Case Study 40.1 — The Boom, the Hiring, and the Contraction: Origination Employment Through the 2020–2022 Rate Cycle

Type: real, public market and industry history Sources: Tier 1 for the rate cycle, the monetary policy actions, the named corporate events, and the existence of the data series named below. Tier 2 for every magnitude.

A deliberate omission, stated up front. This case study contains no employment statistics, no attrition rates, and no volume figures, and that is not an oversight. Every number you would want here exists in a public series — the Freddie Mac Primary Mortgage Market Survey, the Mortgage Bankers Association's origination estimates and its Quarterly Mortgage Bankers Performance Report, the Bureau of Labor Statistics' industry employment series, and the NMLS annual reports on licensed originators. Those series are revised, they are seasonally adjusted differently by different publishers, and quoting one from memory is exactly the habit §7.1 of this book exists to prevent. Read them at the source. What this case study supplies is the structure, which does not get revised, and which is what you actually need.

Why this case study exists

You are going to work through at least one of these. Probably several.

The 2020–2022 cycle is the most recent, the best documented, and the most instructive, because it ran its full arc — expansion, euphoria, contraction, and consolidation — inside roughly thirty-six months, fast enough that the same people were present for all four phases. Loan officers who started in 2020 experienced a market in which almost nothing they did could fail, and then a market in which almost nothing they did could succeed, without changing anything about how they worked.

§40.8 makes the argument in a paragraph: rates fall, volume arrives without being asked for, a great many people conclude they are good at this, and then rates rise. This case study is that paragraph with its mechanism exposed — what specifically happened, in what order, to whom, and why the industry's cost structure turned a revenue decline into an employment collapse.

Background: what the expansion actually was

The sequence is public record and is worth stating as a sequence.

Rates fell to the lowest levels in the history of the series. The Freddie Mac Primary Mortgage Market Survey, which has tracked average offered rates since 1971, recorded successive record lows for the 30-year fixed rate through 2020 and into the first weeks of 2021. The context was the Federal Reserve's response to the COVID-19 economic disruption, which included a near-zero policy rate and large-scale purchases of Treasury securities and agency mortgage-backed securities. Chapter 28 explains the transmission: the Federal Reserve's purchases of agency MBS are a direct bid into the market that ultimately prices mortgage rates, and the effect on primary rates was immediate and sustained.

Refinance volume arrived in a wave. Chapter 37's arithmetic explains why the response is so violent. A refinance is triggered by a comparison, and when the comparison changes for tens of millions of households simultaneously, tens of millions of households become candidates in the same month. No outreach is required. The phone rings.

Capacity, not demand, became the binding constraint. This is the part loan officers who lived through it remember most vividly and the part newer originators find hardest to believe. Lenders could not process what was arriving. Turn times lengthened. Appraisers and underwriters were booked out. Lenders responded the way any manufacturer responds to demand it cannot serve: they widened margins and they hired.

The hiring was aggressive and it was industry-wide. Underwriters, processors, closers, funders, post-closing staff, and licensed originators were recruited at scale, frequently with signing bonuses, frequently from competitors, and frequently with compensation set against the volume then arriving. Several of the largest lenders expanded headcount substantially. Chapter 31's channel map matters here: retail lenders, wholesale lenders, and correspondents were all expanding at once, which meant experienced staff could move for a raise more or less at will.

Read the mechanism, not the mood. Nothing in the paragraph above required anyone to behave foolishly. A lender facing demand it cannot serve should add capacity. The error is not hiring during a boom; it is hiring during a boom on the assumption that the boom is the baseline.

The turn

Monetary policy reversed, and it reversed hard. Beginning in March 2022, the Federal Reserve raised its policy rate repeatedly and at an unusually rapid pace, and it began reducing its balance sheet. Mortgage rates, which had already begun rising in late 2021, rose sharply through 2022. By the autumn of 2022 the Primary Mortgage Market Survey's 30-year fixed average had crossed seven percent — a level it had not reached in roughly two decades.

Refinance volume did not decline. It substantially disappeared. This is the single most important structural fact in the case study, and it follows directly from Chapter 37. A refinance requires the new rate to beat the existing one. When a large majority of outstanding mortgages carry rates from the 2020–2021 window, a market at seven percent does not offer a worse refinance — it offers none, for most of the book. The demand did not soften. The population of candidates was eliminated.

Purchase volume fell too, but for different reasons and by less. Purchase demand is damped by household necessity — people move for jobs, marriages, births, deaths, and divorces regardless of the rate — and it is also constrained by affordability and by the supply of listings, which itself tightened as existing owners declined to trade a low rate for a high one. The asymmetry between the two channels is the entire reason §40.8's first protection is a purchase referral base.

What happened to employment

Here the record is public, specific, and worth knowing by name rather than by number.

• In December 2021 — before the worst of the rate move — the digital lender Better.com laid off a large group of employees in a single video call. The event was reported worldwide and became a shorthand for the industry's turn, partly because of how it was conducted and partly because of its timing: it happened while many originators still believed the market was merely cooling.
• In June 2022, First Guaranty Mortgage Corporation filed for Chapter 11 bankruptcy protection. A lender ceasing operations mid-cycle is a different event from a layoff: loans in process must be moved or die, and the originators holding them discover that their pipeline is not their asset.
• In July 2022, Sprout Mortgage — a non-QM lender — shut down abruptly. Chapter 34's account of a thin investor base is the mechanism: when the buyers for a category of loan withdraw, the lender originating them has no product, and no product means no company.
• In January 2023, Wells Fargo announced a significant retrenchment in home lending, including exiting the correspondent channel. When a bank of that size withdraws from a channel, the effect is not confined to its own staff — it removes an outlet for other lenders' loans and reshapes the channel map in Chapter 31.
• Layoffs, hiring freezes, voluntary buyouts, and branch consolidations were announced across the industry through 2022 and into 2023, at bank lenders and independent mortgage banks alike.

Alongside the named events, two public series recorded the shape of the contraction and are the places to look for magnitude. The Mortgage Bankers Association's Quarterly Mortgage Bankers Performance Report tracks per-loan production revenue and expense at independent mortgage banks; during the contraction it recorded the industry moving from historically strong per-loan profitability to per-loan production losses. The NMLS publishes annual reports on the number of licensed mortgage loan originators, and the count declined from its cycle peak.

(Direction, in both cases, is documented. Magnitude is Tier 2 and is revised. Look it up.)

What it shows: five structural facts

1. Origination revenue is per-unit, and almost nothing else

A mortgage business earns when a file closes. There is no subscription, no recurring fee, no maintenance contract — servicing aside, and most originators do not own servicing. Revenue is therefore a direct multiple of units, exactly as §40.2's identity says. When units fall by two thirds, revenue falls by two thirds, in the same quarter, with no lag and no cushion.

2. The revenue is variable, but the manufacturing cost is fixed

This is the asymmetry that turns a revenue decline into a solvency problem. Originator compensation is a commission and therefore self-adjusting: fewer files, less paid. But underwriters, processors, closers, funders, compliance staff, licensing costs, technology contracts, and leases are fixed. They cost the same in a quarter with 400 units as in a quarter with 120.

Case Study 40.2 works this arithmetic on a composite branch, and the finding there is the finding here: an origination business's break-even is set almost entirely by its manufacturing cost, and management's only real lever is how much of that cost it agreed to carry.

3. Both adjustments lag, and both lags cost money

Capacity is added after volume arrives, because that is when the need is visible — which means a lender pays to recruit, onboard, and train precisely as volume peaks, and the new staff reach full productivity as it turns. Capacity is removed after the pipeline empties, because managers reasonably wait to confirm the decline is not seasonal, and because the loans in process still have to be closed by somebody. The industry pays on the way up and again on the way down.

4. The skills a refinance boom rewards are not the skills a purchase market requires

A refinance is originated without a real estate agent, without a purchase contract, without a listing, and frequently without a competing offer or a hard closing date. It rewards throughput, speed of response, and marketing spend. A purchase transaction requires a referral relationship, a contract with dates in it, coordination with an agent and a title company, and the calendar discipline in Chapter 39.

An originator who built a business entirely on the first set of skills owned nothing transferable when the second market arrived. That is what §40.8 means by mistaking a market condition for a skill, and it is not an insult — it is a description of what was rewarded.

5. The individual's exposure is the firm's exposure, one level down

The originator's income is variable. Their rent or mortgage, their car payment, their childcare, and their insurance are not. Every structural point above applies to a household with the numbers made smaller, which is precisely why §40.8's fourth protection is reserves and why §40.1 insists on them before the first day.

Outcome

The market consolidated. Some lenders failed, some were acquired, some exited channels, and a great many originators left the business — some to adjacent roles inside it, some to other industries. The survivors, at both the firm and the individual level, tended to share a small number of characteristics, and they are the same characteristics §40.8 lists:

• purchase business that predated the boom, because it did not have to be rebuilt from nothing at the worst possible moment;
• a database, which produced the transactions that still existed — the moves, the life events, the second homes, and eventually the cash-out and renovation files in Chapter 35;
• a niche, because a specialist's borrowers are chosen for a reason other than rate;
• a cost structure that could be operated at a fraction of peak volume;
• and reserves, at the firm level and the household level, sufficient to fund the gap between the revenue decline and the cost adjustment.

None of those five can be acquired quickly, which is the whole point. Each one is an asset built in a good market and spent in a bad one.

The lesson

A boom is a test you cannot fail and therefore cannot learn from.

That is the uncomfortable form of it. In 2020 and 2021 a loan officer could quote a rate off the top of a sheet, take an application from a borrower who called them, and close it, and the file would work — not because the work was good but because the market was carrying it. Every discipline this book teaches was economically optional for about twenty-four months. Payment shock did not bite because payments were falling. Referral relationships did not bite because the phone rang anyway. Cost structure did not bite because revenue covered anything.

Then the conditions changed and all of it bit at once.

The transferable rule is not "be pessimistic." It is narrower and more useful: when your results improve without your practice improving, write down what is carrying you. That sentence is the practical form of §40.8's closing question — what am I assuming that I have not written down? — and it is answerable in a good market, which is the only time it is worth asking.

There is a second lesson, smaller and kinder. The people who left were mostly not bad at this. Many were competent originators who started at the wrong point in a cycle, were hired against volume that was never going to persist, and ran out of runway before their referral base could mature. §40.1 exists so that the reader can see that outcome coming and price it — in months of reserves — before accepting the job. That is not cynicism about the industry. It is the same discipline the book asks you to apply to a borrower's file, turned around and applied to your own.

Discussion questions

1. This case study contains no employment statistics, deliberately. Would it be a better case study with them? Argue both sides, then state what you would have to do before quoting one in a presentation to your own branch.

2. Lenders hired aggressively into a boom, and the case study says that was not, by itself, an error. Where exactly is the line between "adding capacity to serve demand" and "scaling a cost structure on boom volume"? Write the test you would apply, and then apply it to the seven-year office lease in Exercise 40.25.

3. Refinance volume did not decline, it substantially disappeared, and the case study attributes that to the rate on the existing book rather than to the rate on offer. Explain the mechanism to somebody outside the industry in three sentences. Then state what it implies about the next contraction's timing.

4. The case study argues that a refinance boom rewards skills that do not transfer to a purchase market. Take Chapter 38's list of what builds a referral base and Chapter 39's pipeline disciplines, and identify which specific ones a pure refinance business never has to develop.

5. First Guaranty Mortgage Corporation's bankruptcy and Sprout Mortgage's shutdown affected originators who had done nothing wrong and had loans in process. Write the two-paragraph plan you would want to already have in a drawer for the morning your employer stops funding. Name the documents you would need and where they are today.

6. The five survival characteristics all have to exist before they are needed, which makes them expensive in a good market and priceless in a bad one. Rank them by how hard each is to build while busy, and state which one you would build first with a fixed and inadequate amount of time.

7. §40.1 says a hiring manager is paid on headcount and production, so the honest version of the first year is not a recruiting pitch. Given this case study, what should a manager have told a candidate in mid-2021 — and would a manager who told them that have hired anybody?