The Real Unit of the Business: The Deal

If a bank is an allocator of scarce capital, strategy is not decided in the annual committee alone. It is executed—or betrayed—every time someone approves or declines a transaction. The Real Unit of the Business: The Deal is not the strategic plan. It is the individual credit facility.

The previous article established that premise: a bank is not an intermediary of flows, but an allocator of scarce capital under risk constraints.

If that holds, strategy is not defined solely in the annual planning sessions, nor in the quarterly reviews. It is executed, distorted, or betrayed in the smallest decision unit of the business.

The problem is that many organizations live with a deep disconnect: at the top, capital and strategy are debated, but when the decision reaches the committee, the question that organizes the discussion tends to be far simpler:

Is the deal approvable?

That difference seems minor.

It is not.

The Illusion of the Approvable Deal

In traditional banking, the success of a transaction is usually measured through a binary filter: approve or decline. If it meets policy, collateral is sufficient, the client has a track record, and the documentation is complete, the deal moves forward.

That approach mistakes compliance for decision quality.

Policy defines what can be approved. Strategy defines what should be approved.

Confusing those two questions is one of the costliest mistakes in banking.

A transaction can be formally viable and strategically weak. It can be added to the balance sheet without triggering any immediate alert and still contribute to a portfolio that is less flexible, less selective, and more dependent on favorable conditions.

A bank does not deteriorate solely through obviously bad deals. Far more often, it deteriorates through reasonable deals that no one questioned deeply enough.

That is the trap of the approvable deal: treating the absence of objections as evidence of quality.

What is approvable is not always what is right. That difference eventually shows up on the balance sheet.

A bank is not what its strategy declares. It is what its credit committees approve.

The Transaction as an Economic Cell

If the bank is an allocator of capital, each transaction is its fundamental economic unit. It is neither paperwork nor a sale; it is the point where client, risk, tenor, collateral, structure, commercial relationship, and use of the balance sheet all converge.

That is why it must be evaluated not only by its individual attributes, but by the kind of institution it helps build.

In a financial organization, decisions accumulate as institutional memory. No single transaction seems capable of defining an institution's risk culture.

Precisely for that reason, none does so alone: they do it all together, quietly, until a succession of exceptions shifts the approval standard without any committee having decided so explicitly.

The exception becomes precedent. The precedent becomes practice. The practice becomes culture. That accumulation rarely appears on the dashboards as an aggregate risk, even though it ends up being one of the most important determinants of the balance sheet's future quality.

Exceptions approved as "just this once" end up becoming, two years later, the unwritten policy of an entire business line. It never happened through one big, wrong decision. It happened because no one reviewed the pattern the exceptions were building.

The Trap of the Overall Relationship

One of the most common arguments for approving weak deals is the so-called "overall relationship."

The logic is familiar: this transaction, taken on its own, is not the best; but the client is important, the relationship is broad, there is shared history, and the future upside offsets today's concession.

It is not always a wrong argument. A universal bank must manage relationships, not just isolated transactions. The problem arises when the overall relationship stops being an exceptional criterion and becomes a permanent justification.

When the potential value of the relationship substitutes for the merit of the transaction, discipline erodes. Every exception approved in its name commits capital without anyone measuring its marginal return.

The future payoff is rarely reviewed with the same rigor applied when the initial concession was approved: the promise stays in the narrative; the decision stays on the balance sheet, and its effects tend to surface years later, once portfolio quality already reflects decisions no one questioned at the time.

The overall relationship should not be banned; it should be governed. If a transaction is approved for reasons that go beyond its individual merit, that decision must be made explicit, traceable, attributable to an owner, and subject to follow-up. Otherwise, it stops being a competitive advantage and turns into a gray area where strategy dissolves.

The Architecture Decision

If every transaction is where strategy gets executed, the implication is direct: push economic discipline down to the first line of decision, even if that reduces convenience, speed, and volume. Governing each transaction as an act of institutional architecture requires three changes.

1. Raise the approval standard. It is not enough for the deal to be possible; it must be consistent with the bank you want to build.
2. Make individual accountability visible. Diffuse responsibility is the natural refuge of accumulated bad decisions.
3. Review exception cohorts annually by business line and originating unit—not to punish the inevitable error, but to identify which exceptions repeat, where the standard is being relaxed, and whether, over time, those decisions strengthened or weakened the return on the capital committed.

The resistance will be predictable: the organization will protect speed, the commercial relationship, and volume, because that is what it has historically rewarded.

But the test is not whether each transaction is defensible on its own. It is whether, taken together, those exceptions strengthened the return on the capital committed or merely lowered the approval standard.

The first sign of deterioration is not a rise in delinquency. It is that exceptions begin to require less and less discussion and are justified with the same arguments. At that point, the organization is no longer exercising judgment: it is normalizing exceptions.

Banks rarely deteriorate through a single big, wrong decision. They deteriorate transaction by transaction, when each one is judged solely by its approvability.

Because a bank is built—or eroded—one transaction at a time.