Chapter 1: The Mortgage Industry: How Home Lending Works and Why Loan Officers Exist

"A house is bought with somebody else's money. Everything else in this business is a consequence of that sentence." — constructed; the working premise of this book

Overview

The call comes in at 8:40 on a Wednesday and it is always the same call. "What's your rate?"

You can answer that question in four seconds and lose the customer in five, because whatever number you say, somebody's website says a smaller one. The number on that website is real, in the sense that some borrower somewhere could get it: a borrower with a 780 score putting twenty-five percent down on a single-family home in a county with no overlays, locking for fifteen days, buying a point and a half. The person on your phone has a 706, five percent down, and a closing date forty-five days out. The rate they were quoted does not exist for them and will never exist for them, and nobody has told them that yet.

What you do in the next ninety seconds decides whether they find out from you — now, while it is free — or from a Closing Disclosure three days before they are supposed to get keys.

That is the job. Not the rate. The rate is the thing people call about. The job is everything that happens after the call.

To do that job you need to know something the caller does not: where the money is actually coming from. Not your employer — your employer is a middleman, however large. The money for the loan on 4412 Linden Street will be supplied by an investor several steps removed from this conversation, who will never know your borrowers' names, and who has published, in advance, the exact terms on which they are willing to lend to people with a 706 score putting five percent down. Those terms are why your rate is what it is. They are why the underwriter will ask for a letter explaining a \$4,900 deposit. They are why there are eleven conditions on the approval and why one of them cannot be waived.

This chapter draws that picture. By the end of it you will be able to follow the money from a pension fund to a kitchen table and back, name every party it passes through, and say what each of them is for. It is the least glamorous chapter in the book and the one everything else assumes.

In this chapter, you will learn to:

• Explain what a mortgage is as a legal structure, and why it takes two documents rather than one
• Distinguish a mortgage from a deed of trust, and say why the difference matters to a borrower
• Trace the money in a residential mortgage from the investor who supplies it to the borrower who spends it
• Distinguish the primary market from the secondary market and explain what each is for
• Describe what an originator, processor, underwriter, closer, and servicer each do
• Compare retail, broker, and correspondent origination in terms of whose money funds the loan
• State what a loan officer is actually paid to produce

Learning Paths

🎓 Exam — §1.2 and §1.4 are directly testable. Know the note-versus-security-instrument distinction cold, and know that Fannie Mae, Freddie Mac, and Ginnie Mae do not originate loans. 🏠 New LO — §1.1 and §1.7. The rest is context; these two are your job description. 🤝 Partner — §1.3 and §1.5. If you understand who actually decides and who merely relays, you will stop calling the wrong person when a transaction is in trouble. 📊 Operations — §1.5 and §1.6, and pay attention to §1.6's effect on turn times.

1.1 The call that starts everything

Let us stay on the phone a moment longer, because the whole book is in this call.

The caller is a buyer's agent you have closed four files with. She is calling on behalf of clients who are, at this moment, sitting at her conference table writing an offer on a house they saw Saturday. They have not spoken to a lender. The listing agent has told her the sellers will not look at an offer without a pre-approval letter attached, which is standard and which is also why this call is happening at 8:40 rather than next week.

Here is what you actually know:

• There is a house. Somebody has priced it at \$385,000.
• There are two people who want it and have some money.
• There is a deadline measured in hours.
• There is no file, no application, no credit report, and no verified fact of any kind.

And here is what you are being asked for: a letter, on your company's letterhead, stating that these two people can borrow enough money to buy this house.

📞 On the Phone

Agent: "Hey — can you get me a pre-approval by two o'clock? They're writing at \$385."

The wrong answer: "Sure, send me their info." You now owe a letter you cannot support, and the first fact you learn about these people will be learned after you have vouched for them.

The other wrong answer: "I can't do anything without a full application, three years of tax returns, and two months of bank statements." True, thorough, and you have just cost your agent a transaction and yourself the next four.

What actually works: "Yes. I need twenty minutes on the phone with them and their permission to pull credit. Get me that today and you'll have a letter in the morning that the listing agent will take seriously. Can you three-way me in at ten?"

Note what that answer does not do: it does not promise the letter by two o'clock because the agent asked for two o'clock. The twenty minutes and the credit pull have to happen first, and on this file they happened on day 1 — the offer went out the night of day 0 and the letter reached the listing side well before the offer was accepted on day 4. Commit to a time you control. An agent remembers a letter that arrived when you said it would; nobody remembers the two hours.

The third answer is not a compromise between the first two. It is a different thing: you are naming the minimum you need to say something true, and committing to a time. Everything in Chapter 8 elaborates on those twenty minutes.

Notice what has already happened. Before you know a single number about these borrowers, the transaction has imposed a clock on you. That clock will not let go for fifty-one days. This is worth sitting with, because new loan officers consistently model their job as analysis punctuated by deadlines, and it is the reverse: it is a deadline, continuously, inside which analysis has to happen.

The second thing that has happened is that you have been asked to make an assertion about strangers' finances to a third party who will rely on it. That is not a sales activity. It is closer to what an auditor does, and it is why this job is licensed.

⚠️ Where Deals Die

The unsupported pre-approval letter is the single most common self-inflicted wound in residential lending. It costs nothing to issue and it can cost a family their earnest money.

The mechanism: you issue a letter on a conversation. The sellers accept the offer and take the house off the market. The buyers waive inspection to be competitive. Twenty days later the verified income comes back lower than what the borrower told you — not because anyone lied, but because they quoted you their gross including a bonus that does not have a two-year history. The loan is now \$40,000 smaller than the letter said. If the financing contingency has expired, the earnest money is at risk.

Chapter 8 draws the line between pre-qualification and pre-approval precisely. Chapter 20 explains what a financing contingency actually protects. For now: a letter is a statement of fact, and you should be able to point at the document supporting every fact in it.

What the listing agent is actually reading

The letter you are being asked for will be read by someone whose interests are opposed to your borrower's. That is worth saying out loud, because new loan officers draft pre-approval letters as though the audience were their own client.

It is not. The audience is a listing agent with three offers on a kitchen counter and forty minutes before a seller wants an opinion. That person is not evaluating your borrowers as human beings. They are estimating one thing: how much of their seller's time this offer is likely to waste. Every feature that makes a letter persuasive follows from that single question.

So a letter that works says what was actually checked. Not "we have reviewed the borrowers' financial information" — that sentence appears on every letter ever written and therefore carries no information at all. It says: credit was pulled on this date and the representative score supports this program; income was reviewed from these document types; assets were reviewed from these document types; and the following remain unverified. A listing agent who reads two letters, one of which claims everything and specifies nothing and one of which admits exactly what is still open, will believe the second one. Specificity reads as competence because it usually is.

WHAT A CREDIBLE PRE-APPROVAL LETTER CONTAINS        [constructed teaching example]

  ┌─────────────────────────────────────────────────────────────────────────┐
  │  LETTERHEAD  lender legal name · NMLS company ID · your NMLS ID ·       │
  │              a phone number that a stranger can call today             │
  │  DATE        the date issued — a letter with no date is a rumor         │
  │  PROPERTY    "4412 Linden Street" or, deliberately, "not property       │
  │              specific" — say which, don't leave it ambiguous            │
  │  AMOUNT      the loan amount and the purchase price it assumes          │
  │  STRUCTURE   program, term, down payment, occupancy                     │
  ├─────────────────────────────────────────────────────────────────────────┤
  │  WHAT WAS VERIFIED                                                      │
  │    credit report pulled [date]; representative score supports the       │
  │      program named above                                                │
  │    income reviewed from: paystubs, W-2s                                 │
  │    assets reviewed from: two months of account statements               │
  │  WHAT IS NOT YET VERIFIED                                               │
  │    employment reverification at closing; appraised value; title;        │
  │      final underwriting decision                                        │
  ├─────────────────────────────────────────────────────────────────────────┤
  │  SIGNATURE   a human being's name, title, direct line, and email        │
  └─────────────────────────────────────────────────────────────────────────┘

  The two blocks in the middle are the entire letter. Everything above them is
  identification and everything below them is courtesy.

Notice the block labeled WHAT IS NOT YET VERIFIED. Loan officers resist writing it, on the theory that admitting weakness weakens the offer. The opposite is true, and the reason is structural: a listing agent has been burned before by a letter that promised more than the file could deliver, and they are reading yours for signs that it is the same kind of letter. A letter that names its own limits is the only kind that can be trusted, because it is the only kind whose author has clearly thought about what could go wrong.

Two more practical points, both learned expensively. Put a direct phone number on the letter and answer it. Listing agents call, and on a competitive property that call is frequently the actual decision — not the letter, the call. And never issue a letter for more than the offer. A letter that says your borrowers are approved to \$430,000 attached to an offer at \$385,000 tells the seller exactly how much more they could have asked for, and you have just cost your own client money with your own stationery.

The twenty minutes, in outline

Chapter 8 does this conversation properly. What follows is only the shape, because you should know what you are committing to before you commit to it on the phone.

You need six things, and they come in this order for a reason. One: the transaction — the address, the price, the closing date they are about to name, and whether anyone has asked for anything unusual. Two: income — what kind, how long, and paid how. Not "how much do you make," which invites a number nobody can document, but "walk me through how you get paid," which produces the structure. Three: assets — what they have, where it sits, and where it came from. Four: debts — what they pay every month whether they want to or not. Five: permission, on the record, to pull credit. Six: what happens next, with a time attached.

The last one is the deliverable. Twenty minutes on the phone does not produce a yes. It produces a commitment to a time, which is the only thing you can honestly give somebody at ten in the morning about a file that does not exist yet.

There is one more question, and it belongs at the end when the borrower has relaxed: "Is there anything about your finances that would surprise me if I found it in a document?" Ask it exactly that way. Not "is there anything I should know," which people answer honestly and incompletely, because they do not know what you should know. The document framing works because it puts the borrower in the underwriter's chair for four seconds, and it is where you learn about the co-signed car loan, the eleven-month-old collection, the money that is in a parent's account, and the second job that started in March. Every one of those is survivable on day 0 and expensive on day 33.

And know what the twenty minutes does not license you to do. It does not license a rate quote presented as a price, because you have not priced anything. It does not license naming a program as settled. It does not license telling them what they qualify for before the credit report is on your screen. What it licenses is a letter you can defend, and a date.

1.2 What a mortgage actually is: the note and the security instrument

Almost everyone, including many people who work in this industry, uses "mortgage" to mean "the loan on my house." That is fine in conversation and wrong in every context where it matters.

A residential mortgage is two documents, and they do two different jobs.

The note

The note — formally the promissory note — is the borrower's promise to repay. It states the amount borrowed, the interest rate, the payment, the payment due date, the term, what happens if a payment is late, and what happens if the borrower stops paying. It is signed by the borrower. It is a debt instrument, and it would be perfectly valid even if the borrower owned nothing at all.

If you strip everything else away, the note is the loan.

The security instrument

The security instrument is a completely separate document that pledges the property as collateral for the note. It creates a lien — a legal claim against a specific piece of property, recorded in the public land records of the county where the property sits, which gives the lender the right to force a sale of that property if the note is not paid.

The security instrument is recorded. The note is not. This surprises people. The promise to repay is private; the claim against the land is public, because the entire point of a land record system is that anyone can find out what is attached to a parcel.

THE TWO DOCUMENTS                                      [constructed teaching example]

  ┌────────────────────────────────────┐   ┌────────────────────────────────────┐
  │            THE NOTE                │   │      THE SECURITY INSTRUMENT       │
  │  (promissory note)                 │   │  (mortgage OR deed of trust)       │
  ├────────────────────────────────────┤   ├────────────────────────────────────┤
  │  "I promise to pay $365,750 at     │   │  "To secure that promise, I pledge │
  │   6.625% in 360 monthly payments   │   │   4412 Linden Street. If I don't   │
  │   of $2,341.94 beginning Dec 1."   │   │   pay, it can be sold."            │
  ├────────────────────────────────────┤   ├────────────────────────────────────┤
  │  NOT recorded.                     │   │  RECORDED in the county land       │
  │  Held by whoever owns the debt.    │   │  records. Public. Creates a LIEN.  │
  │  Can be sold, and will be.         │   │  Follows the property, not the     │
  │                                    │   │  borrower.                         │
  └────────────────────────────────────┘   └────────────────────────────────────┘
            │                                            │
            └──────────────── together ──────────────────┘
                                 ↓
                    what everyone calls "the mortgage"

Most loan officers never read a note. They should read one, once, carefully — because every argument in this book about documentation, ratios, and structure is ultimately an argument about whether a household can perform the promise on this single page.

📄 Read the File

text FIGURE 1.1 — "The promise itself" [the Linden Street file] THE DOCUMENT Multistate Fixed Rate Note, one page plus signature page, signed at closing on day 51. Not recorded; held by whoever owns the debt. THE CONTEXT A $385,000 purchase, 5% down. The note is executed after fifty-one days of work by six people, and it says nothing about any of it. WHAT IT SHOWS Principal $365,750.00. Rate 6.625%, fixed for the full term. 360 monthly payments of $2,341.94, due the 1st, beginning December 1. Late charge after a 15-day grace period. No prepayment penalty. Acceleration on default after notice and a cure period. The borrowers are jointly and severally liable — each owes all of it, not half. WHAT IT DOESN'T It does not mention the property. It does not mention taxes, insurance, or mortgage insurance, so the $2,341.94 it names is NOT the payment the borrowers will actually make ($3,033.72 — see Chapter 4). It does not name the investor who supplied the money, does not name the servicer who will collect the payment, and does not promise that either will stay the same. It says nothing about income, credit, or ratios — every fact this book spends thirteen chapters documenting is absent from the document those facts existed to produce. THE DECISION At the closing table: confirm the rate, term, and payment match the Closing Disclosure the borrowers received three days earlier. A note that disagrees with the CD stops the closing — it does not get fixed afterward. THE LESSON The note is the only document in the file that lasts thirty years, and it contains almost none of the information the file was built out of. Underwriting is the process of becoming confident about a promise that, once made, is enforceable without reference to any of the evidence.

Constructed. Real notes use standardized agency forms; the terms above are the frozen figures for this book's running file.

One line in that figure deserves unpacking, because borrowers ask about it and most loan officers answer it wrong. The borrowers are jointly and severally liable — each owes all of it, not half.

"Jointly and severally" is not a formality and it is not a statement about the marriage. It means the holder of the note may collect the entire balance from either borrower, in any proportion, without first pursuing the other. Two people did not each borrow \$182,875. Two people each borrowed \$365,750, and the lender is entitled to be paid once.

The practical consequences arrive years later and are worth knowing on day one, because you will be asked. If the borrowers separate, the divorce decree can assign the debt to one of them — and it binds the two of them, not the lender, who was not a party to it and whose note says what it says. The only ways a name comes off a note are a refinance into a new loan, a formal release of liability if the program permits one, or paying it off. "I'll just take my name off the mortgage" describes nothing that exists. Similarly, if one borrower dies, the obligation does not shrink; the survivor owes the whole payment, which is the entire reason anyone should think about life insurance and the reason a thoughtful loan officer raises it once, gently, and then leaves it alone.

Note also what the figure says about prepayment: there is no penalty on this note. A borrower can pay it off tomorrow, in full, at par, and owe nothing extra. Hold on to that fact — §1.3 is going to show you that it is one of the most expensive rights in the whole transaction, and that the borrower did not pay for it directly.

Why does this matter to a loan officer on a Wednesday morning? Three reasons, and all three come up in real files.

First, because the note and the lien can travel separately. The debt gets sold — routinely, and usually within weeks of closing. The lien stays recorded where it is. When a borrower calls you in February saying "I got a letter from a company I've never heard of, is this a scam?", you are explaining this separation. Chapter 28 follows the whole journey.

Second, because lien position is everything in a foreclosure. Liens are ranked, generally by the date and time they were recorded — this is lien priority, and Chapter 21 takes it apart. A first-position lien gets paid first out of a forced sale. A second gets what is left. This is the entire reason a title search exists, and it is why a mechanic's lien from a prior owner's contractor, recorded three years before your loan, is a genuine problem rather than a paperwork annoyance.

Third, because the security instrument is what makes a mortgage cheap. A thirty-year unsecured loan to a household at 6.625% does not exist and would be insane to make. The reason a mortgage carries a rate in single digits while a credit card carries one in the twenties is not the borrower's virtue. It is that the lender can take the house.

Why it takes two documents and not one

The two-document structure is not a drafting habit. It exists because the promise and the claim live under two different bodies of law, move by two different mechanisms, and are addressed to two different audiences.

The note is a debt. It is governed by contract and negotiable-instrument principles, it binds the people who signed it, and it moves the way a check moves — by endorsement and delivery. Nobody outside the transaction needs to know it exists, and nobody outside the transaction is bound by it. If your borrower signs a note and the lender sells the debt three times, the borrower's obligation is unchanged and no stranger's rights are affected.

The security instrument is an interest in land. It is governed by the property law of the state where the land sits, and it binds the world — including people who have never heard of your borrower. A future buyer, a future lender, a judgment creditor, a title insurer: all of them are affected by whether this lien exists. And the only way a claim can bind people who were never party to it is if those people had a place to go and look. That place is the county recorder's office. Recording is not a formality; it is the entire mechanism by which a private agreement becomes enforceable against strangers.

Now notice the asymmetry, because it explains a great deal of what underwriting is for. A note without a lien is still a debt. It is a bad debt — unsecured, expensive, hard to collect — but it exists and can be enforced against the borrower personally. A lien without a note secures nothing. It is a claim in support of an obligation that does not exist, and a court will discharge it. The promise is the thing; the property is the backstop.

This is why the underwriting file is organized the way it is. Everything about the borrower — income, assets, credit, ratios — is evidence about the note: will this promise be performed? Everything about the property — appraisal, title, insurance, flood determination, survey — is evidence about the security instrument: if the promise is not performed, is there something here worth taking, and will our claim on it come first? Two documents, two evidence files, two ways for a loan to die. A file with perfect income and a clouded title is as dead as a file with clean title and no documentable income, and the two failures do not substitute for each other in either direction.

What the security instrument actually says

Almost no loan officer has read one, and the ones who have are noticeably better at their jobs, because roughly a third of the questions a borrower asks in the first year are answered by a document the borrower already signed.

The agency security instruments are uniform instruments: standardized forms, with a state-specific rider where local law demands one, containing a numbered list of promises the borrower makes about the property. The numbering has shifted across form revisions, so read the covenants in the actual instrument in your file rather than quoting a paragraph number from memory. The substance, in plain language:

THE UNIFORM COVENANTS, IN PLAIN LANGUAGE       [constructed teaching example — read
                                                the instrument in your own file]

  THE COVENANT                    WHAT THE BORROWER PROMISES
  ─────────────────────────────────────────────────────────────────────────────
  payment                         pay principal, interest, late charges when due
  funds for escrow items          pay 1/12 of taxes and insurance with the payment
  charges and liens               pay anything that could become a lien ahead of us
  property insurance              keep it insured, name us, don't let it lapse
  occupancy                       live here as your principal residence
  preservation and maintenance    don't let the house fall down
  the loan application            the statements you made to get this loan were true
  protection of lender's interest  if you don't, we may — and bill you for it
  mortgage insurance              maintain it while it is required
  transfer of the property        if you sell or transfer, we may call the loan due
  right to reinstate              you can cure a default and stop the process
  sale of note; change of servicer  the note can be sold and the servicer can change
  acceleration and remedies       notice, a cure period, then the whole balance
  release                         when it's paid, we release the lien
  ─────────────────────────────────────────────────────────────────────────────
  Wording and numbering vary by form revision and by state. Verify against the
  instrument in the file, and confirm state-specific riders with your closing
  attorney or title officer.

Four of those deserve a sentence of mechanism, because they are the four you will be asked about.

Escrow. Borrowers ask why they cannot simply pay their own taxes, and the honest answer is not "because we don't trust you." It is that in most states an unpaid property tax creates a lien that takes priority over a recorded mortgage regardless of when the mortgage was recorded — the one exception to the first-in-time rule that Chapter 21 develops. A borrower who stops paying taxes can therefore quietly move the lender out of first position without missing a single mortgage payment. Escrow is the lender protecting its own lien priority, and the borrower gets an involuntary savings plan out of it. Verify how your state treats tax liens; the priority rule and the redemption period vary.

Insurance. The collateral has to keep existing. An uninsured house that burns converts a secured loan into an unsecured one overnight, which is why the covenant permits the lender to buy coverage if the borrower lets it lapse — force-placed insurance, which is expensive, protects only the lender, and is one of the most common reasons a servicing relationship turns hostile.

Occupancy. The borrower promises to occupy the property as a principal residence, generally for a stated period after closing. This is not sentiment. Occupancy drives pricing and it drives default behavior — when money is short, households pay for the roof they sleep under before they pay for the one they rent out — and the whole rate structure assumes the statement is true. Signing this covenant while intending to rent the property is a breach of the security instrument and, if the statement was made to obtain the loan, potentially a crime. Chapter 27 handles it from the detection side.

Transfer of the property. This is the due-on-sale clause, and it is the answer to a question you will hear within your first month: "Can I just take over the sellers' loan? Their rate is better than anything you can give me." On a conventional loan, generally no — the covenant lets the lender accelerate on transfer, federal law permits enforcement, and there is a list of statutory exceptions for certain family and estate transfers that you should verify with counsel rather than recite from memory. Government loans behave differently, which is a genuine planning point rather than trivia.

Three things get sold, and they are not the same thing

The chapter said the note and the lien can travel separately. The full picture is that three distinct assets come out of one closing, each with its own buyer, its own transfer mechanism, and its own paper trail.

ONE CLOSING, THREE SALEABLE ASSETS                     [constructed teaching example]

  THE DEBT                THE LIEN OF RECORD          THE SERVICING
  (the note)              (the security instrument)   (the right to collect)
  ─────────────────────   ─────────────────────────   ─────────────────────────
  transfers by            transfers by ASSIGNMENT,    transfers by contract;
  ENDORSEMENT and         recorded in the county —    the buyer pays for the
  delivery; often on an   OR the instrument names     right to service and
  attached allonge        an electronic registry as   earns a fee out of each
                          nominee, and the transfer   payment
                          is tracked there instead
  ─────────────────────   ─────────────────────────   ─────────────────────────
  NOT public              PUBLIC (or in the registry) NOT public — but the
                                                      borrower gets written
                                                      notice, from BOTH the
                                                      old and new servicer
  ─────────────────────────────────────────────────────────────────────────────
  Which is why a borrower can end up with three different company names in
  their paperwork and nothing whatsoever has gone wrong.

An assignment is the recorded transfer of the security instrument — the public record catching up with a private sale. An endorsement is the transfer of the note itself, which is why a loan can change hands without anything appearing in the land records at all when the instrument names an electronic registry as nominee for the lender and its successors. And the servicing is a third thing again: the contractual right to collect the payment, administer the escrow, and be paid a fee for doing it. It is bought and sold in its own market, and it is the reason your borrower's payment address changes even when the loan did not.

So the February phone call — "I got a letter from a company I've never heard of, is this a scam?" — has a complete answer, and the answer is in the document they signed. One of the uniform covenants told them in advance that the note could be sold and the servicer could change. That is not fine print doing something sneaky; it is fine print doing exactly what it said.

What you should also teach that borrower, in the same call, is what a real transfer looks like and what a fraud looks like, because the two are distinguishable and the distinction is worth money. A legitimate servicing transfer arrives as written notice from both the old servicer and the new one, on a schedule set by federal law, naming an effective date; and for a protective window after the transfer a payment sent in good faith to the old servicer cannot be treated as late. Verify the current notice timing and grace period with your compliance department, because the details are prescriptive. A fraud looks different: an email only, no notice from the current servicer, urgency, a new wire instruction, and a request to send money somewhere that is not the address on the last statement. The instruction to give the borrower is one sentence long — call the number on your most recent statement, not the number in the message — and it costs you thirty seconds to say.

Where the two documents physically go

On day 51 the borrowers sign a stack of paper and then have no idea what happens to any of it. Here is where the two that matter actually go, because a borrower will ask and because the answer explains a delay they will otherwise find alarming.

The security instrument goes to the county. The closing agent records it — along with the deed conveying the property from the seller to the buyer — in the land records of the county where the property sits, and the order matters: the deed goes first, because the buyers cannot pledge property they do not yet own. In most transactions this happens the same day as signing and funding. The recorded originals come back weeks or months later, stamped with a book and page or an instrument number, and by then the borrower has usually forgotten they were coming.

The note goes to a custodian. It is an original, and the original matters — it is the instrument that proves the debt. It travels to the lender's document custodian and, once the loan is sold into a pool, to the custodian for that pool, endorsed along the way. The borrower gets a copy. Nobody records anything.

The practical residue for you: the gap between signing and a clean recorded record is a real window, and it is the window in which title insurance is earning its premium. It is also why the "closing" your borrower experienced on Friday afternoon may not be the moment the transaction was legally complete, which is exactly the distinction §1.9 draws between closing, funding, and recording. Verify how your state and your market sequence it before you tell a family when the truck can be unloaded.

🎓 NMLS Exam Watch

The note-versus-security-instrument distinction is heavily tested, and the exam likes to phrase it as a trap: "Which document creates the lien?" The security instrument. "Which document is the evidence of the debt?" The note. "Which is recorded?" The security instrument.

A related favorite: the mortgagor is the borrower and the mortgagee is the lender. The "-or" is the one giving the mortgage. Candidates reverse this constantly under time pressure. One memory hook: the borrower is the one who offers the property — mortgor gives.

Mortgage or deed of trust?

There are two forms the security instrument takes in the United States, and which one you use is determined by the state the property is in — not by the lender, the borrower, or anyone's preference.

A mortgage has two parties: the mortgagor (borrower) and the mortgagee (lender).

A deed of trust has three: the trustor (borrower), the beneficiary (lender), and a neutral third party called the trustee who holds legal title, or a power of sale, until the debt is satisfied.

That third party is the whole point. States that use deeds of trust generally permit non-judicial foreclosure — the trustee can conduct a sale under the power granted in the document, following a statutory notice process, without a lawsuit. States that use mortgages generally require judicial foreclosure, meaning the lender must sue.

There is a further layer, which the exam asks about: lien theory versus title theory. In a lien-theory state, the borrower holds title and the lender holds only a lien. In a title-theory state, the lender (or trustee) holds legal title until the debt is paid, while the borrower holds equitable title and possession. A few states take an intermediate position. In practice the distinction shows up mostly in foreclosure procedure.

What happens when the promise is not kept

You will never conduct a foreclosure and you should still know its shape, for two reasons. The first is that every guideline you will spend this book learning was written by somebody who was trying to avoid this outcome. The second is that borrowers ask, usually obliquely — "what happens if something goes wrong?" — and the answer they deserve is not "let's not think about that."

The sequence, generically, and with the warning that every step of it is state law and varies enormously:

A payment is missed. The note's grace period runs — commonly fifteen days on the agency form — and a late charge attaches. The delinquency is reported to the credit bureaus once it reaches thirty days, and it is worth knowing that the credit damage begins here, long before anything legal happens. The servicer, not the original lender, is the party doing all of this, and the servicer's first obligation is not to foreclose but to attempt loss mitigation: forbearance, a repayment plan, a modification of the note's terms, or, if the house has to go, a short sale or a deed in lieu. Most delinquencies end here and never become a legal proceeding.

If they do not, the security instrument's acceleration covenant is what makes the next step possible. Acceleration is the mechanism by which a missed payment of \$3,033.72 becomes a demand for the entire outstanding balance — without it, a lender could only ever sue for the payments actually missed, one at a time, for thirty years. The covenant requires notice and a cure period first: the borrower is told what the default is, what it takes to fix it, and by when. That cure right, and the right to reinstate by paying everything past due plus costs, is why a borrower who finds the money late can frequently still stop the process.

Then the two paths diverge, exactly along the mortgage/deed-of-trust line drawn above. In a judicial state the lender files a lawsuit, the borrower may answer and defend, a court enters judgment, and the property is sold under court supervision — a process that can take a very long time. In a non-judicial state the trustee exercises the power of sale granted in the deed of trust, following a statutory notice-and-publication schedule, and no court is involved unless somebody brings it in. Afterward, the questions that vary most by state are whether the borrower has a statutory period to redeem the property after the sale, and whether the lender may pursue a deficiency — the gap between what the sale brought and what was owed — or is barred from doing so.

None of that is a loan officer's job. All of it is the reason the job exists: the entire apparatus of verification you are about to spend thirteen chapters learning is an attempt to make sure this sequence never starts. Verify your own state's procedure, timelines, redemption rights, and deficiency rules with counsel or your compliance department before you describe them to anyone.

⚖️ Compliance Check

Which instrument your state uses, whether foreclosure is judicial or non-judicial, whether an attorney must conduct the closing, and how long a borrower has to redeem after a sale are all state law, and they vary enormously.

This is the first of many places in this book where the honest answer is: learn the structure here, then find out what your state does. Your compliance department knows. So does any closing attorney or title officer in your market, and they will usually explain it for free because it makes their job easier when the loan officer understands it.

Verify current requirements with your compliance department and your state regulator.

1.3 Where the money comes from — the four-party system

Ask a room of new loan officers where the \$365,750 comes from and you will hear "the bank." That is approximately never the complete answer, and understanding why is the difference between an originator who can explain their own pricing and one who can only recite it.

Follow the money backward.

FOLLOWING THE MONEY — one loan, four parties            [constructed teaching example]

   ┌──────────────┐
   │   INVESTOR   │  buys mortgage-backed securities. A pension fund, an insurer,
   │              │  a mutual fund, a foreign central bank. Wants yield, hates risk.
   └──────┬───────┘  Sets, in aggregate, what mortgage money costs today.
          │  supplies capital
          ↓
   ┌──────────────┐
   │  AGGREGATOR  │  Fannie Mae / Freddie Mac / Ginnie Mae issuers and large banks.
   │   / AGENCY   │  Buys loans that meet published guidelines, pools them, and
   └──────┬───────┘  guarantees the securities. Publishes THE RULEBOOK.
          │  buys loans that fit the guidelines
          ↓
   ┌──────────────┐
   │    LENDER    │  Your employer, or the wholesaler behind your broker. Funds the
   │              │  loan at closing, usually with borrowed money, then sells it.
   └──────┬───────┘  Makes money on the sale, not on the interest.
          │  funds the loan at the closing table
          ↓
   ┌──────────────┐
   │   BORROWER   │  Buys the house. Pays for thirty years. Never meets three of
   │              │  the four parties above.
   └──────────────┘

   And running alongside, after closing:
   ┌──────────────┐
   │   SERVICER   │  collects the payment, pays the taxes and insurance out of
   │              │  escrow, handles hardship. Often not the original lender.
   └──────────────┘

Work down that chain and notice what each party actually wants.

The investor wants a predictable return with very little risk of loss. They are not in the housing business and have no opinion about your borrower. They buy a security backed by thousands of loans, and what they require is that those loans behave the way the prospectus said they would. When investors get nervous, they demand more yield; when mortgage yields rise, mortgage rates rise. This happens every day, sometimes several times a day, and Chapter 30 explains what to do about it.

The aggregator or agency is the party that translates the investor's appetite into rules. Fannie Mae and Freddie Mac publish enormous, free, public guides describing exactly which loans they will buy. Ginnie Mae guarantees securities backed by government-insured loans. These entities do not lend money to homebuyers — this is worth repeating, because it is one of the most common misunderstandings in the business and a reliable exam question. Fannie Mae has never made a mortgage loan to a consumer. It buys loans other people made.

The lender is the party that hands over money at the closing table. That money is usually borrowed — on a warehouse line of credit, which Chapter 31 explains — and it is repaid within days or weeks when the loan is sold. This is why your employer cares intensely whether a file meets guidelines: a loan that no aggregator will buy is a loan the lender is stuck holding with borrowed money, and that is how mortgage companies die.

The borrower is the only party in the chain who thinks the transaction is about a house.

🧮 Run the Numbers

Why the lender is not primarily in the interest business.

On the Linden Street loan — \$365,750 at 6.625% — the first month's interest is:

$$\$365{,}750 \times \frac{0.06625}{12} = \$2{,}019.24$$

If the lender held this loan, that \$2,019.24 a month would be its revenue, against the cost of the money it borrowed to fund it. A very thin spread, for thirty years, with all the credit risk.

Instead the lender sells the loan within a few weeks, typically for more than the loan amount — because an investor is buying a thirty-year stream of 6.625% payments on a well-documented loan, and will pay a premium for it. (Careful: this is the lender's execution in the secondary market, not the borrower's rate sheet. Par on the borrower's sheet for this file is 6.750%, which is why they paid half a point to reach 6.625%. Chapter 29 takes the two prices apart.) Suppose the loan sells at a price of 101.500:

$$\$365{,}750 \times 1.01500 = \$371{,}236.25$$

a gain of \$5,486.25, realized in weeks rather than decades, plus the origination charge the borrower paid at closing. The lender then does it again with the next loan.

This is the actual business model of most mortgage lending in the United States, and it explains almost everything else. It is why guidelines are treated as absolute; why a loan that cannot be sold is a catastrophe rather than an inconvenience; and why the lender cares what your rate sheet says this morning, not what it said last week. Chapters 28 and 29 take this apart properly. (Illustrative price; secondary market execution varies daily.)

Where the money physically is on closing day

The four-party diagram is a picture of ownership. Here is the picture of cash, on one day, in one file, and it is worth studying because almost nobody in this business can draw it from memory even though it is the single event they are all paid to produce.

On day 51, money does not move from the lender to the borrower. It never does. The lender wires funds to the closing agent, who is neutral between the parties, and the closing agent disburses. The borrower's own money arrives the same way, by wire, usually a day earlier. Everything then leaves that one account, and the account has to balance to the penny or nothing records.

THE CLOSING TABLE — the buyer's side of the ledger        [the Linden Street file]

  FUNDS IN                                                          AMOUNT
  ─────────────────────────────────────────────────────────────────────────
  loan proceeds, wired by the lender from its warehouse line     365,750.00
  borrower's wire (the "cash to close")                           25,376.34
  earnest money already on deposit with the closing agent          5,000.00
  seller credit toward closing costs                               3,000.00
  ─────────────────────────────────────────────────────────────────────────
  TOTAL IN                                                       399,126.34

  FUNDS OUT                                                         AMOUNT
  ─────────────────────────────────────────────────────────────────────────
  purchase price, to the seller's side of the statement          385,000.00
  closing costs (origination, points, appraisal, credit, flood
    cert, tax service, title, settlement, recording, owner's
    title, survey, pest)                                           9,720.25
  prepaids and escrow deposit (8 days of interest, 12 months
    of homeowners insurance, 5 months tax + 3 months insurance)     4,406.09
  ─────────────────────────────────────────────────────────────────────────
  TOTAL OUT                                                      399,126.34

  Buyer's side only; the seller's payoffs, commissions, and net are on the
  other half of the settlement statement. This file carries no property-tax
  proration — a teaching simplification. Constructed; figures are this book's
  frozen file.

Read down the FUNDS IN column and you can see the whole chapter in it. Ninety-two percent of the money on that table belongs to somebody who is not in the room and will never meet the people signing. The borrower's own contribution — the number they have been terrified about for seven weeks, the number they moved from three accounts and asked their parents for help with — is \$25,376.34 out of \$399,126.34. The seller's \$3,000 credit, negotiated in about ninety seconds by two agents on day 4, covers roughly a third of the loan costs. And the \$5,000 they paid on day 4 comes back to them here, not as a refund but as a credit, because they already spent it.

That last point causes more confusion at the closing table than anything else on the page. Borrowers look at the earnest money line, see a positive number, and think they are getting money back. They are not. They are being given credit for money they parted with seven weeks ago. Say it before they ask.

Why the investor pays what they pay

Here is the mechanism the whole chain hangs from, and it is not the one most people guess.

Ask a new loan officer why mortgage rates are higher than Treasury yields and they will say credit risk — investors charge more because borrowers might default. That is a reasonable guess and, for agency mortgage-backed securities, mostly wrong. The guarantee has already removed credit risk from the investor: if the borrower defaults, the guarantor makes the security's holder whole. The investor is not lying awake about your borrower's 706.

What the investor actually holds is a bond whose issuer can hand the money back at any time, for free, with no notice, and who will reliably choose to do so at the worst possible moment. Your borrower can prepay this loan in full, at par, on any day of the next thirty years — by refinancing, by selling the house, by inheriting money. There is no penalty on this note. That right is worth something, the borrower did not pay for it separately, and somebody has to.

Work the two branches:

WHY A MORTGAGE INVESTOR CANNOT WIN            [constructed teaching example]

  RATES FALL                              RATES RISE
  ──────────────────────────────────      ──────────────────────────────────
  borrowers refinance and prepay          borrowers keep the loan forever
        ↓                                       ↓
  investor gets cash back early            investor is stuck holding a
        ↓                                  below-market bond
  and must reinvest it at the new,               ↓
  LOWER rate                               and cannot sell it except at a
        ↓                                  loss
  the gain they should have had                  ↓
  from falling rates is taken away         the loss from rising rates lands
                                           in full
  ──────────────────────────────────────────────────────────────────────────
  Upside capped. Downside open. That asymmetry has a price, and the price
  is the extra yield mortgage investors demand over a comparable Treasury.

An ordinary bond gains value when rates fall. A mortgage security barely does, because the borrowers whose loans are in it hand the money back precisely then. The investor's upside is capped by the borrower's option and their downside is not capped by anything. So they will not buy at a Treasury yield. They demand a spread, and that spread — not your lender's greed and not your borrower's credit — is the largest single reason the rate you quote is what it is.

Three consequences follow immediately, and all three will show up in your first year:

• Mortgage rates do not track the ten-year Treasury one for one. A borrower who watched the news and calls to ask why their rate did not fall when "rates fell" is asking a real question with a real answer, and the answer is the spread, not evasion.
• When rate volatility rises, the spread widens. An option is worth more when the underlying moves more. That is why mortgage pricing can worsen on a day when the bond market did nothing dramatic, and why your rate sheet can be repriced twice before lunch. Chapter 30 lives here.
• Shorter and less prepayable is cheaper. A fifteen-year loan prices better than a thirty in part because there is less optionality to give away. It is also the honest reason prepayment penalties ever existed — they claw back the option — and why consumer protection law now sharply restricts them on covered consumer mortgages. Chapter 34 draws that line precisely; do not summarize it from here.

Where the 6.625% actually goes

The borrower thinks they are paying 6.625% to the lender. They are paying it to at least three parties, and knowing the split is what lets you talk about pricing without hand-waving.

DECOMPOSING THE NOTE RATE      [constructed teaching example — guarantee fees are
                                negotiated and revised; verify current values]

  6.625%  the borrower's note rate
    −  0.250%   SERVICING FEE — retained by whoever services the loan, out of
                each payment, customarily a quarter point on conventional loans
    −  0.500%   GUARANTEE FEE — paid to the guarantor for standing behind the
                loan's credit performance (illustrative rate)
  ─────────
  = 5.875%  NET RATE passed through toward the security's coupon

  ON THIS LOAN'S OPENING BALANCE OF $365,750, FOR ONE YEAR
    servicing        0.250%  ×  $365,750  =     $914.38
    guarantee fee    0.500%  ×  $365,750  =   $1,828.75
    to the investor  5.875%  ×  $365,750  =  $21,487.81
  ───────────────────────────────────────────────────────
    total interest   6.625%  ×  $365,750  =  $24,230.94

  Pools are formed at coupons set in half-point steps, so the remainder
  between a loan's net rate and its pool's coupon is handled by the
  aggregator. Chapter 28 does that mechanically.

Two honest footnotes on that box, because the book's rule is that every number resolves. The annual figures are computed on the opening balance, before any amortization; the loan's actual first twelve months of interest is smaller, because the balance falls a little every month. You can get that number from figures already in this book: twelve payments of \$2,341.94 total \$28,103.28, the balance falls from \$365,750.00 to \$361,757.88 over the same twelve months, and the difference — \$28,103.28 − \$3,992.12 — is \$24,111.16 of interest. The split is unchanged; the base shrinks.

And notice what the decomposition does not contain: a line for your employer's profit on the interest. There isn't one, which is the point §1.3 has been making. Your employer's revenue is the origination charge and the gain on sale, both collected within weeks. The 6.625% belongs to the people who hold the loan afterward.

Learning to price in your head

One arithmetic habit is worth building on your first day, because it makes every pricing conversation faster and it stops you from saying expensive things by accident. A point is one percent of the loan amount, and on any given file you should know what the ladder looks like without a calculator.

THE POINT LADDER ON A $365,750 LOAN                       [the Linden Street file]

  1.000 point   =  1.000%  ×  $365,750  =  $3,657.50   ← this file's origination charge
  0.500 point   =  0.500%  ×  $365,750  =  $1,828.75   ← this file's discount point
  0.250 point   =  0.250%  ×  $365,750  =    $914.38   ← this file's lock extension
  0.125 point   =  0.125%  ×  $365,750  =    $457.19   ← the smallest step most
                                                          rate sheets quote

  And in the other direction, the payment:
  $2,341.94 ÷ 365.75 thousands = $6.4031 per $1,000 borrowed, at 6.625% for 30 years.
  So $6.40 per thousand is close enough for a phone call: 365.75 × $6.40 = $2,340.80,
  about a dollar light. Never quote the estimate as the payment.

Three of those four lines are figures that actually appear in this file, which is not a coincidence — it is what a rate sheet's units look like once you can see them. An eighth of a point on this loan is \$457.19. A quarter is \$914.38, which happens to be exactly what the fifteen-day lock extension cost on day 42 when the original lock ran out. Chapter 30 has a great deal to say about how that came to happen and who paid for it; for now, note only that a quarter of a point is not an abstraction. It is nine hundred and fourteen dollars, and somebody wrote a check.

The payment-per-thousand shortcut is the one you will use most. It lets you answer "roughly what would that be a month?" in three seconds without opening anything, and it lets you sanity-check a number your software just produced. Software is wrong more often than you would like, usually because somebody typed a term or a rate into the wrong field, and a loan officer who can see that \$2,341.94 is about right for \$365,750 at 6.625% will catch it. One who cannot will send it to the borrower.

What each party is afraid of

One last pass down the chain, from a different angle, because it turns the whole apparatus into something you can reason about instead of memorize. Every requirement you will ever be asked to satisfy is somebody's fear, written down. Learn the fears and most of the rulebook stops feeling arbitrary.

Read across the top row and then the bottom one. The investor's fear and the borrower's fear are the same fear expressed at opposite ends of a very long pipe: somebody is going to be surprised, and it is going to cost money. Everything in the middle — the guidelines, the conditions, the documents, the disclosures, the second underwrite, the callback on the wire — exists to move surprises earlier, where they are cheap.

That is also the most useful thing you can do with a condition you think is stupid. Before you argue it, ask whose fear is this? A condition without an answer to that question is worth escalating. A condition with one is worth clearing, and once you can name the fear out loud you can explain it to a borrower in a sentence, which is most of what this job's difficult conversations consist of.

1.4 Primary market and secondary market

Now name the two halves of that chain.

The primary market is where a loan is made: a borrower and a lender, an application, an underwriting decision, a closing. You work in the primary market. Everything in Parts II, III, and IV of this book is primary-market activity.

The secondary market is where a loan is sold: lender to aggregator, aggregator to security, security to investor. You do not work in the secondary market, but it sets your prices, writes your guidelines, and determines which of your borrowers can be served at all.

PRIMARY vs. SECONDARY                                  [constructed teaching example]

   PRIMARY MARKET                    │        SECONDARY MARKET
   "making the loan"                 │        "selling the loan"
   ──────────────────────────────────┼──────────────────────────────────────
   borrower ←→ loan officer          │   lender → aggregator → security
   application, credit, appraisal    │   pooling, guarantee, TBA trade
   underwriting decision             │   investor purchase
   closing and funding               │   servicing retained or released
   ──────────────────────────────────┼──────────────────────────────────────
   YOU ARE HERE                      │   SETS YOUR PRICE AND YOUR RULES

The relationship between the two is the most important structural fact in American mortgage lending, and it is not obvious. In most countries, a bank lends its depositors' money and keeps the loan. The bank's own risk tolerance determines who gets a mortgage.

In the United States, a lender can make a loan and sell it within a month, replacing its capital and lending again. That arrangement is why thirty-year fixed-rate mortgages exist here at scale — no sane depository would voluntarily hold thirty years of fixed-rate interest-rate risk funded by deposits that can leave tomorrow. It is also why the guidelines are published in advance and uniform: a security is only sellable if the buyer knows what is in it.

The consequence for you is direct. Your borrower's eligibility is not determined by your employer's opinion of them. It is determined, mostly, by a rulebook written by whoever is going to buy the loan — with your employer's own additional rules layered on top. Chapter 14 names that distinction (guideline versus overlay), and it is one of the most useful things in this book.

What a mortgage-backed security actually is

Strip the jargon and it is a simple object. Take several hundred mortgage loans with similar characteristics — same product, similar note rates, all meeting the same published guidelines. Put them in a pool. Sell undivided shares of that pool to investors. Every month, the payments the borrowers make flow into the pool and are passed through to the shareholders in proportion to what they hold, less the servicing fee and the guarantee fee. When a borrower prepays, that money passes through too, as a return of principal.

That is it. No leverage, no tranching, no exotic structure. An agency pass-through security is a pipe with a guarantee wrapped around it.

Two features of that object explain most of what the primary market does to you.

First, the pool has to be describable before it is full. An investor buying a share of loans that have not been originated yet needs to know what will be in it. That requirement — not regulatory zeal — is why the guidelines are published in advance, uniform across lenders, and enforced absolutely. A guideline is a product specification. When your underwriter refuses to waive a condition, they are not being rigid; they are refusing to put something in a box that was sold on the promise of what is in the box.

Second, individual loans stop mattering and portfolios start. No investor reads your file. Nobody will ever look at your borrower's letter of explanation about the \$4,900 deposit. What the investor buys is a statistical promise about several hundred households, and the guarantee makes even that promise about timing rather than about credit. This is genuinely liberating once you see it: your job is not to persuade anybody that your borrowers are good people. It is to make the file indistinguishable from what was promised. Those are different tasks, and only one of them is achievable.

Ginnie Mae is not Fannie Mae, and the difference is on the exam

§1.3 named three entities in one box, which is convenient and slightly misleading. They do different things, and candidates lose points on this every year.

Fannie Mae and Freddie Mac buy loans. They are government-sponsored enterprises — chartered by Congress, privately structured, and in federal conservatorship since 2008. They purchase closed loans that meet their published guidelines, pool them, issue securities backed by those pools, and guarantee the securities. Their guidelines are the conventional rulebook, and Chapter 14 is built on them.

Ginnie Mae does not buy loans at all. It is a government corporation within the Department of Housing and Urban Development, and it does not purchase mortgages, does not issue securities, and has no guidelines of its own about who qualifies. What it does is guarantee the timely payment of principal and interest on securities issued by approved private issuers, where every loan in the pool is already insured or guaranteed by a government program — FHA, VA, or USDA.

So the layering is different, and that is the whole point:

TWO DIFFERENT SHAPES OF GUARANTEE           [constructed teaching diagram]

  CONVENTIONAL                          GOVERNMENT
  ─────────────────────────────         ─────────────────────────────
  lender originates                     lender originates
        ↓                                     ↓
  Fannie/Freddie BUY the loan           FHA insures / VA guarantees the
        ↓                               LOAN itself (first layer)
  pool it, issue the security,                ↓
  and GUARANTEE the security            an approved issuer pools the loans
        ↓                               and issues the security
  investor                                    ↓
                                        Ginnie Mae GUARANTEES THE SECURITY
                                        (second layer, full faith and credit)
                                              ↓
                                        investor
  ─────────────────────────────────────────────────────────────────────
  One guarantee on the conventional side. Two layers on the government
  side — the loan and then the security — and Ginnie Mae never owns a loan.

The exam framing is usually a single sentence: which of these does not purchase mortgage loans? The answer is Ginnie Mae. The practitioner framing is more useful — when an FHA borrower asks who is backing their loan, the honest answer involves both FHA and Ginnie Mae, and they are doing different jobs at different points in the chain.

The rulebook is free, and you should download it

One habit separates loan officers who can answer a question from loan officers who can only forward it.

The guidelines are public. The Fannie Mae Selling Guide and the Freddie Mac Seller/Servicer Guide are published, searchable, and free. HUD Handbook 4000.1 covers FHA. The VA and USDA publish theirs. These are not summaries or vendor cheat sheets — they are the actual documents your underwriter is working from, and reading the paragraph in question takes about ninety seconds.

Three rules for using them, learned the hard way:

Search the guide, not the internet. A forum post about what Fannie Mae requires is a description of what somebody remembered about a version of the guide that may no longer exist. The guides are updated continuously, and each has an effective date and a change log. Check the date on anything you rely on.

When your underwriter cites a section, read it. Not to argue — to learn. Half the time you will find that the requirement is narrower than the condition implied, which is a legitimate basis for a written exception request. The other half you will understand why the requirement exists, which is worth more, because you will structure the next file to avoid it.

Remember that the guide is not the whole answer. The agency guideline is the floor; your lender's overlays sit on top of it, and only your own credit policy department can tell you the combined rule. A loan officer who quotes the Selling Guide at an underwriter enforcing a company overlay has misunderstood the argument they are in. Chapter 14 makes that distinction precisely, and it is one of the most useful things in this book.

Why you can promise a rate for thirty days

This one puzzles nearly everybody, including experienced loan officers who have never asked. On day 12 you told two people they had 6.625% for thirty days. The loan did not exist. Underwriting had not seen it. The appraisal had not come back. Where did the promise come from?

It came from the secondary market's forward market. A lender can sell a mortgage-backed security before the loans in it exist, agreeing today to deliver, on a future settlement date, a pool with specified characteristics — coupon, term, agency — but not specified loans. The pool is to-be-announced. That forward sale is what lets your employer take the other side of your borrower's rate: the lender has already sold the rate it just gave away, so when the market moves, the loan and the hedge move together and cancel.

Three practical consequences you will meet before your third file:

• A lock is a two-sided commitment, and only one side is bound. Your borrower may walk away; the lender's forward sale does not walk away with them. That gap is called fallout, and it is managed with money — which is why "just re-lock me at today's lower rate" is never free and why renegotiation policies exist and vary by lender.
• An expiring lock costs real money to extend, because the hedge has to be rolled forward. On this file, the thirty-day lock taken on day 12 expired on day 42 and was extended fifteen days at 0.250 point — \$914.38 — which the lender paid rather than the borrower. Chapter 30 owns the full critique of how that situation arose, and it is worth waiting for.
• The lock desk is a real desk with real exposure. When you ask for a floating-to-locked conversion or a rate concession, you are asking a person with a position to take a loss. Ask through the right channel, with the file's facts attached, and you will get a hearing. Ask by complaining, and you will not.

What happens when a loan cannot be sold

The chapter has asserted twice that an unsaleable loan is a catastrophe. Here is the actual machinery, because it is the source of the conditions on your approvals and it explains behavior that otherwise looks paranoid.

When a lender sells a loan, it makes representations and warranties: the income was documented this way, the appraisal met these standards, the disclosures went out on time, the borrower occupied the property. Those promises survive the sale. If the buyer later finds that one of them was not true, it can issue a repurchase demand — return the money, take the loan back. The loan comes back onto the lender's books at par, frequently at a moment when it is worth considerably less than par, and now the lender owns a loan that may also be delinquent.

There is a second, faster version. Many purchase agreements include an early payment default provision: if the borrower misses one of the first few payments, the buyer can put the loan back with no argument about whether anything was misrepresented. That single clause is the reason for two things you will otherwise find inexplicable. It is why the lender re-verifies employment within days of the note date rather than relying on the verification from week two — a borrower who was laid off on day 47 is an early payment default waiting to happen. And it is why a pre-closing credit refresh exists at all.

Look at what that produced on this file. Condition 11 on the day-28 approval was a pre-closing credit refresh with a ceiling on the approved debt-to-income ratio. It was written sixteen days before the borrowers financed \$5,200 of furniture on day 41 — an event nobody could have predicted, by people who had done nothing wrong and did not know they were doing anything at all. The condition caught it on day 44. That condition is not bureaucratic residue. It is an early-payment-default clause, walked backward through the process until it became a line on a stip sheet.

A loan that fails review before the sale has softer outcomes: it can be cured, or it can be sold into the scratch-and-dent market — the market for loans with a defect — at a discount that comes straight out of the lender's margin. None of these outcomes is good, and every one of them is somebody's fault. This is the emotional physics of underwriting departments, and understanding it will make you a better colleague than half the loan officers you will meet.

🔍 Check Your Understanding

1. Which document does the borrower sign that would still be a valid obligation if they owned no property at all?
2. A lender funds a loan on Tuesday and sells it three weeks later. In which market did each event occur?
3. Fannie Mae denies your borrower. Is that sentence possible? Why or why not?

(3 is the interesting one. Fannie Mae does not decide on individual borrowers in the way the sentence implies — an automated underwriting system evaluates the file against Fannie's published guidelines and returns a recommendation, and your lender decides whether to lend. Chapter 15.)

1.5 Who does what: originator, processor, underwriter, closer, servicer

A file passes through five sets of hands. Borrowers believe they are dealing with one person — you — and one of your quieter jobs is making that belief true enough to be useful without being false.

The loan officer (mortgage loan originator)

You. You find the borrower, take the application, advise on structure, and own the relationship and the timeline. Legally, a mortgage loan originator is someone who takes a residential mortgage loan application or offers or negotiates terms of a residential mortgage loan for compensation or gain. That definition is from the S.A.F.E. Act and it is precise on purpose; Chapter 3 explains why the wording matters and who is covered.

You are the only person in the chain the borrower chose.

The processor

The processor assembles the file. They order the appraisal, the title work, the verifications of employment and deposit, the flood certification, and the payoffs; they chase documents; they review what comes back for completeness and internal consistency; and they submit to underwriting.

A good processor is worth more to a loan officer's income than almost anything else they could buy. A processor who catches on day six that the borrower's paystub year-to-date does not support the income you calculated has saved a file that would otherwise have died on day thirty.

Processors do not have authority to approve anything. They also, in most shops, do not have a license, which has a specific consequence: there is a line they may not cross into offering or negotiating terms, and Chapter 3 draws it.

How to actually work with a processor

Since the claim above is that a good processor is worth more to your income than almost anything you could buy, it is worth saying what "working well with one" concretely consists of. Six practices, and loan officers who follow them get their files worked first — not because of favoritism, but because a complete file is faster to work and everyone's queue is finite.

Submit complete, or say plainly that you have not. The single most destructive habit in this business is submitting a partial file to make a milestone look met. It does not accelerate anything; it converts one review into three and moves your file to the bottom of a queue twice.

Hand over the story, not just the documents. Three sentences at the top of the file: what this borrower does, where the money comes from, and what you already know is going to be a question. "B2 is commissioned and paid quarterly, so the deposits are lumpy and there will be a large one. Gift from B1's parents, letter is coming Thursday. Reserves are thin after closing." A processor who knows where the trouble is will find it on day six instead of day thirty.

Own your own conditions. Borrower conditions are yours. It is not the processor's job to chase your borrower for a paystub, and every hour they spend doing it is an hour not spent on the title work only they can move.

Never let them learn something from the borrower first. If the borrower tells you on Tuesday that they are changing jobs, the processor hears it Tuesday. Information that reaches operations through the borrower instead of through you is information you have already lost control of.

Escalate to help, not to blame. "Can we get this to underwriting today, or should I reset the agent's expectation?" is a question that gets answered. "Why isn't this submitted yet" is a question that gets a defensive answer and a slower file.

Tell them when a file is dead. Loan officers hate this and let dead files sit in a pipeline for weeks out of optimism. Every one of those files is occupying somebody's queue.

The underwriter

The underwriter decides. They evaluate the file against the applicable guidelines and the lender's overlays and issue a decision: approved with conditions, suspended, or denied. In practice almost every approval is a conditional approval — approved provided a specified list of items is delivered and satisfies them.

The underwriter is not your adversary, though a first-year loan officer will experience them that way. They are the person who has to certify that this file is what the investor was told it would be, under representations and warranties that can require their employer to buy the loan back years later if it was not. Chapter 14 explains reps and warrants, and it will substantially improve your relationship with underwriting.

Underwriters rarely speak to borrowers. This is deliberate.

Why the underwriter never calls

New loan officers assume this is a staffing decision, or rudeness, or a company being cheap with people's time. It is none of those. There are four reasons and they are all good ones.

The decision has to rest on the file. An underwriter who forms an impression from a conversation has introduced something into the decision that is not documented, cannot be reviewed, and cannot be relied on by anyone downstream. If it is not in the file, it did not happen — and a fact that was "explained on a call" is a fact that will be missing when a quality control auditor opens this loan in two years.

It is a fair-lending protection, and this is the reason that matters most. A voice on a phone carries information an underwriter is not permitted to consider: accent, apparent age, whether children are audible in the background, how fluently somebody speaks English, how anxious they sound. Nobody has to intend anything improper for that information to influence a judgment. The cleanest way to keep it out of the decision is to keep the decision-maker out of the conversation. Chapter 25 develops this seriously; here, notice that a structure that looks bureaucratic is in fact a deliberate protection for exactly the borrowers who are most likely to be harmed without it.

It protects the borrower from talking themselves into trouble. A frightened person, asked an open question by the person who decides, will explain. In the explaining they will volunteer things that are irrelevant, imprecise, or actively harmful — a plan to change jobs, a relative's money, a side arrangement with the seller. A written condition asks for a specific document; a phone call invites a narrative, and narratives create conditions.

It keeps one version of the truth. If the underwriter, the processor, and you are all talking to the borrower, the file acquires three accounts of the same fact. One channel, documented, is how a file stays coherent across fifty-one days.

So the underwriter's silence is not distance. It is the discipline that makes the decision reviewable — and it is also exactly why the translation job in §1.5 falls to you and cannot fall to anybody else.

The closer and the closing agent

Two different roles, often confused.

The closer works for the lender. Once the file is clear to close, the closer prepares the closing package and the Closing Disclosure, coordinates figures with the settlement agent, and authorizes funding.

The closing agent — a title company, escrow company, or in some states an attorney — is neutral between buyer and seller. They conduct the signing, hold and disburse funds, and record the documents. They work for the transaction, not the lender.

The wire, and the fraud that targets it

The closing agent's other function is that they are the one party in the transaction who is holding everybody's money at once, which makes the moment your borrower sends their cash to close the single most dangerous moment in the whole fifty-one days.

Real estate closings are a standing target for wire fraud, and the pattern is consistent enough that you can inoculate a borrower against it in ninety seconds. The essential shape: somewhere in the transaction an email account is compromised — it might be the buyer's, the agent's, or the settlement office's — and shortly before closing the buyer receives what looks like a routine message from a name they recognize, containing revised wiring instructions. The money goes where the instructions say. By the time anyone notices, it is frequently gone and frequently unrecoverable, and the family that just lost their down payment has no house and no recourse.

Nothing about that requires anyone to be careless. It requires only that somebody's email was read and that a family in the most stressful week of their financial life received an instruction that looked exactly like the twenty instructions they had already received.

What a competent loan officer does about it costs one conversation, on day 5, at application, while nobody is under time pressure:

• Say it out loud, early: "Wiring instructions for your closing will never change by email. If you ever receive an email changing them — even from me, even from your agent, even from the title company — assume it is fraudulent."
• Give them the callback rule and a number to call. They call the settlement office at a number they obtained independently, earlier, from a source that is not the message in question, and they confirm every digit of the account with a human being before sending anything.
• Tell them urgency is the tell. Legitimate closings do not require money to move in the next twenty minutes.
• Tell them to call you if anything at all seems off, at any hour, including on a Friday afternoon.

Say all of it again the week of closing. It is repetitive, it feels slightly insulting to a competent adult, and it is one of the most protective things a loan officer does in the entire transaction.

The servicer

After closing, the servicer takes over: collects the monthly payment, maintains the escrow account and pays the property taxes and homeowners insurance out of it, sends the annual statements, and handles delinquency, loss mitigation, and payoff.

The servicer is frequently not the lender whose name was on the note, and servicing can be transferred more than once over a loan's life. To the borrower this feels like something going wrong. It is not. Chapter 23 covers the transfer notices, and Chapter 28 explains why servicing has its own market and its own value.

What servicing is worth, and why anyone would buy it

Servicing sounds like a chore. It is an asset, it is bought and sold, and understanding its economics explains a whole category of borrower experience that otherwise looks random.

The servicer is paid out of the interest the borrower is already paying — customarily a quarter of a percent a year on a conventional loan, taken off the top before the rest passes through. On this file that is \$914.38 in the first year, falling a little each year as the balance amortizes. For that money the servicer must run a payment system, a call center, an escrow administration that correctly pays a tax bill in one county and an insurance premium to one carrier, an annual escrow analysis, a year-end statement, and a loss mitigation department.

Now the part that makes it interesting. When a borrower stops paying, the servicer generally must keep advancing the payment to the investor anyway, out of its own pocket, along with the taxes and the insurance, until the loan performs again or the loss is resolved. The investor's guaranteed timely payment is guaranteed partly because somebody is fronting it.

So the arithmetic of servicing is stark: a performing loan is a small, dependable annuity; a delinquent one is a cash outflow plus a very expensive human being on a phone. That produces effects your borrower will experience directly:

• Servicing is worth more when rates rise, because the loans in the portfolio are less likely to be refinanced away and the annuity lasts longer. It is worth less when rates fall. This is why servicing portfolios change hands in waves, and why transfers cluster.
• The escrow account is administered strictly, because the servicer is the one who eats an unpaid tax bill.
• A servicer's economics reward volume and standardization, which is why the borrower who calls with a genuinely unusual question gets a scripted answer — and why the loan officer who is still willing to take that call three years later is remembered.

The last point is the commercially useful one. The servicer is contractually obligated to your borrower and structurally uninterested in them. You are structurally uninterested and not obligated at all. Which of you they call in year three is entirely a matter of who behaved like the other thing.

WHO TOUCHES THE FILE — and who ever speaks to the borrower

  ROLE            WORKS FOR      DECIDES?     TALKS TO BORROWER?
  ─────────────────────────────────────────────────────────────────────────
  loan officer    the lender     no           YES — constantly
  processor       the lender     no           sometimes, for documents
  underwriter     the lender     YES          almost never
  closer          the lender     no           rarely
  closing agent   the deal       no           YES — at signing
  servicer        the investor   no           YES — for thirty years
  ─────────────────────────────────────────────────────────────────────────
  Note who decides and who talks. They are almost disjoint sets. That gap
  is where a loan officer's real value lives.

That last line is the point of the section. The person who decides never meets the borrower, and the person the borrower trusts cannot decide. A loan officer's job is to stand in that gap: to present a file that answers the underwriter's questions before they are asked, and to translate the underwriter's requirements into something a frightened human being can act on by Thursday.

📞 On the Phone

Borrower: "Why do they need a letter explaining where the \$4,900 came from? It's my money. It's in my account."

The unhelpful answer: "It's just what underwriting requires."

The answer that works: "Because from their side of the desk, a deposit that appeared six days before you applied looks exactly like a loan somebody made you — and a loan would change your debt-to-income and they'd have to count it. They're not doubting you. They literally cannot tell the difference between your commission check and a loan from your brother unless we show them. Send me the commission statement and the deposit slip and this is over in an hour."

Every good version of this conversation has the same shape: name what the underwriter can't see, then hand the borrower a specific, finite task. Chapter 12 works the large-deposit rule in full; Chapter 19 covers condition-clearing as a discipline.

The people who touch the file and never appear on that table

Five sets of hands is the structure. It is not the population. A residential purchase file is worked by ten to fifteen people and four or five separate companies, most of whom the borrower never hears of and several of whom your own employer does not control.

THE ORG CHART THE BORROWER NEVER SEES                  [constructed teaching example]

                        ┌──────────────────────┐
                        │   YOU (the LO)       │  the only person they chose
                        └───────────┬──────────┘
                                    │
      ┌────────────────┬────────────┼────────────┬──────────────────┐
      ↓                ↓            ↓            ↓                  ↓
 ┌─────────┐   ┌─────────────┐ ┌─────────┐ ┌──────────┐    ┌───────────────┐
 │PROCESSOR│   │ UNDERWRITER │ │ CLOSER  │ │ LOCK DESK│    │ POST-CLOSING  │
 └────┬────┘   └──────┬──────┘ └────┬────┘ └──────────┘    │ / QC / SHIPPING│
      │               │             │       prices and     └───────────────┘
      │               │             │       hedges the      finds the defect
      │               │             │       lock            AFTER the fact
      │               │             │
      │  ORDERS FROM OUTSIDE THE COMPANY (your employer does not control these)
      │               │             │
      ├──→ APPRAISAL: ordered through an appraisal management company or an
      │    independent panel. YOU MAY NOT SELECT OR INFLUENCE THE APPRAISER.
      ├──→ TITLE: examiner, title officer, closing agent — often one company
      ├──→ CREDIT VENDOR: the tri-merge report and any rescore
      ├──→ FLOOD DETERMINATION VENDOR
      ├──→ VERIFICATION VENDORS: employment, income, deposits
      ├──→ MORTGAGE INSURANCE UNDERWRITER: a separate approval, on top of yours
      └──→ THE BORROWER'S OWN INSURANCE AGENT: the one link in this chain that
           nobody in the transaction can order, chase, or expedite

   AND ON THE OTHER SIDE OF THE DEAL: the buyer's agent, the listing agent,
   the seller, the seller's lender's payoff department, and in attorney
   states, two attorneys.

Three of those deserve a note, because each is a place where new loan officers reliably do damage.

The appraiser is walled off from you on purpose. After 2008 the rules were rewritten so that the person producing the value cannot be selected, coached, pressured, or rewarded by the person whose commission depends on the number. In practice this means orders route through an appraisal management company or an independent assignment process, and the correct response to a value you disagree with is a reconsideration of value submitted through the required channel with comparable sales attached — never a phone call. Chapter 18 handles the mechanics; verify your lender's process and the current independence requirements with compliance, because the details are prescriptive and the penalties for getting them wrong are personal.

Mortgage insurance is a second underwrite. On a loan above eighty percent loan-to-value, a mortgage insurance company evaluates the file too, and it can decline what your lender approved. On this file that was condition 9 — a mortgage insurance certificate at the approved coverage and factor — and it cleared on day 33 without drama. It does not always.

The borrower's insurance agent is the only vendor nobody can push. Your processor can escalate a title order and lean on an appraisal management company. Nobody can make a borrower's insurance agent call them back. Homeowners insurance evidence was condition 8 on this file. Ask for it on the day the application is taken, not the week before closing, because it is the single item on the list whose timing you have no leverage over at all.

Who to call when a file is in trouble

Here is the rule, and it is worth more than most of what people are taught in their first year: call the person who can change the answer, not the person who delivered it.

New loan officers violate it constantly, usually by arguing with a processor about a condition the processor did not write and cannot waive. It feels like doing something. It accomplishes nothing, and it burns the one relationship that saves your files week after week.

The map:

Two habits attach to that table. Escalate with a date attached. "This is taking too long" is noise; "the contract closes on the fourteenth and the title commitment was ordered nine days ago" is a request somebody can act on. And never make the borrower carry a message between two departments of your own company. If the underwriter and the mortgage insurance underwriter want different things, that is your problem to reconcile, not a puzzle to forward to a frightened household by email at nine at night.

1.6 The three business models in one paragraph each

The same loan, to the same borrower, at the same rate, can be originated three structurally different ways. The difference is whose money funds it and whose name is on the note — and it changes your job in ways nobody explains at the interview. Chapter 31 does this properly; here is the shape.

Retail. You work for the lender. The company you work for takes the application, underwrites it, funds it with its own money or its warehouse line, closes it in its own name, and sells it. You have one set of guidelines, one underwriting department, one rate sheet, and one price. Simplest, and the model most new loan officers start in. Your product menu is whatever your employer offers, which is its main limitation.

Broker. You work for a mortgage brokerage, which does not lend. You take the application and then place it with a wholesale lender — one of many you are approved with — who underwrites, funds, and closes it in their name. Your advantage is choice: if one lender's overlay kills a file, another's may not, and pricing varies. Your cost is complexity — multiple portals, multiple guideline sets, multiple submission processes — and less control over turn times, because you are a customer of the underwriting department rather than a colleague of it.

Correspondent. Your employer is a lender that underwrites and closes loans in its own name using its own warehouse line of credit, then sells the closed loans to investors — sometimes to the agencies directly, sometimes to larger aggregators. It looks like retail from the borrower's side and like a small wholesale operation from the capital-markets side. Correspondents get more pricing control than brokers and more product flexibility than pure retail, at the cost of needing real capital and a secondary marketing operation.

There is a fourth distinction that cuts across all three and matters more than any of them for your licensing: whether your employer is a depository institution (a bank or credit union) or a non-bank. If you originate for a depository, you are generally registered rather than licensed — no state license, no SAFE MLO test, no continuing education, but also no license to take with you if you leave. Chapter 3 explains this thoroughly, and it is the most consequential career decision most new originators make without knowing they are making it.

The same file, three ways

Take the Linden Street file exactly as it is — \$385,000 purchase, 5% down, \$365,750 conventional thirty-year fixed at 6.625% — and run it through each model. The borrower's payment is identical in all three. Almost nothing else is.

ONE FILE, THREE STRUCTURES                                [the Linden Street file]

                        RETAIL            BROKER              CORRESPONDENT
  ────────────────────────────────────────────────────────────────────────────
  who takes the app     you               you                 you
  whose 1003 system     your employer's   the wholesaler's    your employer's
                                          portal
  who prices it         your employer's   you shop several    your employer's
                        rate sheet        wholesale sheets    secondary desk
  who underwrites       your employer     the wholesaler      your employer
  whose guidelines      one set           one set PER         one set, chosen
                        + overlays        LENDER              by your employer
  who orders the        your employer     the wholesaler,     your employer
    appraisal                             usually
  whose money at        your employer's   the wholesaler's    your employer's
    closing                                                   warehouse line
  whose name on the     your employer     the wholesaler      your employer
    note
  who the borrower      "my lender"       the wholesaler      "my lender"
    sees on the CD                        they never met
  where the loan goes   sold after        already the         sold after
                        closing           wholesaler's        closing
  ────────────────────────────────────────────────────────────────────────────
  Constructed comparison; specific practices vary by company.

The row that surprises borrowers is the one about the name on the Closing Disclosure. In a brokered transaction the borrower has spent fifty-one days talking to you and then sits down at a table where the creditor named on the paperwork is a company they have never heard of and cannot picture. Handle it on day 5, not on day 48: "You're working with me, and I'm placing your loan with a lender called X. Their name will be on your documents. That's normal, and here's why I chose them for you." Said early, that is a sign of expertise. Said at the closing table, it sounds like something you were hiding.

The row that matters most to you is guidelines. A retail loan officer learns one rulebook deeply and can answer most questions from memory. A broker learns the shape of many rulebooks and learns where they differ — which is a genuinely different skill and takes longer, but it is what makes a broker useful. Consider a constructed but entirely ordinary case: a file with variable income and reserves of a little over four months of the new payment. Suppose one wholesale lender's overlay requires six months of reserves when more than twenty percent of qualifying income is variable, and another follows the agency guideline as written. Same borrower, same agency rulebook, same application — declined at one, approved at the other. Nobody did anything wrong. The overlay is not a guideline; it is a company's private caution layered on top of one, and Chapter 14 gives that distinction the treatment it deserves.

How to tell which one you are working for

This sounds like it should be obvious and, from inside the building, frequently is not. Companies describe themselves generously, and job postings use "lender" for all three. Four questions settle it, and you should ask them in an interview:

1. Whose name is on the note at closing? This is the definitive test. If it is your employer's, you are retail or correspondent. If it is another company's, you are a broker.
2. Who underwrites the file, and are they employees of the company paying me? If your underwriting decision comes from a portal belonging to someone else, you are a broker regardless of what the sign says.
3. Does the company have a warehouse line, and does it sell closed loans? That is the correspondent's signature. A pure retail shop owned by a depository may hold loans instead.
4. How many wholesale lenders am I approved with? One is not a menu.

The reason to care is not tidiness. It is that the three models fail in different ways and reward different behavior, and knowing which building you are standing in tells you where your files will die. A broker's files die in the gaps between portals, on submissions that were formatted for the wrong lender. A retail loan officer's files die on the one overlay their company has and cannot route around. A correspondent's files die in the secondary marketing operation, when a product the desk was buying last month is no longer being bought.

The account executive, and what a wholesale relationship actually is

If you ever work on the broker side, one relationship will shape your week more than any other and nobody explains it at the interview.

Each wholesale lender you are approved with assigns you an account executive — a salesperson whose job is to win your loans. That is worth stating plainly, because it explains both what an account executive is enormously useful for and what they cannot do. They can tell you before you submit whether their guidelines fit a file, which is the most valuable ninety-second phone call available to a broker. They can find out where a file is sitting. They can sometimes get a second look at a decision, or a pricing exception on a file worth having. They can teach you a lender's quirks faster than any portal documentation will.

What they cannot do is approve anything, and a broker who mistakes an account executive's encouragement for an underwriting decision will eventually promise a borrower something that does not survive contact with the actual guideline. The discipline is simple: get it in writing, from the lender's own guideline or from an underwriter, before it goes in a letter or on a phone call with an agent.

The other thing worth understanding is your position in the relationship. A broker is a customer of the underwriting department, not a colleague of it — which cuts both ways. You cannot walk down the hall. But you also have the one thing retail loan officers do not: if a lender's service becomes unworkable, you can send your next twenty files somewhere else, and everyone involved knows it.

The exception that proves the rule: the loan nobody sells

Everything above assumes the loan gets sold. Some do not, and the exception is worth knowing because it is where the hardest files sometimes find a home.

A portfolio loan is one the lender makes with its own money and keeps on its own balance sheet. Because nobody downstream is buying it, nobody downstream writes the rules — the lender's own credit committee does. That single fact reverses the chapter's central argument in a narrow and useful way. The borrower's eligibility here really is determined by your employer's opinion of them, and a banker who knows a depositor of twenty years can act on knowledge that is invisible to an automated underwriting system.

Depositories are the usual home for this: a community bank or credit union lending its members' deposits. So are lenders serving borrowers the agency rulebook simply does not describe — an unusual property, a complex ownership structure, an income stream that is genuinely stable and genuinely undocumentable by the standard method.

Two cautions, and they are not small. Portfolio lending is rarely cheaper, because the lender is holding the interest-rate risk that the secondary market otherwise absorbs, and it charges for it. And "portfolio" is not a synonym for "no rules" — the Ability-to-Repay requirements apply to the loan regardless of who ends up holding it, and a lender that keeps a loan has more exposure to a borrower who cannot pay, not less. Chapter 34 develops what sits outside the agency box; the point here is only that the box has an outside, and that "the guidelines say no" and "nobody will lend on this" are not the same sentence.

1.7 What a loan officer is actually paid to do

We can now answer the question the chapter opened with.

A loan officer is paid to convert an unverified household into a saleable file.

That is the whole job, stated as narrowly as possible. Every task decomposes into it:

• Finding the borrower — there is no file without one, and they do not arrive on their own. (Chapters 7, 38)
• Choosing a structure that fits — because a file that does not fit the guidelines is not saleable, and a structure the borrower cannot live with will not be paid. (Chapters 5, 13)
• Documenting every claim — because an undocumented claim is not saleable at any price. (Chapters 10, 11, 12)
• Managing the calendar — because the contract has dates, the lock has an expiration, and both cost money. (Chapters 19, 30, 39)
• Complying — because a file with a disclosure violation may be unsaleable or may cost a cure, and a pattern of them ends the license. (Chapters 22, 24, 25, 26)
• Telling the truth early — because the alternative is telling it late, when it is expensive. (Chapters 8, 18)

Notice what is not on that list. "Getting the lowest rate" is not on it. You do not set rates. You cannot beat a competitor's price by wanting to. What you can do — and what a competitor with a lower advertised number frequently cannot — is build a file that actually closes at the price you quoted.

This is the book's first theme, and it is not a slogan. Consider what happens on the call we started with.

🧮 Run the Numbers

What a quarter point is worth, and what a failed transaction costs.

On a \$365,750 loan, the difference between 6.625% and 6.375% is:

Rate	Monthly P&I
6.625%	\$2,341.94
6.375%	\$2,281.80
Difference	\$60.14/month

Sixty dollars a month is real money — \$721.68 a year, and over the first five years, \$3,608.40. Nobody should pretend otherwise. That is why borrowers shop.

Now price the other side. If a transaction falls apart at day forty-five because the rate quoted was not achievable for this file, the borrower may lose:

• earnest money at risk, if the financing contingency has expired — on this file, \$5,000
• the appraisal fee already spent — \$650
• the house, which cannot be priced
• and any rate advantage, since they now restart in whatever market exists that week

The quarter point is worth \$60.14 a month. The failed transaction is worth the house. This is not an argument that rate does not matter. It is an argument that a rate you cannot deliver is worth less than zero, and it is the entire reason this book spends thirteen chapters on documentation before it spends one on pricing.

There is a second-order version of this that takes longer to see. A borrower closes a mortgage roughly every seven years. They talk about the experience for thirty. An agent does twelve transactions a year and remembers which lender made the hard one work. The loan officer who tells a borrower on day one that their 706 score prices differently than the advertised rate has done something that looks, in that moment, like losing the deal on price — and has in fact just started a referral relationship. That is the book's fourth theme, and Chapter 38 turns it into arithmetic.

⚠️ Where Deals Die

Quoting a rate you have not priced. New loan officers do this constantly, usually by quoting the rate sheet's headline number without applying the file's adjustments.

The Linden Street file is a good example. Read from the top of a rate sheet, the day's 30-year fixed number might look like 6.250%. Applied to this file — a 706 representative score at 95% loan-to-value on a 30-day lock — it is not 6.250%, and the difference is not small. Chapter 29 rebuilds this quote from base price so you can see exactly where every basis point comes from.

The discipline: never quote a rate without the four facts that price it — representative credit score, loan-to-value, occupancy and property type, and lock period. If you do not have all four, you do not have a quote. You have a range, and you should say so.

What "saleable" actually means

"A saleable file" is easy to say and worth unpacking, because the definition is also the syllabus of this book. A file review — whether it is your own underwriter, the buyer's due diligence, or a quality control audit two years later — asks four questions, and a file dies if any one of them fails.

One: are these people who they say they are, and did they agree to this? Identity, consent to pull credit, a signed application, signed disclosures, the right people on the right documents. It sounds procedural until a file is short one signature at nine o'clock on closing morning.

Two: do the documents support the numbers? Not "is the borrower honest" — do the documents support the numbers. The income used to qualify must be reconstructable from the paystubs, the W-2s, and the verification of employment by somebody who has never met the household and is working three years later from a scanned file. Same for assets, same for the debts. This is Part II, and it is why this book spends three consecutive chapters on credit, income, and assets before it spends one on pricing.

Three: is the collateral worth it, and is our lien first? An appraisal that supports value, a title commitment that comes back clean or gets cleaned, hazard insurance in force at closing, a flood determination. On this file that meant a prior owner's mechanic's lien discovered on day 19 and not resolved until day 30. Nobody's fault, nobody's income, and it could have ended the transaction.

Four: were the disclosures right and on time? Delivered within the required windows, with fees inside the tolerances, and the timing documented. A violation here can make a loan unsaleable or force a cure that somebody pays for, and — unlike the first three — this one is entirely within your own company's control and entirely your own fault when it goes wrong.

Now look back at the list of six tasks above. Finding the borrower feeds nothing but the existence of a file. Everything else on that list maps onto one of these four questions. That is not a coincidence and it is not a rhetorical trick; it is the reason the job decomposes the way it does.

Why you are not paid more for a higher rate

There is a structural fact here that new loan officers are rarely told plainly, and it changes how the whole job feels once you know it.

Under the loan originator compensation rules in Regulation Z, an individual loan originator's compensation on a transaction may not vary based on the terms of the loan — not the interest rate, not the points, not the product. Your compensation plan is set in advance, it is generally expressed as a percentage of the loan amount, and you cannot increase it by talking a borrower into a higher rate. You also cannot be paid by both the borrower and the lender on the same transaction. Chapter 26 develops this properly, including how compensation plans are structured and what the anti-steering requirements actually require; verify the current rule with your compliance department, because it has been refined since it was written.

Two things follow. The first is practical: because your compensation moves with loan amount and not with rate, the fastest honest way to increase what you earn is to close more files and larger ones, which is a business-development problem, not a pricing problem. The second is about posture. The most common suspicion a borrower carries into this conversation — this person makes more if I take a worse deal — is one you can answer truthfully and completely. Answer it. Very few loan officers do, and the ones who do are remembered.

The rules exist because the alternative was tried. Compensation that rose with the rate is exactly the incentive structure that produced a great deal of what went wrong before 2008, and the rewrite that followed was aimed squarely at it. This is the first instance of a pattern you will see in every compliance chapter of this book: the rule is not arbitrary, it is a scar.

What you own, and what you do not

A last frame for the job, and the one that will keep you sane in your first year.

You own three things. The expectations — what the borrower and the agent believe is going to happen, and when. The calendar — every date in the transaction and the distance between them. And the truth-telling — being the person who says the difficult thing while it is still cheap to hear. Nobody else in the transaction owns any of these. The processor owns documents, the underwriter owns the decision, the closing agent owns the signing. Expectations, the calendar, and candor are yours alone, and every catastrophe in this business is downstream of one of them being dropped.

You do not own the rate. You do not own the appraised value. You do not own the underwriter's decision, the seller's behavior, the title company's queue, or whether an insurance agent returns a call. Loan officers who confuse the two lists burn out in about eighteen months, because they take personal responsibility for outcomes they cannot influence while quietly neglecting the three they control completely.

The test is simple and you can apply it to any bad day. When a file goes sideways, ask which list the failure came from. If it came from the second list, your job was to communicate it early and accurately, and you either did or you did not. If it came from the first list, it was yours, and the honest thing is to say so and fix the system that let it happen. Most loan officers are far too hard on themselves about the second list and far too easy about the first.

1.8 What this book will teach you, and in what order

A short map, because knowing why a chapter exists makes it easier to read.

Part I — The Foundation (1–6). What a mortgage is, how the industry got this shape, how you become licensed to do the work, the arithmetic every later chapter assumes, the program map, and the process itself end to end.

Part II — The Borrower (7–13). Where loans come from, the pre-qualification conversation, the application, and then the verification core — credit, income, assets — followed by structure.

Part III — Underwriting (14–19). The conventional rulebook, automated underwriting, FHA, VA and USDA, the appraisal, and conditions.

Part IV — The Transaction (20–23). The purchase contract, title and insurance, TRID and the Closing Disclosure, and closing day and everything after it.

Part V — Compliance and the License (24–27). RESPA and TILA, fair lending, compensation, and fraud.

Part VI — The Money Behind the Loan (28–31). The secondary market, how a rate is actually built, rate locks, and the three business models in full.

Part VII — Specialized Lending (32–35). Self-employed borrowers, first-time buyers and assistance, non-QM, and construction, renovation, and reverse.

Part VIII — The Business and the Career (36–40). Technology, the purchase-versus-refinance pivot, building a book of business, running a pipeline, and the career — ending with the complete Linden Street file.

Four constructed files recur. Linden Street is the one you originate yourself. Cypress Court is an appraisal \$35,000 under contract eleven days before closing. Fulton Avenue is a self-employed contractor whose accountant's number and whose underwriter's number are not the same number. Harlow Street is a first-time buyer at a 641 score using county down-payment assistance. Each returns when its subject is in play.

1.9 Words that mean something different inside the building

Two last pieces of ground-level equipment before Chapter 2, and this is the first: a translation guide.

Mortgage lending has a vocabulary problem that is worse than most industries', because the confusing words are not technical jargon the borrower has never heard. They are ordinary English words that mean one thing in the world and something narrower — sometimes something entirely different — inside a lending company. Borrowers and real estate agents use them in the ordinary sense. Underwriters, closers, and servicers use them in the narrow one. Nobody notices the mismatch until a promise has been made.

Each of the pairs below gets its full treatment in the chapter that owns it. What follows is only the distinction, so that you stop conflating them on the phone this week.

"Approved." To a borrower, "approved" means yes, it's done, we can relax. To an underwriter it means yes, if — a conditional approval, which is what nearly every approval actually is, and which comes attached to a list of things that must still be delivered. This file's day-28 approval carried eleven of them. When you say the word "approved" to a household, say the whole sentence: "approved with eleven conditions, six of which are yours, and here they are." Chapter 19 is about that list.

Pre-qualification and pre-approval. These are used interchangeably by everyone outside the building and they are not the same object. One is arithmetic on what you were told; the other is arithmetic on what you verified. The difference is entirely in whether documents were examined, and it is precisely the difference between a letter you can defend and a letter you cannot. Chapter 8 draws the line, and it is one of the most consequential lines in the book.

"Escrow" — two unrelated meanings. In the transaction, escrow is the neutral holding of money and documents by a third party before closing; "we're in escrow" is how agents in many markets say "under contract." After closing, the escrow account is something else entirely: the impound account into which one-twelfth of the annual taxes and insurance is collected with every monthly payment, for thirty years. The earnest money sits in the first kind. The \$385.00 of monthly taxes and \$130.00 of monthly insurance in this file's payment go into the second. Chapter 23 handles both, and you should never use the bare word without knowing which one you mean.

Closing costs and cash to close. Not synonyms, and not close. On this file the closing costs — the fees for making the loan and transferring the property — are \$9,720.25. The cash to close — the number the borrowers actually have to wire — is \$25,376.34, because it adds the down payment and the prepaid items and then subtracts the earnest money and the seller's credit. A borrower who has been told "closing costs are about ten thousand" and shows up expecting to wire ten thousand has been badly served. Quote the cash to close, always, and say what is in it.

Points. Three different things share the word. A discount point is one percent of the loan amount paid to buy the interest rate down — on this file, 0.500 point, \$1,828.75. The origination charge is what the lender charges for making the loan, which on this file happens also to be quoted in points — 1.000 point, \$3,657.50. And a basis point is a hundredth of a percent, which is the unit the secondary market and your lock desk actually speak in. Three quantities, one word, and a borrower who thinks they are being charged twice for the same thing. Chapter 29 sorts it out.

Rate and APR. The note rate prices the payment; on this file it is 6.625%, and it is what determines the \$2,341.94. The annual percentage rate is a disclosure figure that expresses the cost of credit as a yearly rate including certain prepaid finance charges; on this file it is 7.253%. They are different numbers measuring different things, and the single most common pricing mistake a borrower makes is comparing one lender's rate to another lender's APR. Chapter 24 shows exactly which charges create the gap.

PMI and MIP. Private mortgage insurance is what a conventional loan above eighty percent loan-to-value carries, it is paid by the borrower to a private insurer, and it can end. This file carries it at a 0.58% annual factor — \$176.78 a month — and it terminates automatically at payment 137. FHA's mortgage insurance premium is a different animal: an upfront premium that is usually financed into the loan, plus an annual premium, and at the loan-to-value ratios most FHA buyers use, it does not terminate at all. Chapters 13 and 16 do the comparison; the vocabulary point is only that "PMI" is not a generic term for mortgage insurance and using it that way will mislead an FHA borrower about a thirty-year obligation.

Appraisal, inspection, and assessment. Three numbers, three purposes, and no two of them are the same. The appraisal is an opinion of market value produced for the lender, by an appraiser the lender's process selected, to answer one question: is the collateral worth what the loan assumes? The home inspection is bought by the buyer, for the buyer, and it is about condition, not value — an inspector will not tell you what the house is worth. The tax assessment is a county's valuation for property tax purposes and frequently bears no useful relationship to market value at all. Borrowers cite the assessment as evidence of value constantly. Chapter 18 handles the appraisal; here, just do not let the three words be used as one.

Title, deed, and deed of trust. Title is not a document — it is the legal status of ownership, an abstraction, the answer to "who owns this and subject to what." A deed is the instrument that conveys title from the seller to the buyer. A deed of trust is a security instrument, and despite the shared word it conveys nothing to the buyer at all; it is the thing that lets the lender force a sale if the note is not paid. A borrower who says "when do I get the title?" is usually asking about the deed, and Chapter 21 explains why the answer involves a title insurance policy rather than a piece of paper in the mail.

Processing and underwriting. Confused constantly, including by people in the industry. Processing assembles and verifies; underwriting decides. When a borrower asks "where is my loan?" the answer is materially different depending on which one it is sitting in, and telling them "it's in processing" when it has been in underwriting for four days is the kind of small inaccuracy that destroys trust when it surfaces.

Closing, settlement, funding, and recording. Four different moments, and in some states they are not the same day. Closing or settlement is the signing. Funding is when the lender releases the money. Recording is when the security instrument and deed hit the county land records. Whether the buyer gets keys at signing or after funding and recording is a matter of state practice and of what the contract says, and getting it wrong means a family with a moving truck in the driveway and nowhere to go. Verify how your market does it before you promise anybody a time of day.

"Locked." To a borrower, "we locked your rate" sounds like the loan is settled. It is not. A lock is a commitment about price, for a period, on a set of assumptions — it is not an approval, not a guarantee that the file will close, and not immune to the file changing underneath it. If the representative score, the loan-to-value, the occupancy, or the loan amount moves, the price moves with it. And a lock has an expiration date that arrives whether or not anybody is ready: this file's thirty-day lock was taken on day 12 and ran out on day 42. Say the expiration date out loud whenever you say the word "locked," and say it to the agent too, because the lock's clock and the contract's clock are two different clocks and somebody has to be watching both.

Conventional, conforming, and qualified mortgage. Used as synonyms and describing three separate axes. Conventional means not government-insured — not FHA, VA, or USDA. Conforming means meets the agency requirements, including the loan limit, which is revised annually and which you should verify at the Federal Housing Finance Agency rather than quote from memory. Qualified mortgage is a regulatory category under the Ability-to-Repay rule and is about the loan's features and the lender's verification, not about who insures it. A loan can be conventional and non-conforming. A non-QM loan can be conventional. Chapters 14 and 34 develop each properly.

"The bank." Finally, the word your borrower will use for the entire apparatus described in §1.3 — you, your employer, the underwriter, the aggregator, the investor, and the servicer, compressed into two syllables. It is not worth correcting most of the time. It is worth hearing, because when a borrower says "why does the bank need this," they are asking a question about a system they cannot see, and §1.3 is the answer.

QUICK REFERENCE — the fourteen that cause the trouble

  WHAT THEY SAY            WHAT IT MEANS INSIDE           OWNED BY
  ────────────────────────────────────────────────────────────────────
  "approved"               approved WITH CONDITIONS       Ch. 19
  "pre-approved"           verified, not just asked       Ch. 8
  "escrow"                 two unrelated things           Ch. 23
  "closing costs"          ≠ cash to close                Ch. 22
  "points"                 discount / origination / bp    Ch. 29
  "rate"                   ≠ APR                          Ch. 24
  "PMI"                    ≠ FHA's MIP                    Ch. 13, 16
  "appraisal"              ≠ inspection ≠ assessment      Ch. 18
  "title"                  ≠ deed ≠ deed of trust         Ch. 21, §1.2
  "processing"             ≠ underwriting                 §1.5
  "closing"                ≠ funding ≠ recording          Ch. 23
  "locked"                 a price, not an approval       Ch. 30
  "conventional"           ≠ conforming ≠ QM              Ch. 14, 34
  "the bank"               all six parties at once        §1.3

1.10 The answers you will need before Friday

And this is the second piece of equipment: nine questions you will be asked in your first weeks, each answerable entirely from this chapter, with the answer written the way it should actually sound out loud. Read them as scripts to adapt, not to recite. The point is not the wording — it is that every one of these has a real answer, and a loan officer who reaches for "that's just how it works" in front of a borrower has thrown away the only thing that distinguishes them from a website.

"Who's actually lending me the money?"

"Technically, we are — at the closing table the money comes from us. But we're not the ones who keep it. Within a few weeks your loan will almost certainly be sold to an investor, and the rules your file has to satisfy were written by whoever buys it, not by me. That's not a bad thing for you; it's the reason a thirty-year fixed rate exists at all. But it's why I can't waive a requirement even when I agree with you that it's annoying. I'm not the one who set it, and neither is my underwriter."

That answer takes eleven seconds and it prevents about half the arguments you would otherwise have over the next fifty days. Say it in the first conversation, not the twentieth.

"Why did my loan get sold? Did I do something wrong?"

"Nothing at all. It has nothing to do with you or how you've paid. Almost every mortgage in this country gets sold, usually within a few weeks of closing, and it's how the money keeps circulating — if lenders had to hold every loan for thirty years, there wouldn't be enough money left to lend to the next family. Your rate, your payment, your term, and the address of the house don't change. What changes is where you send the check. And you agreed to it, in the sense that one of the paragraphs you signed says exactly this — sale of the note and change of servicer."

Then the safety instruction, every time, in one sentence: "If you ever get a notice like that, call the number on your most recent statement — not the number in the letter or the email — and confirm it before you send money anywhere new."

"Why do I have to escrow? I've been paying my own taxes for twenty years."

"Because of what happens if you don't, and it's not about trusting you. In most states, an unpaid property tax bill creates a lien that jumps ahead of ours, even though ours was recorded first. So a homeowner who falls behind on taxes can push the lender out of first position without ever missing a mortgage payment. We collect one-twelfth a month so that can't happen. The side effect is that you never get a surprise bill for a year's taxes — on your file that's \$4,620 a year, and \$385 a month is a lot easier than \$4,620 in one November."

Some borrowers can waive escrow on some programs at some loan-to-value ratios. Do not promise it until you know your lender's rule and the state's, and Chapter 23 explains what a waiver actually costs.

"Can I just take over the sellers' loan? Their rate is way better than anything you can offer."

"On a conventional loan, almost never. There's a clause in their security instrument that lets their lender call the whole balance due if the property transfers — a due-on-sale clause. It's enforceable. There are narrow exceptions for certain family transfers, but a sale to a buyer isn't one of them. Government loans work differently and can sometimes be assumed by a qualified buyer, so it's a real question worth asking when the seller has one, and I'll find out. What I won't do is let you write an offer that depends on it before we've confirmed it in writing."

That last sentence is the one that matters. This question usually arrives from an agent, in a hurry, looking for a way to make an unaffordable house affordable, and the honest answer is often no.

"My brother-in-law got a much lower rate. Why can't I?"

"He might have. Two things are probably going on. The first is timing — if that was a year or two ago, that rate belonged to that week, and neither of us can get it back. The second is that a rate doesn't belong to a person or a company; it belongs to a file. Four things price it: the representative credit score, how much you're putting down, what kind of property it is and whether you'll live in it, and how long we need to hold the rate. Change any one of those and the number changes. If you want, tell me his four and I'll tell you honestly whether we could have done the same thing — and if we could, I'll tell you that too."

The offer at the end is the whole answer. You are not defending your number; you are inviting a comparison on terms that make sense, which is something the borrower has never been offered and which almost nobody follows up on.

"Is the pre-approval a guarantee?" — the listing agent, who is deciding whether to trust it

"No, and anybody who tells you it is, is selling you something. Here's what it is: I've pulled credit, I've looked at income documents and asset statements, and the number in that letter is supported by what I've seen. What's still open is the appraisal, the title work, and a final underwriting decision — and I've listed those on the letter so you know exactly where the risk is. If anything changes I'll call you the same day, before you hear it from anyone else. Here's my cell."

Notice that the answer is more candid than the question invited, and that this is what makes it persuasive. Compare it to "yes, they're fully approved," which is both untrue and, to an experienced listing agent, an immediate signal that the letter is worthless.

"Why can't you just approve it? You work there."

"I do, and I still don't get a vote. The person who decides is an underwriter I'm not allowed to influence, and they're not deciding whether you seem trustworthy — they're certifying that this file matches what an investor was promised, under warranties that can come back on this company years from now. Which is actually good news for you, because it means the decision runs on documents, not on whether somebody likes you. My job is to make sure the documents say what your life actually says. That's what I'm asking you for."

This is the single most useful thing in the chapter to be able to say naturally. It reframes the underwriter from an obstacle into a standard, and it reframes your document requests from bureaucracy into advocacy — which is what they are.

"Who is this 'trustee'? I didn't hire anyone."

"You didn't, and in a deed-of-trust state there's a third name on your security instrument that isn't you and isn't us. The trustee holds a power of sale until the loan is paid off. Nothing about that gives them any right to the house while you're paying — they're a neutral placeholder. It's there because your state handles the enforcement side through a trustee instead of through a lawsuit. In a mortgage state there'd be two names instead of three and the process would run through a court. Neither one is better or worse for you at closing; it only matters if something goes very wrong, and it's worth knowing your state's version."

"When do we actually get the keys?"

"Signing day and key day are not always the same day, and I'd rather over-explain this than have you sitting in a truck. Here's the sequence: you sign, we fund, and the county records the documents. In some markets that all happens within a few hours and you get keys that afternoon. In others, funding and recording land the next business day, and the contract says when possession transfers — which may not be the same moment as recording either. I'll confirm the exact sequence for your closing and tell you the day before, in writing, so you can schedule the movers around it. What I won't do is guess at a time and have you pay a moving crew to sit in a driveway."

There is a version of this question the agent asks too, usually as "can we do a morning closing so they can move in the same day?" Learn your market's actual sequence and answer it precisely. It is a small thing that experienced agents use to sort loan officers into two piles.

The general shape of all nine answers is the same, and it is the shape of the whole book: name the party whose rules are actually operating, explain what they need and why, and end with a specific thing you will do. A borrower who is told "that's just the requirement" learns nothing and trusts you slightly less. A borrower who is told whose requirement, and why it exists, has been handed a map of a system that has been frightening them for weeks — and they will remember which lender did that for thirty years.

🗂️ The Loan File

Chapter 1 contribution: open the file and name the parties.

The agent's call is at 8:40 Wednesday. Call that day 0. You have a property address, a price, two buyers you have not met, and a deadline of two o'clock.

Before you learn anything about them, fill in what you already know about the structure of the transaction you are about to enter:

What this settles: the map. You know who decides (the underwriter), who owns the clock (the contract), and who owns the relationship (you).

What it does not settle: anything about whether these people can buy this house. You have no income, no assets, no credit, and no idea whether \$385,000 is a reasonable number for them. Every one of those is a chapter.

Open questions carried forward:

• Q1. Can they afford this house, or only qualify for it? (Chapter 8)
• Q2. Which program fits — conventional or FHA? (Chapter 13)
• Q3. Will an appraisal support \$385,000? (Chapter 18)

Your task. In Appendix C's workbook, start the file. Record the day-0 facts, and write one sentence answering: if this loan closes, who will own the debt in a year, and who will the borrowers actually pay? You will not be able to answer it with certainty — that is the point. Write what you can support and mark the rest unknown. That habit is the whole book.

Conclusion

A mortgage is two documents: a promise to repay and a recorded claim against a specific piece of land. The promise can be sold and will be; the claim stays where it was recorded. Which form the claim takes — mortgage or deed of trust — is a matter of state law and determines how a foreclosure proceeds.

The money is not your employer's, in any lasting sense. It comes from investors who buy securities backed by pools of loans, through aggregators who publish, in advance, the exact terms on which they will buy. Your employer funds the loan at closing, usually with borrowed money, and sells it within weeks. That single fact explains why guidelines are treated as absolute, why documentation matters more than sincerity, and why a loan that cannot be sold is a disaster rather than an inconvenience.

Five sets of hands touch the file. The one that decides never meets the borrower. The one the borrower trusts cannot decide. Standing in that gap — anticipating the underwriter's questions, translating their requirements into human tasks, and holding the calendar — is what a loan officer is actually paid for.

Which is why the answer to "what's your rate?" is never just a number. You do not set rates. You build files that close at the rate you quoted, which is a different and much rarer skill.

Next: the thirty-year fixed-rate mortgage that seems so obviously the default product did not exist before the 1930s, and neither did the appraisal, the loan-to-value limit, or the secondary market. Chapter 2 explains where all of it came from — and why nearly every rule you will spend this book learning is a scar from a specific failure.

Key Terms

Mortgage — in strict usage, the security instrument that pledges real property as collateral for a debt; in ordinary usage, the note and security instrument together. (Ch.1)

Note (promissory note) — the borrower's written promise to repay: amount, rate, payment, term, and default terms. Evidence of the debt. Not recorded. (Ch.1)

Security instrument — the recorded document creating a lien on the property to secure the note; takes the form of a mortgage or a deed of trust depending on state law. (Ch.1)

Lien — a legal claim against a specific property that can be enforced by forcing its sale. (Ch.1)

Mortgagor / mortgagee — the borrower who grants the mortgage and the lender who receives it, respectively. (Ch.1)

Deed of trust — a three-party security instrument (trustor, trustee, beneficiary) used in many states, generally permitting non-judicial foreclosure. (Ch.1)

Loan officer / mortgage loan originator (MLO) — the person who takes a residential mortgage loan application or offers or negotiates terms for compensation or gain. (Ch.1)

Origination — the process of creating a mortgage loan, from application through funding. (Ch.1)

Processor — the person who assembles and verifies the file and submits it to underwriting; has no approval authority. (Ch.1)

Underwriter — the person who evaluates the file against guidelines and issues the decision. (Ch.1)

Closing agent — the neutral party (title company, escrow company, or attorney) who conducts the signing, disburses funds, and records documents. (Ch.1)

Servicer — the entity that collects payments, administers escrow, and handles delinquency after closing; often not the original lender. (Ch.1)

Investor — the ultimate supplier of mortgage capital, who buys securities backed by pools of loans. (Ch.1)

Primary market — where mortgage loans are originated: borrower and lender. (Ch.1)

Secondary market — where originated loans are sold, pooled, and securitized. (Ch.1)

Retail lender — a lender whose own employees originate, underwrite, fund, and close loans in its own name. (Ch.1)

Mortgage broker — an intermediary who takes applications and places them with wholesale lenders who underwrite and fund in their own name. (Ch.1)

Correspondent lender — a lender that underwrites and closes in its own name using a warehouse line, then sells closed loans to investors. (Ch.1)

Warehouse line of credit — short-term borrowing a lender uses to fund loans at closing, repaid when the loan is sold. (Ch.1)

Endorsement — the transfer of the note itself, by signature and delivery, often on an attached allonge; nothing is recorded. (Ch.1)

Assignment — the recorded transfer of the security instrument from one holder to another, which is how the public record catches up with a private sale of the debt. (Ch.1)

Due-on-sale clause — the covenant in the security instrument permitting the lender to accelerate the loan if the property is sold or transferred. (Ch.1)

Spaced Review

1. Your borrower calls in February, upset: they received a letter saying to send their payment to a company they have never dealt with. Using only §1.2 and §1.3, explain in two sentences what has almost certainly happened and why it is not a problem.

2. A colleague says, "Fannie Mae denied my borrower." Rewrite the sentence so it is accurate, and name the party that actually issued a decision.

3. A file is originated by a mortgage broker. At the closing table, whose money funds the loan, and whose name appears on the note?

4. Explain to a real estate agent, in under thirty seconds, why the underwriter cannot simply call the borrower and ask about the \$4,900 deposit.

5. The chapter argues that a quarter-point rate difference is worth \$60.14 a month on this file but that the rate is still not the job. Restate that argument in your own words without using the word "relationship."