Case Study 2 — The Option ARM: A Product That Worked Exactly as Designed

A real, public case, approached from the failure side. Option ARMs — also marketed as pick-a-payment or payment-option loans — were widely originated in the United States through the mid-2000s. The mechanics described here are the product's documented structure; the illustrative figures are constructed.

Background

Chapter 2 argued that products are not good or bad, they are appropriate or not. The option ARM is the strongest test of that claim, because it is a product that was genuinely well-designed for a narrow set of borrowers and became one of the most destructive instruments in American consumer finance when sold outside that set.

The design intent was cash-flow flexibility. Some borrowers have lumpy income — a commissioned salesperson, a business owner with seasonal receipts, someone whose compensation arrives largely as an annual bonus. For such a borrower, a loan permitting a small payment in a thin month and a large one in a fat month is a legitimately useful instrument, and a sophisticated borrower who understands the mechanics can use it well.

The operating issue

Each month, the borrower chose among four payment options:

Option 1 is where the product lives or dies. If the minimum payment is less than the interest accrued, the shortfall is added to the loan balance. This is negative amortization: the borrower makes a payment, on time, in full, and owes more afterward than before.

The product contained two brakes:

• A negative amortization cap, commonly 110% or 115% of the original balance. Reach it and the loan recasts — it is re-amortized over the remaining term at the current balance and rate.
• A scheduled recast, typically at five years regardless of balance.

Both brakes work. Both were disclosed. And both produce, at the moment they engage, an increase in the required payment that is very large.

The marketing frequently emphasized a low "start rate" — sometimes near 1% — used to compute the minimum payment for the first year. The actual accrual rate was the index plus margin, which was vastly higher. A borrower who understood the minimum payment as "the rate" was not reading the note carelessly so much as reading the advertisement accurately.

What happened

Illustrative example (constructed; figures chosen to show the mechanism):

A borrower takes a \$400,000 option ARM. The minimum payment for year one is computed at a 1.5% start rate; the loan actually accrues at 7.0%.

Twelve months of that adds \$11,808.31 to the balance — and because the balance is growing, the accrual grows with it, so the shortfall widens every month. The borrower has made twelve on-time payments and owes \$411,808.31.

At a 115% cap on \$400,000 — \$460,000 — the loan recasts. Suppose that happens in year four with twenty-six years remaining, at a then-current rate of 7.5%:

Nothing has gone wrong. The loan is performing exactly as written. Every disclosure was delivered. And the household's payment has more than doubled.

Two conditions turned that into a systemic failure.

The product was sold on the minimum payment. Borrowers were frequently qualified on the minimum payment rather than the recast payment, meaning the file demonstrated ability to repay an amount the borrower would only pay temporarily. Many were told, accurately in the moment and catastrophically in aggregate, that they would refinance before the recast.

It was sold to the wrong borrowers. The design served a borrower with volatile but substantial income and the sophistication to manage a growing balance. A large volume was written for borrowers who used the minimum payment because it was the only payment they could make — precisely the population for whom a growing balance is unrecoverable.

When house prices fell, the promised refinance was unavailable. Negative amortization had increased the balance while the collateral value dropped; borrowers were deeply underwater at the exact moment their payment recast.

What it shows

1. Disclosure is necessary and is not sufficient. The negative amortization, the cap, and the recast were disclosed. Borrowers signed. The information was in the file and did not function as information, because a borrower comparing a \$1,380 payment to a \$2,661 payment on a conventional loan is making a comparison the disclosure did not correct. This is why the modern framework does not stop at disclosure — the Ability-to-Repay rule imposes a substantive obligation, and Qualified Mortgage status flatly excludes negative amortization rather than requiring better explanation of it.

2. Qualifying at the payment the borrower will actually make is the whole ballgame. ATR's requirement to qualify using the fully indexed rate and a fully amortizing payment is a direct response to this product. When Chapter 5 covers ARM qualification and Chapter 24 covers ATR, this is the case behind them.

3. "They'll refinance before then" is a plan that depends on the market. It was true for years. It stopped being true for everyone simultaneously — the same structure as the 1920s balloon in §2.1 and the same category of failure. Any file whose viability depends on a future transaction that requires favorable market conditions is carrying an undisclosed assumption.

4. Product suitability is a real professional obligation even where it is not a legal one. Residential mortgage originators do not owe a fiduciary duty in the way an investment adviser might, and the LO Compensation and anti-steering rules (Chapter 26) impose specific requirements rather than a general suitability standard. But Chapter 35's reverse mortgages, Chapter 34's non-QM programs, and Chapter 5's ARMs all present the same question this product did: is this instrument right for this household, or merely available to them? The originator is the only person in the transaction positioned to ask it.

The outcome for the practitioner

Option ARMs in this form are effectively gone from mainstream lending; negative amortization is excluded from QM and the ATR rule makes minimum-payment qualification impossible.

The pattern, however, is entirely alive. Any product with a payment that changes — a temporary buydown (Chapter 13), an adjustable-rate mortgage (Chapter 5), an interest-only period (Chapter 34), a forgivable second with a recapture provision (Chapter 33) — presents a version of the same question: what does this household pay when the favorable period ends, and have they seen that number in writing?

The professional habit that follows is simple and takes ninety seconds: show the borrower the worst-case payment before you show them the best-case one. If the file only works on the introductory payment, you have learned something important, and you have learned it while it is still free.

Discussion questions

1. This case argues the option ARM "worked exactly as designed." Defend that claim, then explain why it is not a defense of the product's use.

2. Contrast the two brakes — the negative amortization cap and the scheduled recast. Both are consumer protections. Explain why both produce harm at the moment they engage, and what a better design might have done instead.

3. The chapter's §2.9 says products are "appropriate or not," not "good or bad." Identify the borrower profile for whom an option ARM would have been genuinely appropriate, and list three things you would have needed to verify before recommending it.

4. Disclosure was complete and did not work. Name one disclosure in the current framework (Chapters 22 and 24 preview them) that you suspect is similarly complete and similarly ineffective, and say what you would do beyond delivering it.

5. Apply the "show the worst-case payment first" habit to a temporary 2-1 buydown. Write the two sentences you would say to a borrower, and state which number you would put in writing.