When discretion actually gets used

A clause granting absolute discretion over collateral values could, read literally, be exercised on any Tuesday for any reason. It is not, and the reasons it is not are institutional rather than generous.

Exercising discretion is expensive. It generates shortfalls across a whole population of borrowers at once, which generates calls, which generates work. It can cause the liquidation that produces the loss the lender was trying to avoid. It creates a record that has to be governed. And it damages a relationship the institution is usually trying to keep. So the powers sit unused for long stretches.

When they do fire, they fire in clusters. A name gaps on an announcement or is halted. A sector reprices sharply enough that a whole row looks mispriced. Aggregate exposure to one issuer crosses an internal ceiling. Funding costs move enough to justify repricing a class. Each of these touches many accounts simultaneously, which is why a revision feels like a policy event rather than a decision about you: it is one.

The clustering has an uncomfortable property. The conditions that make a revision worth the institutional cost are the conditions in which collateral is already falling. So the discretionary reduction and the market reduction arrive together, which is the same structural point Track 1 lesson 4 made about haircuts, observed from the other side. It also means the observed frequency of revisions during a calm stretch says very little about the frequency during a stressed one, because the two are drawn from different distributions.