Chapter 30 — Key Takeaways
The core claims
A lock is an option the lender wrote, and it hedged it the same afternoon. That one sentence explains why locks have prices, why longer locks cost more, why extensions carry fees, why worst-case pricing is asymmetric, and why a borrower who walks away costs the lender money even though nothing was ever lent. Chapter 28 supplies the market; this chapter supplies the consequence.
A lock is not an approval. It binds the lender's rate and price on a stated loan for a stated period. It does not obligate the borrower, does not mean underwriting has seen the file, and does not survive a change to any input that priced it — loan amount, loan-to-value, score, occupancy, property type, or program.
There is no correct lock decision, only a decision made under uncertainty with somebody else's money. Nobody knows where rates are going. A loan officer who tells a borrower "rates are going down" has made a prediction they cannot make and will own. Refuse the forecast, price both directions in dollars, and put the decision in writing.
Mortgage rates move on mortgage-backed security prices, not on the federal funds rate. The federal funds rate is an overnight bank-to-bank rate; it drives the prime rate and therefore home equity lines, credit cards, and short-term consumer credit. Long-term fixed mortgage rates reflect what investors require over years — inflation, employment, policy expectations, Treasury supply, prepayment risk. Mortgage rates can rise on the afternoon of a Fed cut, and often do.
Prices up, rates down. Say it both ways until it is automatic.
A rate quoted at 9:00 existed at 9:00. Lenders reprice intraday when the market moves enough. Quote with an expiration attached, get lock authority in writing, submit it the moment you have it, and never tell a borrower you locked them when you did not.
Size the lock against the contract, never against your optimism. Measure from the lock date to the purchase contract's closing date, add a buffer for what you do not control, and round up. Nobody has ever complained about a lock that was too long.
When a lock expires because you sized it wrong, the answer to "who pays?" is you. There is no changed circumstance to document, because nothing changed. The cost has one place to go.
The rule of thumb, and the arithmetic that proves it
Contract's closing date − today + buffer = the lock you need. Round up.
Applied to the Linden Street file on day 12:
| Contract's stated closing date | day 45 |
| Lock date | day 12 |
| Days that had to fit inside the lock | 33 |
| Lock actually taken | 30 days, expiring day 42 |
| Short by | 3 days, before any buffer at all |
| Consequence, day 42 | 15-day extension, 0.250 point = \$914.38, to day 57 |
| Paid by | the lender, as a tolerance cure — cash to close unchanged at \$25,376.34 |
| A 45-day lock taken day 12 would have expired | day 57 |
The extension bought, on day 42 and at roughly double the price, exactly the lock that was available on day 12. \$914.38 is half of the \$1,828.75 the borrowers paid in discount and a quarter of the \$3,657.50 origination charge. That is the price of a subtraction nobody did.
The four inputs to the lock decision (§30.2)
- How much time is genuinely left — count to the contract's closing date, not to your hopes.
- How much of the file is done — an unappraised, unsubmitted file has more ways to eat days.
- What the borrower can absorb — in dollars and in ratio. On Linden Street, one eighth of rate ≈ \$30/month ≈ 0.29 percentage points of back-end DTI. From 42.66%, a 45% ceiling is about one full percentage point of rate away.
- What the borrower will feel — regret is asymmetric, and that is a legitimate input to name out loud, not a lever to pull.
What each frozen number does
| Figure | What it is |
|---|---|
| 6.625% at +0.500 point = \$1,828.75 | the locked rate and price |
| \$2,341.94 | P&I at the locked rate |
| day 12 → day 42 | the 30-day lock, and the error |
| day 45 | the contract's closing date — three days past expiration |
| 0.250 point = \$914.38 | the 15-day extension, to day 57 |
| \$25,376.34 | cash to close — unchanged, because the lender cured it |
| day 51 | the actual closing, six days inside the extended lock |
| +\$30.31/mo | an eighth of a point worse at the same cost (6.750%) |
| +\$60.78/mo | a quarter of a point worse at the same cost (6.875%) |
Key terms
Rate lock · float · lock period · lock expiration · float-down · lock extension · worst-case pricing · relock · lock desk · market movement · reprice · lock policy · fallout
Do not confuse worst-case pricing (the worse of your original lock and current market, applied on a relock) with a reprice (a mid-day reissue of the sheet, which can go either way). Do not confuse a float-down (an option you purchased, with a trigger and a window) with a lender's discretionary market improvement policy (a business decision, not a right).
Where deals die in this chapter
- The lock sized against optimism. Three days short on day 12, and no event caused it.
- The 9:00 quote and the 11:00 call. Telling a borrower you locked them when you did not is how licenses end.
- The extension requested at 4:45 p.m. on the day it expires. Ask on day 38 for the day-42 problem. Emergencies get worst-case treatment.
- "Rates are going down." A forecast, delivered by a licensed professional, that you will own.
- The float-down sold in a sentence when the document says something narrower.
- The seller's amendment nobody priced. The delay had a cost and the moment to allocate it was before the amendment was signed.
What you should be able to do Monday morning
Open every live file on your desk. Find the purchase contract's closing date. Find the lock expiration. Subtract. Write the number down. If it is not comfortably positive, fix it today — while it costs a disclosed fraction of a point instead of an undisclosed one, and while it is your choice instead of your employer's problem.
Then, on the next file you quote, ask for the closing date before you say a rate — and size the lock to it, out loud, in front of the borrower.