2008 was not a credit crisis; it was a design crisis

When trust dies, the brand, the rating, and the banker's reputation do not survive; what survives is what can be identified, controlled, and liquidated — the rest was refinancing dressed as wealth.

Lehman Brothers filed for bankruptcy protection yesterday. The market will call this a credit crisis. It is a comfortable description because it allows one to blame an abstraction. Credit does not wake up in a bad mood. Structures fail, balance sheets lie, collateral is overestimated, and institutions discover too late that liquidity was merely someone else's willingness to refinance them. What died was not credit. It was the belief that any obligation could be rolled over as long as asset prices kept rising.

A few months ago, mortgage securities were treated as near-cash because models said millions of borrowers would not stop paying at the same time. Yesterday, the market discovered that correlation is not a mathematical constant. It is an animal that sleeps through prosperity and wakes up hungry in a panic. When everyone needs to sell, different assets become the same thing: collateral in search of a buyer. When nobody trusts the value of the collateral, even a solvent balance sheet looks like fraud.

The phrase "crisis of confidence" will also be repeated. It is correct and useless. Confidence is not a moral variable. It is the residue left after the creditor evaluates the flow, the collateral, the priority, the liquidity, and the capacity to enforce. When those things are clear, little confidence is needed. When they are opaque, confidence has to carry the entire building. Wall Street built fifty floors on top of it and called the result diversification.

The central problem was the design. Fragile mortgages were pooled, sliced, and rebranded. Subordination seemed to absorb predictable losses. The agencies assigned grades based on historical series that did not contain the regime the products themselves were helping to create. Banks financed long assets with short money. Off-balance-sheet vehicles allowed economic risk to exist without looking like accounting risk. Derivatives spread protection among institutions that perhaps lacked the capital to honor it. Each piece looked rational in isolation. The system was insane as a whole.

The lesson is not that securitization is evil. Knives do not commit murders on their own, although knife sellers should not be responsible for certifying the safety of the kitchen either. Securitization can separate origination from funding, broaden the investor base, and reduce cost. But when the originator retains no loss, quantity replaces quality. When the evaluator is paid by whoever wants the grade, opinion becomes a product. When the investor outsources understanding, the rating becomes religion. And financial religions always end in human sacrifice.

Collateral, unlike reputation, does not need to be persuaded to keep a promise. It can be sold, provided it exists, is legally available, and holds value when everyone else is trying to sell it. That last condition is ignored in good times. A house appraised at a million is not a million in collateral if the price depends on abundant credit. The value of a guarantee is not the price observed before the crisis. It is the probable price inside it, discounted by the time, cost, and conflict required to take possession.

That is why it is not enough to say an operation is secured. One must ask: secured by what, in which jurisdiction, with what priority, registered where, enforceable in how long, and under what liquidation scenario? The adjective "secured" is usually deployed to end the analysis exactly when it should begin it. The debtor offers a property; the creditor hears safety. Between the two lie the registry, preference, taxation, occupancy, upkeep, the market, litigation, and time. Fear lives in those intervals.

In Brazil, the crisis will arrive through different channels, but the question will be the same. Companies that depend on short external capital will discover that an available line is not capital. Exporters will watch foreign banks withdraw liquidity. Companies with currency derivatives will discover that protection and speculation can use the same document. Domestic credit will not dry up because every debtor got worse on the same day. It will dry up because creditors will lose the capacity and the willingness to distinguish one risk from another.

That is when the branch charges extra for its own confusion. Without structure, the bank prices the entire class by its worst example. The good company pays for the fear produced by the bad company. The entrepreneur calls that a spread. I call it a tax on indistinguishability. Whoever cannot prove where the money comes in, which asset answers, and how the creditor exits will be treated like everyone else. In calm times, that costs a few points. In a crisis, it costs access.

The companies that survive best will not necessarily be the least indebted. They will be the ones whose debt is compatible with their assets. Working capital financed by verifiable receivables can keep existing even when the originator's balance sheet loses prestige. Equipment financed over a tenor similar to its useful life produces a logic of repayment. An infrastructure project with a long-term revenue contract can be ring-fenced. Generic debt, with no destination, rolled over every ninety days, depends exclusively on the continuation of confidence.

The crisis reveals a distinction the years of abundance erased: liquidity is not solvency, and solvency is not structure. A company can have assets greater than liabilities and break tomorrow because the liability matures first. It can generate enough cash in the long run and die because the collateral is trapped. It can own valuable wealth and be unable to turn it into liquidity without destroying it. Good design does not make the business immune. It aligns clocks, priorities, and exits so that a temporary problem does not have to become a permanent bankruptcy.

My reaction would not be to buy any credit just because prices fell. Declines do not create a margin of safety when the value is unknown. I would look for instruments where I could compute the recovery without depending on growth, refinancing, or state benevolence. Debt with liquid collateral, low leverage against liquidation value, simple documentation, and uncontested priority. I would rather be the creditor of a mediocre asset well protected than the shareholder of an excellent story financed by short-term debt.

I would also look for the infrastructures the crisis will make necessary: registration, custody, independent appraisal, collection, covenant monitoring, and markets able to trade troubled assets. After every fire, regulators demand more extinguishers; investors, however, tend to buy the company that manufactures smoke. The durable money will be in reducing the opacity that made the panic rational. The next decade will be less tolerant of risk that exists only in a spreadsheet.

There is an inevitable political temptation: restoring confidence by decree. Central banks will provide liquidity, governments will guarantee liabilities, and institutions deemed systemic will be preserved. Some measures will be necessary. But public liquidity does not fix a bad asset; it only changes who carries the time until the loss appears. A state guarantee can stop the run, not turn an insolvent borrower into a solvent one. The State can answer fear. It cannot repeal arithmetic.

The easy critique will say greed caused everything. Greed is a constant, therefore it explains little. What changed was the permission to leverage it. Incentives turned volume into bonuses, grades into merchandise, and refinancing into a business model. The architecture rewarded people for originating risk and punished others, far away, for discovering it. When responsibility is separated from decision, recklessness stops being a moral defect and becomes a rational strategy.

This matters to the Brazilian entrepreneur because he also lives inside an architecture that separates price from risk. The bank offers a rate, demands a personal guarantee, and takes every possible security, even when one would suffice. Not because it modeled each flow, but because generic documentation and bargaining power allow it. The debtor accepts because he compares banks, not structures. A global crisis should teach him to ask less "what is the rate?" and more "what event turns this debt into a problem?"

If the answer is "the line not being renewed," the company does not have financing; it has a revocable concession. If it is "the dollar rising," there may be speculation hidden in the treasury. If it is "the property falling twenty percent," the collateral margin was fiction. If it is "the largest client paying late," the concentration should have been in the price. Good debt makes the breaking point visible before it is signed. Shelf debt hides it until collection.

Collateral will survive the death of trust only when it was designed for the day trust dies. That requires discount, control, registration, and real enforcement capacity. It is not enough to hold a promise over an asset. One must hold a path to it. The crisis of 2008 is the difference between the map drawn in the living room and the road found during the storm.

The market will spend years debating culprits. I prefer to debate precedence. Who gets paid first? Who absorbs the initial loss? Which asset is beyond the reach of other creditors? Which obligation can be accelerated? Which information arrives before the cash runs out? These questions look small next to trillions of dollars evaporating. They are exactly the questions that could have prevented part of the evaporation.

2008 was not a credit crisis. It was a design crisis exposed by the withdrawal of trust. When trust dies, the brand, the rating, and the banker's reputation do not survive. What survives is what can be identified, controlled, and liquidated. The rest was just refinancing dressed as wealth.

Leo Bentier