Essay 06 — Bruce Eickelman

What the GFC Taught Me About Structural Failure

The fault signals were visible for years. The structural decisions that would have addressed them were deferred.

I ran a mortgage origination network of more than 60 brokers through the years leading into 2008. The business model of that industry, at that time, had a structural characteristic that the crisis made impossible to ignore. The broker was paid when the loan settled. The consequence of the loan failing landed somewhere else entirely.

That decoupling was not a secret. It was the architecture. Everyone operating inside the industry understood it. The incentive was to originate. Volume was the measure. The quality of the loan, the likelihood of the borrower servicing it, the appropriateness of the product to the specific situation — these were the problem of whoever bought the paper after settlement, not the broker who wrote it.

This is not a story about fraud. Most of the people in the industry were not committing fraud. They were operating rationally inside a structure that pointed them toward a single outcome. Originate. Settle. Get paid. The rationality of the individual decision is not the issue. The structure is the issue.

The signals that were deferred

The structural fault signals were visible well before 2008. Underwriting standards were loosening incrementally — not dramatically, not overnight, but in a direction that was apparent if you were watching the criteria over time. Lenders were competing on approval rates and speed as much as on pricing. Serviceability calculations were being run on introductory rates rather than the rates that would apply after the honeymoon period. Documentation requirements were softening in the name of customer experience.

Each of these shifts was defensible in isolation. The market is competitive. Customers want fast decisions. The regulatory floor permits it. Nobody is being forced to borrow more than they can afford. The logic held for each individual decision. It did not hold for the aggregate.

Running a network of that size, you could see the aggregate. Not with perfect clarity, and not with the benefit of hindsight. But there was enough signal in the pattern to raise the question of what would happen to the portfolio if conditions changed. The question was raised, in various forms, in various rooms. The structural decisions that would have addressed it were deferred. Volume was strong. The market was performing. The decision to tighten would cost revenue immediately and would benefit the business only in a scenario that had not yet arrived.

That is always the arithmetic of the deferred structural decision. The cost of acting is concrete and present. The cost of not acting is probabilistic and future. The human brain, and any rational business operating under competitive pressure, will discount the future cost. It does this consistently and, at scale, disastrously.

The structural decoupling pattern

The deeper lesson of the GFC, as I came to understand it afterward, is not specific to financial services. The mortgage origination industry simply made visible a pattern that exists in many operating environments.

The pattern is this: when the person who makes a decision is structurally insulated from the consequence of that decision, the quality of the decision degrades over time. Not because the person is careless or dishonest. Because the feedback that would correct the decision never arrives where the decision is being made.

In the origination industry, the broker's decision was whether to write the loan. The consequence of a bad loan was a default, which arrived later, at the lender or the investor who held the paper, not at the broker. The broker's income had no relationship to the loan's performance after settlement. That is not an incidental feature of the business model. It is the mechanism that produced the outcome.

Translate this into a founder-led business and the equivalents are not difficult to find. A pricing decision made by the founder, with the consequence landing on the team's workload. A client acquisition decision made by sales, with the consequence landing on operations. A hiring decision made quickly to fill a gap, with the consequence of a poor fit absorbed by everyone who works alongside the hire. A growth commitment made to the market, with the operational cost absorbed by people who had no voice in the commitment.

The people making these decisions are not malicious. They are operating inside a structure that insulates them from the consequence of the decisions they are making. That insulation is the fault. It will produce the same degrading output that the mortgage industry produced, proportional to the stakes involved and the duration of the deferral.

The correction

The structural correction is not complicated to describe, though it is genuinely difficult to implement because it requires founders to build systems that constrain their own future choices.

Make the structure explicit. In any organisation, map where significant decisions are made and where their consequences land. If those are consistently different places, the structure has a fault that will compound over time. This mapping is not a once-done exercise. It needs to be revisited as the business changes, because the decoupling tends to increase with complexity. Growth usually adds more distance between decision-makers and consequences, not less.

Make the consequence land where the decision was made. Not as punishment, but as information. The broker who receives feedback on the performance of their loan portfolio six months after settlement is a different decision-maker than the broker who does not. The feedback does not need to be financial to change behaviour. It needs to be present, specific, and connected to the original decision.

Pre-decide the rules before the pressure arrives. The quality standards, the floor criteria, the minimum acceptable conditions — these need to be written down at a time when the business is not under the pressure that will test them. Written criteria hold under pressure. Oral commitments to quality do not. They bend incrementally, in small ways that are individually justifiable, until the aggregate bears no resemblance to the original standard.

The GFC was a compressed, high-velocity version of a structural failure that happens slowly in most businesses. The mechanism was the same. The decoupling between decision and consequence. The deferred structural decisions. The incremental softening of standards under competitive pressure. The aggregate cost arriving all at once when conditions changed.

The industry I worked in at that time had no monopoly on this pattern. It demonstrated it at a scale that made it impossible to miss. The lessons are portable. Most of the founders I work with are carrying a version of the same structural fault, operating at a scale where the collapse, when it comes, will be personal rather than systemic. That does not make it smaller.