the benefit of crossing an LTV threshold by roughly two-fifths.

Third — and this is the whole section — 10% down is not available.

🧮 Run the Numbers

Why Option C is not an option.

Ten percent of \$385,000 is **\$38,500.00**.

Total verified funds are \$38,000.00.

$$\$38{,}500.00 - \$38{,}000.00 = \$500.00 \text{ short}$$

That is before a single dollar of closing costs, prepaid taxes, prepaid insurance, or per-diem interest. On the 5%-down structure, the total cash to close is \$25,376.34, of which \$19,250.00 is the down payment — leaving \$6,126.34 for everything else Chapter 12 itemized.

Add roughly that same six thousand to a \$38,500.00 down payment and the borrowers need approximately \$44,600** against \$38,000.00 on hand. They are short by roughly \$6,600, and if they somehow found it they would close with zero reserves** on a payment 64% larger than their current rent.

Note what just happened. Option C has the best payment, the best ratios, and the cheapest mortgage insurance in the ladder — and it is not a choice. It is a fact about a bank statement. A structure the borrower cannot fund is not the wrong structure; it is not a structure at all, and presenting it as an option they "decided against" is a small dishonesty that will cost you the next conversation.

Run this test on every file before you build a comparison: can they actually fund each column? Delete the ones they cannot, and say why you deleted them.

Twenty percent is worse: \$77,000.00 plus costs against \$38,000.00, short by roughly \$45,100. It appears in the ladder for one reason — so that when the borrowers say "everybody says twenty percent," you have already run it and can show them exactly how far away it is instead of arguing.

So the ladder collapses. There is one down payment on this file, and it is 5%, and the only remaining question is which program's 5%-or-less structure they should use. Which returns us to the word from §13.2.

\$5,775.00 of FHA down-payment relief on this file is convenience. They close either way. What the \$5,775.00 buys is a bigger cushion afterward: holding all other closing costs equal, reserves would rise from \$12,623.66 to **\$18,398.66, which against the FHA payment of \$3,015.84 is 6.10 months** rather than 4.16. That is not nothing. Six months of reserves is meaningfully more comfortable than four for a household whose housing payment is about to increase 64%, and any loan officer who waves that away is selling rather than advising.

But it is a comparison now, not a gate. And there is a second number attached to it that the borrowers have not seen yet.

FHA hands them \$5,775.00 at the closing table and charges them \$12,276.69 more of debt.

\$371,525.00 base loan + \$6,501.69 of financed upfront MIP = \$378,026.69, against \$365,750.00 conventional. The difference is \$12,276.69 — which is exactly the \$5,775.00 of down-payment relief plus the \$6,501.69 upfront premium that paid for it.

That sentence is the most useful thing in this chapter, because it is the general form. Every down-payment relief in mortgage lending is purchased with something. Sometimes with an upfront premium, sometimes with a higher rate, sometimes with mortgage insurance that never cancels, sometimes with a repayable second lien. The originator's job is to find the price tag, not to be impressed by the discount.

13.4 Reading a rate sheet

Up to this point the rate has been a fact handed to you. It is time to open the document it comes from.

A rate sheet is a menu. It shows, for one program and one lock period, a ladder of interest rates and the price attached to each. It is republished at least once every business morning and frequently intraday. Where the numbers on it come from — the mortgage-backed securities market, the base price, the loan-level adjustments that move a borrower off that base — is Chapter 29's, and Chapter 29 rebuilds this exact quote from the ground up. When you may lock, what a lock costs, and what happens when one expires is Chapter 30's. Right now we are doing something narrower and more immediately useful: reading one.

📄 Read the File

```text FIGURE 13.1 — "The grid, after adjustments" [constructed teaching grid — verify current pricing at the source] THE DOCUMENT Pricing engine output, 30-year fixed conventional, run at 9:12 a.m. on day 14. Four inputs: representative score 706, LTV 95.00%, single-family primary residence, 30-day lock. The engine has already applied the adjustments; this is what this file prices at, not what the program's headline rate is. THE CONTEXT $365,750 loan on a $385,000 purchase. The borrowers are shopping and have seen a lower number advertised by an online lender. WHAT IT SHOWS

  LOCK PERIOD:    15 DAY | 30 DAY | 45 DAY | *60 DAY*
  (pricing worsens as the lock lengthens; the 60-day column is this file's)

  RATE      PRICE      POINTS       COST / (CREDIT)      P&I
  ────────────────────────────────────────────────────────────────
  7.000%   100.750     -0.750         ($2,743.12)     $2,433.34
  6.875%   100.375     -0.375         ($1,371.56)     $2,402.72
  6.750%   100.000      0.000 par           $0.00     $2,372.25
  6.625%    99.500     +0.500          $1,828.75      $2,341.94
  6.500%    99.000     +1.000          $3,657.50      $2,311.79
  6.375%    98.375     +1.625          $5,943.44      $2,281.80
  ────────────────────────────────────────────────────────────────
  Price 100.000 = par. Above 100 = credit to the borrower.
  Below 100 = cost to the borrower. One point = 1% of $365,750 = $3,657.50.

WHAT IT DOESN'T It does not show the borrower's rate — it shows six of them, and which one is theirs is a decision nobody has made yet. It does not show the lender's origination charge, title, appraisal, taxes, insurance, or prepaids, so no line on it is a cost to close. It does not show mortgage insurance, which is priced off a separate card. It is not a lock, and it will not be these numbers tomorrow. THE DECISION Do not pick a row yet. Compute the break-even on each row against par (§13.5), then ask the horizon question (§13.6), then let the borrowers pick. THE LESSON A rate sheet does not contain a rate. It contains a trade, and the trade is always the same one: dollars today against dollars per month for as long as they keep the loan. The number in the newspaper is one row of somebody's grid, for somebody else's file. ```

Four conventions on that sheet are worth learning cold, because they are near-universal and because misreading any of them produces a quote that is wrong by thousands of dollars.

Price is quoted per hundred. 100.000 is par — Chapter 4 defined it — meaning the loan is worth exactly its face amount and no points change hands. 99.500 means the loan is worth 99.5% of face, so somebody has to make up the missing half percent, and that somebody is the borrower: 0.500 points, or \$1,828.75. 100.750 means the loan is worth more than face, and the excess comes back the other way as a lender credit of \$2,743.12.

Points and price move in opposite directions. A higher price is better for the borrower and comes with a higher rate. This trips up every new originator at least once, usually out loud in front of a customer. The mnemonic that works: the investor is buying an income stream, and a bigger income stream is worth more money.

Cost and credit are shown in points; the borrower experiences dollars. Always convert on the actual loan amount before you say anything to anyone. "A quarter point" is an abstraction. "Nine hundred fourteen dollars" is a conversation.

Parentheses mean credit. Every sheet in the business uses parentheses or a negative sign for money flowing toward the borrower. Read them carefully at 7:40 in the morning.

The last convention worth naming is the one that does not appear on the sheet at all: the sheet is priced for a set of file characteristics, and the pricing engine has already applied them. Figure 13.1 is not "today's rates." It is today's rates for a 706 score at 95% LTV on a single-family primary residence with a 30-day lock. Change any one of those four and the entire grid shifts. This is why Chapter 1 insisted you never quote without all four facts, and it is why the number the borrowers saw advertised is not a lie and is also not theirs.

13.5 Points, par, and lender credits

Now work the grid.

There are three positions a borrower can take relative to par, and each is a different transaction.

Pay points and take a lower rate. The borrower hands over money at closing in exchange for a smaller payment every month. This is a permanent buydown — permanent because the note rate itself is lower for the entire life of the loan, as distinct from the temporary version in §13.8, where the note rate never changes at all. It is a purchase, and like any purchase it has a payback period.

Sit at par. No points paid, no credit received, note rate is whatever the sheet says par is. The default, and more often correct than the industry's enthusiasm for points suggests.

Take a lender credit and accept a higher rate. Chapter 4 defined the lender credit; here is what it is for. Money flows toward the borrower at closing, offsetting closing costs, and they pay for it with a higher payment forever. It is a loan against their own future payments, and it is frequently the right answer for a borrower who is short on cash — which describes a great many first-time buyers.

The arithmetic is identical in both directions. Chapter 4 gave you the formula: cost of points divided by monthly payment saved, in months. Apply it to every row of Figure 13.1 against par.

🧮 Run the Numbers

The whole ladder, both directions. All against par at 6.750%, P&I \$2,372.25, on \$365,750.

Rate	Cost / (credit)	P&I	vs. par, monthly	Break-even
7.000%	(\$2,743.12) | \$2,433.34	+\$61.09	credit repaid in 44.9 months
6.875%	(\$1,371.56) | \$2,402.72	+\$30.47	credit repaid in 45.0 months
6.750%	\$0.00 | \$2,372.25	—	par
6.625%	\$1,828.75 | \$2,341.94	−\$30.31	recovered in 60.3 months
6.500%	\$3,657.50 | \$2,311.79	−\$60.46	recovered in 60.5 months
6.375%	\$5,943.44 | \$2,281.80	−\$90.45	recovered in 65.7 months

Work one line so the rest are checkable. At 6.625% the borrower pays \$1,828.75 and saves \$2,372.25 − \$2,341.94 = \$30.31 a month:

$$\frac{\$1{,}828.75}{\$30.31} = 60.3 \text{ months} = 5.0 \text{ years}$$

And the credit side, at 6.875%: the lender hands over \$1,371.56 and the borrower pays \$2,402.72 − \$2,372.25 = \$30.47 a month more:

$$\frac{\$1{,}371.56}{\$30.47} = 45.0 \text{ months}$$

Now look at the asymmetry, because it is the most useful thing on this page. Buying down from par breaks even at about sixty months. Taking credit from par breaks even at about forty-five months. Those are not the same number, and the gap is not an accident.

Buying the first eighth of a point of rate costs 0.500 points. Selling the first eighth of a point of rate pays 0.375 points. The 0.125-point spread is the market's, and it means the borrower is on the wrong side of a bid-ask every time they move off par in either direction.

The practical consequence: the case for taking a lender credit is stronger, at any given horizon, than the case for buying points — because the credit pays back in three-quarters of the time. A borrower who is unsure how long they will stay and is short on cash has an easy answer. A borrower who is unsure how long they will stay and has cash does not, and should probably sit at par.

Three warnings before you use any of this on a live file.

A lender credit has to be spent on something. It offsets closing costs and prepaids; it cannot be taken as cash back to the borrower on a purchase, and if the credit exceeds the costs available to absorb it, the excess is simply lost or must be repriced. Before you promise a borrower a \$2,743.12 credit, know what their actual closing costs are. Chapter 22 covers how the credit is disclosed and what happens when it changes.

Points paid must be genuinely for the rate. A discount point is 1% of the loan amount paid to reduce the rate — that is Chapter 4's definition and it is also a legal characterization with consequences under Regulation Z's points-and-fees treatment. A charge labeled "discount" that does not buy a rate reduction is not a discount point. Chapter 24 is the authority.

Break-even is not the same as "worth it." Sixty months is a fact. Whether sixty months is short or long is a judgment about the borrower's life, and you do not have that information. Which is the subject of the next section, and the reason it exists.

13.6 Break-even, and the horizon question nobody asks

Every number in §13.5 is a fraction with the same denominator problem. Break-even tells you how many months it takes to recover a cost. It does not tell you whether the borrower will be there.

The horizon — how long this household will actually hold this loan — is the single most decisive input in loan structuring, and it is the only one you cannot verify. Income comes from a paystub. Assets come from a bank statement. Credit comes from a bureau. The horizon comes from a conversation, and most loan officers never have it.

Consider what the horizon does to decisions already on the table:

Every row flips. Same borrower, same house, same rate sheet — and the correct structure inverts on a fact you have to ask for.

How to ask it

Badly asked, the horizon question gets a useless answer. "How long do you plan to stay?" invites "forever," because that is what people say about a house they are buying on Saturday.

Ask it three ways instead, in this order:

1. "Is this the house, or is this the first house?" Borrowers answer this one honestly, because it is not about the loan. A three-bedroom in an established suburb bought by two people with a young family is usually an answer.
2. "Is there anything in the next five years that could move you — a transfer, a promotion that relocates, a parent you might need to be closer to, a program either of you might go back to school for?" Specific, finite, and it surfaces the transfer clause in an employment contract that nobody thinks to mention.
3. "If rates dropped a point next year, would you refinance?" This one is about the loan, and it is the question that decides the points. Almost everyone says yes. Say the follow-up out loud: "That's the honest answer and it's the right answer — and it means the money we spend today buying the rate down might only have two years to work. Let's price it both ways."

On Linden Street the answers are clear. Both borrowers have stable local employment — three years at the same regional hospital, four years with the same company — in a metro they are from. They are buying a three-bedroom, two-bath family home in an established inner-ring suburb, not a starter condominium. Neither has any plan to move and neither has a relocation clause.

And to the third question they say yes, they would refinance if rates fell substantially.

That is not a footnote. It is the fact that weakens everything this chapter is about to recommend, and you must say so out loud rather than filing it where the borrower cannot see it.

⚠️ Where Deals Die

Selling a buydown to a borrower who is going to refinance.

This one does not kill the transaction. It closes, funds, and looks perfect. It kills the relationship eighteen months later when rates fall, the borrower refinances, and somebody points out that the \$1,828.75 they paid to buy the rate down went with the old loan.

Quantify the exposure before you recommend anything. The point recovers at \$30.31 a month:

If they refinance at month	Recovered	Unrecovered
12	\$363.72 | **\$1,465.03**
24	\$727.44 | **\$1,101.31**
36	\$1,091.16 | **\$737.59**
48	\$1,454.88 | **\$373.87**
60	\$1,818.60 | **\$10.15**
137 (MI termination)	\$4,152.47 | recovered, +\$2,323.72 ahead
360	\$10,911.60 | recovered, +\$9,082.85 ahead

The worst case is \$1,828.75 — they refinance next month and recover nothing. The realistic downside at a three-year refinance is \$737.59. The upside if they hold to the mortgage insurance termination point is \$2,323.72** net, and to term, **\$9,082.85 net.

That is the entire decision, and it is a decision about their life, not about mortgages. Show them the table. Do not summarize it. A borrower who has read this table and chooses the point has made an informed decision that will survive a rate drop; a borrower who was told "points are a good idea" has not.

(Recovered = \$30.31 × months. Unrecovered = \$1,828.75 − recovered. Net ahead = recovered − \$1,828.75. Figures ignore the time value of money, which makes the point look slightly better than it is — say so.)

The Linden Street answer on points

Their horizon exceeds sixty months. They intend to keep the house, they have no relocation exposure, and the break-even on the half point at 6.625% is 60.3 months.

So the half point is taken. \$1,828.75 at closing, \$30.31 a month back, whole thing recovered five years and one month in, and every month after that is profit for as long as they hold the loan.

Had they answered the horizon question differently — five years or less, a possible transfer, a starter home they expected to outgrow — par at 6.750% is the right answer, and this chapter would end with a \$2,372.25 P&I and \$1,828.75 still in their savings account. Not a close call in that direction either. The arithmetic did not change; the borrower did.

Notice, too, that the refinance answer cuts both ways, and honesty requires saying so:

• It weakens the case for the point, for exactly the reason the table above shows.
• It weakens the conventional case's biggest number, because the mortgage insurance advantage in §13.3 only fully materializes if the loan survives to payment 137.
• It strengthens the conventional case in one specific way: a borrower who refinances out of an FHA loan into a conventional one carries the \$6,501.69 of financed upfront mortgage insurance premium with them as unpaid principal. FHA provides a partial refund of upfront MIP only on an FHA-to-FHA refinance within a limited window and on a declining schedule — verify the current schedule in HUD Handbook 4000.1 — and a conventional refinance gets nothing back. Conventional has no equivalent sunk one-time charge.

Three effects, two directions. Write them down for the borrower in exactly that form. A comparison that only lists the facts supporting your recommendation is not a comparison.

🔍 Check Your Understanding

1. A borrower is offered a \$4,200 lender credit in exchange for a rate 0.375% higher, which raises the payment by \$92.00. How many months until the credit is repaid, and which kind of borrower should take it?
2. Why does buying down from par on Figure 13.1 break even at about 60 months while taking credit from par breaks even at about 45?
3. A borrower says "we'll be here forever" and, ninety seconds later, "we'd definitely refinance if rates drop." Are those answers in conflict? What does each one govern?

(1: \$4,200 ÷ \$92.00 = 45.7 months; a borrower short on cash or with a horizon under about four years. 2: it costs 0.500 points to buy the first eighth and pays only 0.375 to sell it — the spread is the market's. 3: not in conflict at all. The first governs the program and the mortgage insurance comparison; the second governs the points. A borrower can hold a house for thirty years and hold four different loans on it.)

13.7 ARM vs. fixed, honestly

Chapter 5 built the adjustable-rate mortgage: the index, the margin, the fully indexed rate, and the caps. This section does one thing with it — puts it next to the fixed-rate loan on this file and prices the risk in dollars.

And it does so in a specific order. Worst case first.

That ordering is not a stylistic preference. An ARM presented best-case-first is a sales pitch with a disclosure attached, and every originator who lived through 2006 and 2007 watched what that produced. Present the ceiling, then the floor, then let the borrower look at the gap.

Here is Chapter 5's illustration on Linden Street's \$365,750:

THE ARM PAYMENT PATH — 5/6 ARM, $365,750, 30-year term
                                                       [the Linden Street file]
  Initial rate 5.875% fixed for 60 months, then adjusting every 6 months.
  Index 4.25% + margin 2.75% = FULLY INDEXED RATE 7.00%.  Caps 2/1/5.

  $3,500 ┤                                    ████  $3,448.62   LIFETIME CAP 10.875%
         │                                    ████               (+5.000 over initial)
  $3,000 ┤                    ████            ████  $2,651.94   MAX 1st ADJ 7.875%
         │                    ████            ████               (+2.000 at month 61)
         │          ████      ████            ████  $2,433.34   QUALIFYING 7.00%
  $2,500 ┤          ████      ████            ████               (fully indexed)
         │  ████    ████      ████            ████  $2,341.94   the FIXED loan, 6.625%
         │  ████    ████      ████            ████  $2,163.55   INITIAL 5.875%
  $2,000 ┤  ████    ████      ████            ████
         └──────────────────────────────────────────
          months   qualifying  month 61      worst
           1-60      payment    ceiling       case

  WORST CASE IS $1,285.07 PER MONTH ABOVE THE INITIAL PAYMENT.
  Not to scale. Illustrative index value; the index moves and so does the fully
  indexed rate at every adjustment.

Read the ceiling first. The lifetime cap is 5.000 percentage points over the initial rate, which is 10.875%, which on this loan is \$3,448.62** of principal and interest. That is **\$1,285.07 a month more than the initial payment and \$1,106.68 more than the fixed-rate payment they can have today. Whether that scenario is likely is a question nobody can answer; whether it is permitted by the note they would be signing is not in doubt at all.

Then read the first adjustment. The first-adjustment cap is 2.000 percentage points, so at month 61 the rate can go to 7.875% and the payment to \$2,651.94** — an increase of **\$488.39 in a single month, and \$310.00 a month more than the fixed loan.

Only now look at the initial payment: \$2,163.55**, which is **\$178.39 a month less than the fixed-rate \$2,341.94. Over the sixty-month fixed period that is:

$$\$178.39 \times 60 = \$10{,}703.40$$

Ten thousand seven hundred dollars of real savings, against a payment that can rise \$488.39 in one month at the end of it and \$1,285.07 over the life. That is the trade, stated fairly. For some borrowers it is a good trade — a household with a certain five-year horizon, or with income that is rising in a documented way, or with the liquidity to absorb the ceiling.

Now the part that decides it on this file.

🎓 NMLS Exam Watch

The Ability-to-Repay rule requires the creditor to qualify an adjustable-rate borrower at the greater of the fully indexed rate or the introductory rate, using a fully amortizing payment schedule. On this file the fully indexed rate is 4.25% + 2.75% = 7.00%, and the introductory rate is 5.875%. The greater is 7.00%, so the qualifying P&I is **\$2,433.34** — not the \$2,163.55 the borrower would actually pay.

Run the ratios at the qualifying payment, using the same taxes, insurance, and mortgage insurance:

	Fixed 6.625%	5/6 ARM, qualified
P&I used to qualify	\$2,341.94 | **\$2,433.34**
Taxes + insurance	\$515.00 | \$515.00
MI	\$176.78 | \$176.78
Qualifying payment	\$3,033.72 | **\$3,125.12**
Housing ratio	28.89%	29.76%
Back-end DTI	42.66%	43.53%

The ARM's low initial payment does not help them qualify. It hurts. The ARM qualifies them at a higher debt-to-income ratio than the fixed loan, because the qualifying rate is 7.00% and the fixed rate is 6.625%.

This is the trap the exam builds its ARM questions around, and it is also the trap a borrower walks into on their own: "we'll take the ARM to afford more house." Under Ability-to-Repay, on a standard ARM, you generally cannot. The introductory rate buys cash flow, not qualifying power.

(ARM mortgage insurance factors are typically higher than the fixed-rate equivalent; \$176.78 is held constant above only to isolate the effect of the qualifying rate. Verify current ATR/QM requirements and MI rate cards at the source — requirements change and state law varies.)

So on Linden Street the ARM is declined on its own terms, and the reasoning is short enough to say in one breath: it does not improve qualification, the borrowers' horizon is long, they are not liquidity-rich, and the ceiling is \$1,106.68 a month above a fixed payment they can have today. The fixed loan wins for this household.

That is not a general verdict on ARMs, and you should be careful not to deliver it as one. Chapter 5 was explicit that the adjustable-rate mortgage is a legitimate instrument that was misused, not a defective one. The borrower with a documented three-year assignment, the borrower whose income steps up on a contractual schedule, the borrower buying a bridge property they will sell — for these, an ARM's five years of \$178.39 a month is a real and appropriate saving. Present the worst case, price the horizon, and let them decide. The rule is the same rule as everywhere else in this chapter.

13.8 Temporary buydowns and who is paying for them

A temporary buydown is not a rate. It is an escrow account.

Here is the mechanic, and it is worth being precise because almost every borrower and a startling number of agents have it wrong. The note rate does not change. The borrower signs a note at 6.625% for thirty years, exactly as they would have. Alongside it, somebody deposits a sum of money into an escrow account at closing. Each month for a stated period, the servicer draws from that account to make up the difference between the payment at the note rate and a lower, subsidized payment. When the account is exhausted, the subsidy stops, and the borrower pays the note rate they agreed to on day one.

A 2-1 buydown is the common form: the effective rate is reduced by 2 percentage points in year one and 1 percentage point in year two, then the note rate applies for the remaining twenty-eight years. On a 6.625% note that is 4.625% in year one and 5.625% in year two.

Two facts follow immediately, and they are the two facts that matter.

The funds are escrowed up front by whoever pays. Most often a seller or a builder, as a concession. Sometimes a lender. The money is real, it is delivered at closing, and it is a cost to somebody on day one — not spread out, not conditional.

The borrower is qualified at the note rate, not the bought-down rate. The subsidy is temporary and the obligation is not, so the underwriter runs the ratios at the payment the borrower will be making in year three. A 2-1 buydown does not qualify anyone for anything they could not otherwise qualify for. Anyone who tells a borrower otherwise is describing a product that does not exist. (Verify current requirements against the applicable agency guide and your lender's overlays; buydown eligibility, permitted sources of funds, and qualifying treatment vary by program and change.)

Now cost it.

🧮 Run the Numbers

What a 2-1 buydown on the Linden Street loan actually costs, to the dollar.

Note rate 6.625% on \$365,750, 30-year fixed. The subsidized rates are 2 points and 1 point below:

Period	Effective rate	P&I paid by borrower	Note-rate P&I	Monthly subsidy
Months 1–12	4.625%	\$1,880.47 | \$2,341.94	\$461.47
Months 13–24	5.625%	\$2,105.46 | \$2,341.94	\$236.48
Months 25–360	6.625%	\$2,341.94 | \$2,341.94	\$0.00

Year one: $\$2{,}341.94 - \$1{,}880.47 = \$461.47$ per month.

$$\$461.47 \times 12 = \$5{,}537.64$$

Year two: $\$2{,}341.94 - \$2{,}105.46 = \$236.48$ per month.

$$\$236.48 \times 12 = \$2{,}837.76$$

Total escrowed at closing:

$$\$5{,}537.64 + \$2{,}837.76 = \boxed{\$8{,}375.40}$$

Sanity-check it as a percentage, because that is how a seller will think about it:

$$\frac{\$8{,}375.40}{\$365{,}750} = 2.290\% \text{ of the loan} > \qquad \frac{\$8{,}375.40}{\$385{,}000} = 2.175\% \text{ of the price}$$

Eight thousand three hundred seventy-five dollars and forty cents is a very large concession — larger than the 1.625 points (\$5,943.44) that would buy the rate down to 6.375% permanently on Figure 13.1. That comparison is the reason this section exists, and no borrower ever gets shown it.

Note what the buydown does to the borrower's own payment path. The full housing payment — P&I plus taxes, insurance, and mortgage insurance — moves from \$2,572.25** in year one to **\$2,797.24 in year two to **\$3,033.72** in year three, where it stays. Against current rent of \$1,850.00, the payment shock arrives in three stages rather than one, which is genuinely valuable to some households and genuinely dangerous to others: a borrower who budgets around year one's number has built a household on a payment that is contractually scheduled to rise \$461.47 within twenty-four months.

(\$1,880.47 + \$385.00 + \$130.00 + \$176.78 = \$2,572.25; \$2,105.46 + \$691.78 = \$2,797.24; \$2,341.94 + \$691.78 = \$3,033.72.)

The honest framing: three uses of the same dollar

Here is the part that separates an originator from an order-taker. If a seller or builder is willing to contribute \$8,375.40, a 2-1 buydown is one of at least three things that money can do — and it is the one the borrower will be offered, because it produces the most dramatic-looking first-year payment and because it is what the builder's marketing department has already printed.

Show all three.

WHERE $8,375.40 OF CONCESSION CAN GO             [constructed teaching comparison —
                                                  verify pricing at the source]

  (A) 2-1 TEMPORARY BUYDOWN — note rate stays 6.625%
      ────────────────────────────────────────────────────────────────
      Year 1 P&I     $1,880.47   (saves $461.47/mo)  ─┐
      Year 2 P&I     $2,105.46   (saves $236.48/mo)  ─┤ $8,375.40 of relief,
      Year 3+ P&I    $2,341.94   (saves $0)          ─┘ all of it inside 24 months
      Total relief delivered: $8,375.40 over 24 months, then nothing.

  (B) PERMANENT BUYDOWN — buy the note rate down instead
      ────────────────────────────────────────────────────────────────
      6.375% costs 1.625 points = $5,943.44 (vs. $1,828.75 already
      committed at 6.625%): incremental cost $4,114.69
      P&I falls from $2,341.94 to $2,281.80 = $60.14/mo, FOREVER
      Break-even on the incremental cost: $4,114.69 / $60.14 = 68.4 months
      Relief over 24 months:  $60.14 x 24  =   $1,443.36
      Relief over 360 months: $60.14 x 360 =  $21,650.40
      And $2,431.96 of the concession is left over for closing costs.

  (C) PRICE REDUCTION — take $8,375.40 off the contract price
      ────────────────────────────────────────────────────────────────
      Price $385,000.00 -> $376,624.60
      5% down $19,250.00 -> $18,831.23   (frees $418.77 of cash)
      Loan $365,750.00 -> $357,793.37
      P&I $2,341.94 -> $2,290.99         (saves $50.95/mo)
      MI  $176.78   -> $172.93           (saves  $3.85/mo)
      Total monthly relief: $54.80, FOREVER, plus $418.77 back at closing.
      Relief over 24 months:  $54.80 x 24  =   $1,315.20
      Relief over 360 months: $54.80 x 360 =  $19,728.00

  ────────────────────────────────────────────────────────────────────
  SHORT HORIZON  -> (A) delivers six times the relief of (B) or (C) in
                    the first two years.
  LONG HORIZON   -> (B) and (C) deliver two and a half times (A)'s total.
  ────────────────────────────────────────────────────────────────────
  Requires seller agreement; (C) additionally requires a contract
  amendment and may affect the appraisal and the assessed value.

None of those three is the right answer in general. Column A is genuinely better for a household whose income is documented to rise, or who is furnishing an empty house, or who expects to refinance inside two years — it delivers \$8,375.40 of relief in twenty-four months, which is roughly six times what either alternative delivers in the same period. Columns B and C are better for the household that stays, because their relief never stops.

The failure mode is not choosing A. The failure mode is never having priced B and C. A borrower who is handed a builder's 2-1 buydown as though it were the only available shape of a concession has been sold something rather than advised, and if their horizon is long they have quietly given up roughly thirteen thousand dollars of relief over the life of the loan.

Two further cautions, because they come up constantly:

Seller contributions are capped. Every program limits how much an interested party may contribute, generally as a percentage of the price and generally varying with occupancy and loan-to-value. An \$8,375.40 concession is 2.175% of \$385,000, which sits inside the typical caps — but "typical" is not a guideline. Verify the current interested-party contribution limit for your program, occupancy, and LTV in the applicable guide before you promise a borrower any concession. Chapter 20 covers how the concession is written into the contract, and Chapter 14 covers the conventional limits.

Unused buydown funds do not vanish. If the loan is paid off — sold or refinanced — before the escrow is exhausted, the remaining balance is generally applied to the loan rather than kept by the lender. That is a meaningful protection and worth telling the borrower, and the specific treatment should be verified against the note, the buydown agreement, and the servicer's policy.

13.9 Presenting options without steering

You now have everything. Two live programs, a collapsed down-payment ladder, a rate/point grid with break-evens in both directions, an ARM priced worst-case-first, and a buydown costed against its alternatives. One thing remains, and it is the thing this chapter is actually about.

You do not choose. They do.

That sentence is easy to agree with and hard to execute, because every incentive in the room points the other way. The borrowers want to be told. Telling them is faster. You are better at this than they are and you know it. And — this is the part the industry does not like to say out loud — different structures produce different economics for the originator, which is precisely why there is a rule about it.

⚖️ Compliance Check

Anti-steering. Regulation Z's loan originator compensation rule contains an anti-steering provision: a loan originator may not direct or "steer" a consumer to a transaction based on the fact that the originator will receive greater compensation than from other transactions the originator could have offered, unless the transaction is in the consumer's interest.

The rule provides a safe harbor. It is generally satisfied where the originator presents the consumer with loan options for each type of transaction in which the consumer expressed an interest, drawn from a significant number of the creditors with which the originator regularly does business, and the options presented include:

1. the loan with the lowest interest rate;
2. the loan with the lowest interest rate without certain risky features — negative amortization, a prepayment penalty, interest-only payments, a balloon payment in the first seven years, a demand feature, shared equity, or shared appreciation; and
3. the loan with the lowest total dollar amount of origination points or fees and discount points.

The compensation rules themselves — how an originator may and may not be paid, and why paying on loan terms is prohibited — are Chapter 26's, and you should read that chapter before you form an opinion about your own comp plan.

Two things to hold onto. First, the safe harbor's "significant number of creditors" element speaks most directly to originators who can place a file with more than one creditor; a retail originator working from a single menu satisfies it differently, but the prohibition on steering applies to every originator regardless of channel. Second, and independently: presenting different options, or different levels of effort, to similarly situated applicants on a prohibited basis is a fair lending violation under ECOA and Regulation B and the Fair Housing Act, and it does not require any compensation motive at all. Chapter 25 is the authority there.

Requirements change and state law varies. Verify the current text of the rule, your company's anti-steering policy, and your documentation requirements with your compliance department.

The rule tells you what you may not do. The craft tells you what to do instead, and it is a five-step discipline that takes about twenty minutes and makes the compliance question moot:

1. Present every structure they can actually fund, and say which ones you deleted and why. Two columns on Linden Street, not three, because Option C is not fundable — and the borrowers hear that sentence from you rather than wondering later whether you ran it.

2. Hold every assumption identical across the columns. Same price, same closing date, same taxes and insurance, same lock period. A comparison in which one column has a 30-day lock and the other has 60 is not a comparison; it is an argument with a table around it.

3. Quantify each column in the same four units. Monthly payment. Cash to close. Reserves left afterward. Total cost over a stated horizon. Borrowers can hold four numbers. They cannot hold fourteen.

4. State the assumption that decides it, and ask them to correct it. "This comparison assumes you keep the loan more than about eleven and a half years. If that's wrong, tell me now, because it changes the answer." Naming the load-bearing assumption out loud is the single most protective thing you can do, for them and for you.

5. Make a recommendation, give the reason, name what would reverse it — and stop talking. A recommendation is not steering. Refusing to make one is not neutrality; it is abandonment dressed up as caution, and borrowers experience it as an originator who will not stand behind anything. Say what you would do, say why, say what fact would change your mind, and then be quiet and let them answer.

📄 Read the File

```text FIGURE 13.2 — "The one page you hand them" [the Linden Street file] THE DOCUMENT A single-page loan comparison, prepared day 14, emailed the evening before the structure call. Not a Loan Estimate — the LE follows the application and is Chapter 22's. This is a working comparison. THE CONTEXT $385,000 purchase, 4412 Linden Street. Two programs, identical assumptions: same price, same 30-day lock, same closing date, same $385.00 taxes and $130.00 insurance. Verified funds $38,000.00. WHAT IT SHOWS

                                 OPTION A              OPTION B
                              CONVENTIONAL 95%          FHA 96.5%
─────────────────────────────────────────────────────────────────────
Purchase price                    $385,000.00        $385,000.00
Down payment                  $19,250.00 (5%)   $13,475.00 (3.5%)
Base loan amount                  $365,750.00        $371,525.00
Financed upfront MIP                        —          $6,501.69
TOTAL LOAN                        $365,750.00        $378,026.69
LTV                                    95.00%      96.50% (base)
─────────────────────────────────────────────────────────────────────
Note rate                               6.625%             6.250%
Discount points          0.500 = $1,828.75                   par
─────────────────────────────────────────────────────────────────────
Principal & interest                $2,341.94          $2,327.58
Taxes                                 $385.00            $385.00
Insurance                             $130.00            $130.00
Mortgage insurance         $176.78 (0.58%)      $173.26 (0.55%)
─────────────────────────────────────────────────────────────────────
TOTAL MONTHLY PAYMENT               $3,033.72          $3,015.84
                                                ($17.88 LESS)
Housing ratio                           28.89%             28.72%
Total debt ratio                        42.66%             42.49%
Payment vs. current $1,850 rent          +64.0%             +63.0%
─────────────────────────────────────────────────────────────────────
Cash to close                      $25,376.34   $5,775.00 less
Reserves after closing             $12,623.66        ~$18,398.66
  stated in months                 4.16 months        6.10 months
─────────────────────────────────────────────────────────────────────
Mortgage insurance ENDS      payment 137            NEVER - life
                             (11.4 years)           of the loan
Total MI/MIP paid                  $24,218.86         $62,374.40
─────────────────────────────────────────────────────────────────────
TOTAL COST OF CREDIT, 30 yrs      $503,396.01        $528,778.20
                                              ($25,382.19 MORE)
─────────────────────────────────────────────────────────────────────
Reserves shown for B hold all other closing costs equal. MI factors and
MIP factors illustrative; verify current rate cards. Total cost of credit
= points + all interest + all mortgage insurance over 360 payments,
excluding the down payment, which is equity rather than cost.

WHAT IT DOESN'T It does not price either loan at a horizon shorter than 30 years, and the borrowers' real horizon is the assumption the whole page rests on. It does not account for the time value of money. It does not show the 10%-down column, because they cannot fund it. It is not a Loan Estimate, is not a lock, and is not a commitment to lend. THE DECISION Send it the night before. Walk it on the phone in this order: payment, cash, reserves, then the two numbers at the bottom. Ask the horizon question. Recommend. Hand it back. THE LESSON A comparison that fits on one page and holds every assumption constant is the most persuasive document in origination, and it persuades by being checkable rather than by being confident. Put the number that hurts your recommendation on the same page as the number that helps it. ```

Look at what that page does to the compliance question. Both structures are shown. Both are fundable. The assumptions are identical and stated. The number that argues for FHA — \$17.88 a month cheaper, \$5,775.00 less down, nearly two more months of reserves — is on the page in the same size type as the number that argues against it. A borrower who chooses Option B off that page has made an informed choice, and an originator who wrote it has documented that they presented one.

13.10 The structure conversation, scripted

Day 15. Both borrowers on the call, comparison already in their inbox since last night. Budget twenty-five minutes and do not rush the silences.

You: "Do you both have the one-pager in front of you? Good. I want to walk it in a specific order, and I want you to interrupt me anywhere it doesn't make sense. There are no dumb questions on this call — this is a thirty-year decision and you're allowed to make me say things twice."

You: "Two columns. Same house, same closing date, same taxes, same insurance, same lock. The only thing that changes between them is the program. Start at the payment line, because that's what everybody looks at first anyway."

You: "Conventional is \$3,033.72 a month. FHA is \$3,015.84. FHA is seventeen dollars and eighty-eight cents a month cheaper. That's a real number and it's smaller than most people expect."

Borrower: "That's it? I thought FHA was supposed to be the cheap one."

You: "That's the most common misunderstanding in my business, and it's an honest one. FHA's rate is lower — 6.250% against 6.625%. But the FHA loan is bigger, because the upfront mortgage insurance premium gets financed into it, and the two effects nearly cancel. Go down to the cash line."

You: "FHA needs \$5,775.00 less at the closing table. That is the real advantage, and it's not trivial. It would leave you with about \$18,400 in the bank instead of \$12,624 — call it six months of payments instead of four."

Borrower: "So FHA is better."

You: "For a lot of people it is. I have a file on my desk right now where it's the only thing that works — the buyer doesn't have the down payment at all without it. That's not your situation, and the difference matters. You can close either way. So for you, that \$5,775.00 isn't the difference between buying this house and not buying it. It's the difference between a four-month cushion and a six-month cushion. Which is worth something. Hold that thought and look at the bottom two lines."

You: "Conventional mortgage insurance comes off automatically at payment 137 — that's eleven years and five months in — because at that point you'll owe less than 78% of what you paid for the house. FHA's runs the life of the loan. It never comes off. That's the whole comparison, and it's the two lines under it: \$24,218.86 of mortgage insurance versus \$62,374.40."

[Let that sit. Do not fill the silence. On most files this is where the borrower does the arithmetic themselves, and a conclusion a borrower reaches out loud is worth ten that you hand them.]

Borrower: "That's a thirty-eight thousand dollar difference."

You: "\$38,155.54, in mortgage insurance alone. Now let me argue against myself for a second, because FHA's lower rate does give some of that back — about \$17,400 less interest over the full term. When you net everything — the points, all the interest, all the mortgage insurance — the honest number is that the FHA loan costs \$25,382.19 more over thirty years. Not thirty-eight. Twenty-five thousand three hundred eighty-two."

You: "And here is the assumption that whole bottom section rests on, which is why I need you to answer one question honestly rather than the way you think I want."

You: "How long are you keeping this house? Not the loan — the house. Is this the house, or is this the first house?"

Borrower: "This is the house. We're not doing this again."

You: "Anything in the next five years that could move you? Transfer clause, a promotion that relocates, a parent you might need to be near, school?"

Borrower: "No. We're both from here. Her hospital's ten minutes away."

You: "Last one, and this is a different question: if rates dropped a point next year, would you refinance?"

Borrower: "Probably, yeah. Wouldn't everybody?"

You: "Yes, and that's the right answer — I want you to say yes to that. But it changes one thing, so let me put it on the table rather than leave it out. Two things, actually, and they point in opposite directions."

You: "The first: if you refinance in year three, the \$38,000 mortgage insurance advantage never fully happens, because you never get to payment 137 on this loan. The second, and this one runs the other way: if you're in the FHA loan and you refinance into a conventional one, that \$6,501.69 of upfront premium goes with you as money you still owe. There's a partial refund only if you refinance FHA into FHA, within a limited window. Conventional doesn't have anything like that to lose."

You: "So here's what I'd do, and then it's yours. I'd take the conventional at 95%. The reason is that the FHA relief on your file is a convenience, not a necessity — you can close either way — and you're paying for that convenience with mortgage insurance that never turns off, on a house you've just told me you're keeping. Seventeen dollars and eighty-eight cents a month against twenty-five thousand dollars is not a close call once I know your horizon."

You: "What would change my mind: if either of you had any real doubt about staying past about eleven years, or if four months of reserves feels too thin to you. That second one is a feeling, not a calculation, and you're allowed to weigh it. If you tell me the six-month cushion lets you sleep and the four-month cushion doesn't, that is a completely legitimate reason to take Option B and I will write it that way without arguing."

Borrower: "No — four months is fine. We've got both incomes. Let's do the conventional."

You: "Good. One more decision and it's a smaller one. Look at the rate grid I sent underneath."

You: "Par today on your file is 6.750%, no points. For \$1,828.75 you can have 6.625%, which saves you \$30.31 a month. That's a sixty-point-three month break-even — five years and one month before that \$1,828.75 comes back. Given what you just told me about staying, I'd buy it. But understand the downside: if you refinance at three years you'll have gotten \$1,091.16 of it back and lost \$737.59. Worst case is you refinance next month and lose the whole \$1,828.75."

Borrower: "And if we stay?"

You: "If you're still in this loan at payment 137 you're \$2,323.72 ahead. At thirty years, \$9,082.85. Your call."

Borrower: "Take the point."

You: "Done. I'm going to email you a two-paragraph summary tonight that says exactly what you chose and why, including the FHA option you turned down and the number that made it close. Read it. If anything in it doesn't match your memory of this call, tell me tomorrow, not in three weeks."

Four things in that script are doing structural work, and they are worth naming because they transfer to every structure conversation you will ever have.

The order. Payment, cash, reserves, total cost, horizon, recommendation. Payment first because that is what they are already thinking about; total cost late because it only means something once they trust the earlier numbers.

The counterexample. "I have a file on my desk right now where it's the only thing that works." One sentence, and it converts "the loan officer doesn't like FHA" into "the loan officer picks programs based on borrowers." It is also true, which is why it works.

Arguing against yourself. Volunteering the \$17,446.29 of interest FHA gives back — and revising your own headline from \$38,155.54 down to \$25,382.19 — costs you nothing and buys you every subsequent number on the page. An originator who corrects their own strongest argument is an originator a borrower stops fact-checking.

The written confirmation. Two paragraphs the same night, in the borrower's own decision language, naming the option they declined. It protects them, it protects you, and in four years when somebody's brother-in-law says they should have gone FHA, it exists.

🗂️ The Loan File

Chapter 13 contribution: the decision.

Day 15. Everything from Chapters 9 through 12 is verified. Three structures were built; one was deleted for lack of funds; two went to the borrowers on one page; the borrowers chose.

Step 1 — Which columns are fundable?

The 10% column is arithmetically unavailable before closing costs are even considered: \$38,500.00 of down payment against \$38,000.00 of total verified funds is a \$500.00 shortfall on the down payment line alone. Two columns survive. (FHA cash figure holds all other closing costs equal; the precise figure is Chapter 12's.)

Step 2 — The rate/point grid, and the break-even.

[constructed teaching grid — verify current pricing at the source]

$$\text{Break-even} = \frac{\$1{,}828.75}{\$30.31} = 60.3 \text{ months} = 5.0 \text{ years}$$

Step 3 — The two-column comparison. Figure 13.2, in full. The four decisive lines:

Step 4 — Total cost of credit, decomposed. Over 360 payments, excluding the down payment (which is equity, not cost):

Three components, two directions, one answer. The headline everyone quotes — \$38,155.54 more in mortgage insurance — is real, and it overstates the case by \$12,773.35, because FHA's lower note rate gives back \$17,446.29 of interest against \$4,672.94 of extra upfront charge. The honest number is \$25,382.19. Use the honest number. (Interest totals treat every payment as exactly level for 360 months; real final payments differ by pennies.)

Step 5 — The outcome.

THE DECISION: Conventional 30-year fixed, 5% down. Loan \$365,750.00, LTV 95.00%, rate 6.625% with 0.500 discount point (\$1,828.75), 30-day lock, borrower-paid monthly mortgage insurance at a 0.58% factor. PITI + MI = \$3,033.72. Housing 28.89%, back-end 42.66%.

The borrowers decided, on the numbers above, on day 15.

The reasoning, in the order it was reached:

1. FHA's \$5,775.00 of down-payment relief is convenience here, not feasibility. They have \$38,000.00 verified and close either way, with \$12,623.66 — 4.16 months — left over. That distinction is the whole question, and for a borrower where it went the other way, FHA would be the right answer. It is why the Harlow Street file goes FHA.
2. \$17.88 a month against \$25,382.19 of total cost is not a close call once the horizon is known. It would be a close call if the horizon were short.
3. They intend to keep the loan. Stable local employment of three and four years, a family home rather than a starter, no relocation exposure, no plan to move. They would refinance if rates fell substantially — the one fact that weakens this recommendation, stated to them out loud.
4. 10% down was arithmetically unavailable, not rejected.
5. The half point at 6.625% was taken because break-even is 60.3 months and their horizon exceeds it.

What would have changed it:

• If their verified funds had been \$32,000 instead of \$38,000, the \$5,775.00 becomes feasibility and the file goes FHA. Same programs, same rate sheet, different answer.
• If the representative score were below the point where conventional mortgage insurance prices competitively — as on Harlow Street at 641 — the file goes FHA.
• If they had answered the horizon question with five years or less, or named a possible transfer, par at 6.750% is the right answer and the \$1,828.75 stays in their savings account.
• If a four-month cushion had felt too thin to them, Option B was theirs to take and would have been written without argument. That one is a judgment about how they sleep, not an arithmetic error.

What this settles: the program, the down payment, the term, the rate and point position, the mortgage insurance structure, and that there is no buydown on this file. All six structure decisions, made and documented. Q2 is answered.

What it does not settle: any of it, until the appraisal supports \$385,000 (Q3) — a low appraisal re-opens the entire down-payment analysis, as the Cypress Court file shows. Nor does it settle where 6.625% came from (Chapter 29), when the rate gets locked (Chapter 30), or whether the file survives underwriting (Part III). The rate is not locked as of day 15, and a rate that moves before it is locked moves every number in Figure 13.2.

Open questions:

• Q2. RESOLVED — conventional 95%, 6.625% with 0.500 point. (This chapter.)
• Q3. Will an appraisal support \$385,000? (Chapter 18)
• New: The borrowers have been shopping and have an online lender's advertised rate in hand, lower than 6.625%. Nothing on Figure 13.2 addresses it, because an advertised rate is not a quote for a specific file. (Chapter 29 rebuilds where a rate comes from; Chapter 40 prices that competing quote honestly against this file.)
• New: The rate is not locked. (Chapter 30)

Your task. In Appendix C's workbook, complete the structure page. Record all six structure decisions with the number that drove each. Then do the thing this chapter is really teaching: write the two-paragraph confirmation email in the borrower's language, naming the option they declined and the number that made it close. If you cannot write that email without hedging, you have not finished the analysis.

Conclusion

A loan is six decisions, not one. Program, down payment, term, rate and point position, mortgage insurance structure, and concessions. The first three mostly determine whether the loan is possible; the last three mostly determine what it costs, and they are the ones nobody runs.

Program selection is a statement about a borrower, never about a product. The four questions — cash, credit, property and occupancy, horizon — narrow the field in that order, behind a gate that asks about entitlement first. On the cash question, everything turns on a single distinction: does the lower down payment buy feasibility or convenience? On Harlow Street it is feasibility, and FHA is unambiguously right. On Linden Street it is convenience, and \$5,775.00 of relief at the closing table costs \$12,276.69 of additional debt and mortgage insurance that never turns off.

Points and lender credits are the same arithmetic run in opposite directions, and they are not symmetric: on this grid a buydown recovers in about sixty months and a credit repays in about forty-five, because it costs more to buy an eighth of a point of rate than it pays to sell one. Both break-evens are meaningless without the horizon, which is the only load-bearing input you cannot document. You have to ask, and you have to ask three ways.

An ARM must be presented worst case first, and on a standard ARM the Ability-to-Repay qualifying rate means the low introductory payment does not improve qualification — on this file it made the debt-to-income ratio worse. A temporary buydown is an escrow account, not a rate; the borrower qualifies at the note rate; and the \$8,375.40 it costs on this loan is more than the money that would have bought the rate down permanently.

And then you hand it back. You present every structure they can fund, hold the assumptions identical, quantify in the same four units, name the assumption that decides it, recommend, say what would reverse your recommendation, and stop talking. The borrowers chose conventional 95% at 6.625% with a half point on day 15, on numbers they could check.

Next: the structure is decided, and now it has to survive somebody who has never met these people and will not take anyone's word for anything. Part III opens with the conventional rulebook — what Fannie Mae and Freddie Mac actually require, what your own employer layers on top of it, and why the distinction between a guideline and an overlay is one of the most useful things a loan officer can know.

Key Terms

Loan structure — the specific combination of six decisions that turns a purchase price and a household into one particular note: program, down payment, term, rate and point position, mortgage insurance structure, and concessions or buydowns. (Ch.13)

Program fit — whether a program's eligibility rules, cost structure, and mortgage insurance cancellation terms match a specific borrower's cash, credit, property, and horizon; a statement about a borrower rather than about a product. (Ch.13)

Permanent buydown — discount points paid at closing to reduce the note rate for the entire life of the loan; the rate on the note itself is lower. (Ch.13)

Temporary buydown — an escrow account funded at closing by the seller, builder, or another party, drawn down monthly to subsidize part of the borrower's payment for a stated period while the note rate remains unchanged; the borrower is qualified at the note rate. (Ch.13)

2-1 buydown — the common form of temporary buydown: the effective rate is reduced by 2 percentage points in year one and 1 percentage point in year two, then the note rate applies for the remaining term. (Ch.13)

Total cost of credit — the sum of everything a borrower pays to obtain and carry a loan over the period they actually hold it: points and origination charges, all interest, all mortgage insurance, and any financed one-time charge. Distinct from the note rate, from the monthly payment, and from APR — APR is a standardized yearly rate over the full term, while total cost of credit is a dollar total over a stated horizon. The down payment is excluded, because it is equity rather than cost. (Ch.13)

Loan comparison — a side-by-side presentation of two or more complete, fundable structures for the same borrower and property, on identical assumptions, quantified in the same units, with the load-bearing assumption stated. (Ch.13)

Spaced Review

1. (Ch. 5 + Ch. 13) A borrower is offered a 5/6 ARM with an initial rate of 5.500%, an index of 4.25%, a margin of 2.75%, and 2/1/5 caps. State the fully indexed rate, state the rate the file must be qualified at under Ability-to-Repay and why, and state the highest rate the note permits over its life.

2. (Ch. 12 + Ch. 13) The Linden Street borrowers have \$38,000.00 verified, of which \$10,000 is a documented gift. Cash to close on the chosen structure is \$25,376.34. Compute the reserves remaining, express them in months, and explain in one sentence why 10% down is not a structure the borrower declined.

3. (Ch. 13) Par on Figure 13.1 is 6.750% at \$2,372.25. A borrower is considering 6.500% for \$3,657.50. Compute the break-even, then state the two facts you would need from the borrower before recommending either way.

4. (Ch. 5 + Ch. 13) Conventional mortgage insurance on the chosen structure ends at payment 137; FHA's runs the life of the loan. Explain why that single difference outweighs a monthly payment that is \$17.88 lower — and then state the horizon at which it would stop outweighing it.

5. (Ch. 12 + Ch. 13) A builder offers a 2-1 buydown worth \$8,375.40 on this loan. The borrower's agent asks whether the buyer could take the money as a price reduction instead. Answer in three sentences: what each choice is worth in the first twenty-four months, what each is worth over thirty years, and what fact about the borrower decides it.