Case Study 37.1 — The Wave and the Cliff: the 2020–2021 refinance boom and the cycle that followed

Type: real public event, described qualitatively Sources: Tier 1 (the event and its direction are well documented); Tier 2 (any specific figure — go to the published series and read it yourself)

A note on numbers before we start. This case study deliberately contains almost no statistics. The event is thoroughly documented — the Federal Reserve, the Federal Housing Finance Agency, the Federal Reserve Bank of New York's household debt reporting, and the Mortgage Bankers Association all publish relevant series, and several are free. What this case study will not do is quote a volume figure, a market share, or a headcount reduction from memory, because a textbook that prints an unsourced number teaches a habit that will eventually cost a reader a deal. Read the direction here. Get the magnitudes from the source.

Background: the conditions that made the wave

Beginning in 2020, in response to the economic disruption of the COVID-19 pandemic, monetary policy in the United States moved sharply toward accommodation. Among other actions, the Federal Reserve resumed large-scale purchases of Treasury securities and agency mortgage-backed securities. Chapter 28 explains the transmission mechanism properly; the short version is that when a very large and price-insensitive buyer enters the mortgage-backed securities market, the yield investors require falls, and mortgage rates follow.

Mortgage rates reached the lowest levels in the recorded history of the modern American mortgage market and stayed there, roughly, for the better part of two years. This book's narrator describes the full cycle as taking the market from under three percent to over seven in nineteen months. The first half of that sentence is this case study's setup; the second half is its subject.

Now apply §37.1's arithmetic. At any moment, the population of refinance candidates is the set of outstanding mortgages whose note rate exceeds today's achievable rate by enough to justify the cost of a new loan. When the achievable rate falls to a historic low, that population becomes enormous — because it includes essentially every mortgage written in the preceding two decades.

That is what a boom is. Not a marketing success. An inventory of existing loans, suddenly all in the money at once.

What happened

Three things, in sequence, and each one is a lesson.

One: the wave

Refinance origination surged. Not marginally — refinancing came to dominate the origination market for an extended period, and the industry built capacity to serve it. Lenders hired processors, underwriters, and closers; some hired aggressively enough that experienced operations staff could change employers for large increases. Loan officers who had never originated in a rising market discovered that the phone rang without any effort at all.

Two features of the wave matter for a loan officer's education.

It was self-consuming. Every household that refinanced left the population of candidates and re-entered the outstanding book of mortgages at the new, lower rate. The pool refilled only when rates fell further, which they repeatedly did — which is why the wave lasted as long as it did and felt, from inside, like a permanent condition rather than the draining of a reservoir.

It was easy in a way that hid what the job is. A refinance boom is the one market in which the loan officer's least durable skills — speed, availability, and a competitive rate — are the ones that get rewarded. Nobody needs to be talked through a contract deadline. Nobody's earnest money is at risk. There is no listing agent, no seller, and no inspection. Enormous numbers of loans closed in that period without anyone in the chain having to solve a hard problem, and an originator who entered the business in 2020 could reasonably have concluded that this is what mortgage origination is.

Two: the turn

Beginning in 2022, monetary policy reversed sharply in response to inflation. Mortgage rates rose over a span of months by an amount that, in the modern era, is without close precedent for its speed.

The population of refinance candidates did not shrink. It vanished. Every loan written during the low-rate period was now, permanently, far out of the money — not by a few basis points, but by hundreds. There was no rate at which those borrowers could benefit from refinancing, and there would not be one for years unless rates returned to something near where they had been.

This is §37.8's diagram in real life, and the shape of it is the entire point. A business whose demand curve is a threshold does not decline. It stops.

Three: the aftermath, and the part nobody predicted

Two consequences followed, and the second one surprised a great many people who should have seen it coming.

The industry contracted, publicly and sharply. Capacity built for boom volume became fixed cost against a fraction of the revenue. Headcount reductions were widespread across the industry, and several lenders exited entirely or were absorbed. The specifics are documented in public filings, industry press, and regulatory data; read them rather than a summary.

And purchase volume fell too, for a reason that had nothing to do with purchase demand. All those households now held mortgages at rates far below the market. Selling the house meant giving up the rate — replacing a mortgage in the threes with one in the sixes or sevens on a house that had also appreciated. Many households that would ordinarily have moved simply did not. The supply of existing homes for sale contracted, and that contraction, rather than a collapse in buyer demand, became a principal constraint on purchase origination.

This is the lock-in effect, and it is worth sitting with, because it violates the simple story a loan officer tells themselves during a boom: when refinances go away, I will do purchases. The rate cycle did not move business from one column to the other. It reduced both columns, on different timetables, through different mechanisms.

What it shows

First, that refinance demand is a stock rather than a flow. Ordinary businesses draw from a flow of new customers. The refinance business draws down a stock of existing loans, and a stock can be exhausted. Everything strange about refinance volume behavior — the suddenness, the burnout, the years-long flat periods — follows from that one structural fact.

Second, that capacity is the industry's real risk. The mortgage business does not fail because loans go bad, most of the time. It fails because volume falls faster than fixed cost can be removed. Chapter 31 covers the business models; every one of them has this exposure, and the correspondent and non-bank models have it most acutely because they lack a deposit base to carry them.

Third, that a loan officer's channel concentration is a personal version of the same risk. A shop with excess capacity lays people off. An originator with a single lead source is the excess capacity.

Fourth, that the two markets are not substitutes. They are correlated in a way that is occasionally negative on the timescale that matters. The best hedge against the refinance business disappearing is not the purchase business appearing — it is the purchase business already being there, built through a period when it was the worse use of the next hour.

Outcome

Origination volume, industry employment, and the composition of the origination market all moved substantially through this cycle, and the composition change — from a refinance-dominated market to a purchase-dominated one — was the most dramatic in decades. The specific magnitudes are published and worth looking up, and they are worth looking up now, while you are reading this, rather than carrying a remembered impression.

The individual outcomes are harder to document and are the ones this chapter cares about. A large number of people entered residential mortgage origination during the boom, and a large number left during what followed. That is a real pattern, widely reported, and it is not a statistic this book will invent a number for. What can be said with confidence is the mechanism: the originators most exposed were the ones whose entire funnel had been supplied by their employer, and the ones least exposed were the ones with referral relationships that predated the boom.

The lesson

The lesson is not "diversify," which is advice nobody has ever acted on because it costs something today and pays something later.

The lesson is sharper and it is about timing: the only period in which a purchase business can be built is the period in which building it is obviously the worse use of your time. During the boom, an hour spent developing a referral relationship produced measurably less income than an hour spent on the refinance in front of you. That was true. It remained true right up until it became catastrophically false, and by then the option had expired — because a referral relationship takes quarters to establish, and the month you need one is the month every other originator in your market is calling the same agents with the same story.

The originators who came through this cycle intact did not make a better prediction than anyone else. Most of them were not predicting anything. They were simply running two channels because they had always run two channels, and when one stopped they weighted toward the other. That is what a pivot actually is: a reweighting of a mix you already have, executed in a quarter. It is not a rescue mission, and it cannot be performed by someone who does not already have both halves.

Discussion questions

1. The case study argues that refinance demand is a stock rather than a flow. Name two other consumer credit products with the same property, and one that does not have it. What does the difference imply about how a lender should staff each?

2. During the boom, the rewarded skills were speed, availability, and price. Name three skills a purchase-focused originator uses constantly that a refinance-only originator may never develop. For each, name the specific way its absence shows up on a first purchase file.

3. The lock-in effect meant that a rate rise constrained purchase origination through supply rather than through demand. If you were advising a real estate agent partner during that period, what would you have told them to do differently? Be specific about who they should have been talking to.

4. A loan officer says: "I'll build referral relationships when things slow down." Using §37.9 and this case study, write the two-sentence reply. Then write the version of the reply you would actually say to a colleague you like, which is a different problem.

5. This case study deliberately quotes no volume figures. Go find three — total origination volume, refinance share, and industry employment — for the years 2020 through 2023, and cite where you got each. Then write one sentence about what surprised you. (This exercise is the point of the case study. The habit of sourcing a number before repeating it is the difference between a loan officer who can be trusted and one who cannot.)

6. Suppose you are hired tomorrow into a shop that is 90% refinance. You believe the boom has two years left. Write your first-year plan, allocating your week in hours. Defend the allocation to a manager whose compensation depends on this quarter's volume.