Recovery: why terms normalise more slowly than prices

Prices are forward-looking and can retrace an entire decline in weeks. Terms are backward-looking and structurally cannot, which is why the second half of a cycle feels so much longer than the first.

Start with the estimator. The trailing window from lesson 11 works in reverse exactly as it worked forwards: five shock days stay inside a sixty-day window for sixty trading days, whatever prices do afterwards. So the measure holds at 16.48 percent for fifty-five further trading days and only then falls, one day at a time, to 15.25, 13.91, 12.43, 10.74 and finally 8.73 percent on the sixtieth day. That is twelve weeks on a calendar set by an averaging convention rather than by any market.

Then add the publication cycle. Schedules are republished on a fixed cadence, so a reading that normalises in week twelve waits for the next version, which on a quarterly cadence can be thirteen weeks away. Nothing is withheld; a number cannot change on a day when no publication occurs.

Finally there is an asymmetry in the decision itself. Being tight for too long costs a lender some volume and some goodwill. Being loose too early costs it a loss that will be examined afterwards by people who know how it ended. Those two costs are not equal, so the behaviour is not symmetric: advance rates tend to fall in a single publication and return across two or three.

The shape this produces for a pledged portfolio is specific. Illustratively, prices recover in week seven, the volatility input normalises in week twelve, and the schedule is republished in week twenty-four. For seventeen weeks the portfolio sits at its old market value with a blended advance rate of 54 percent against the 66 it had — on $2,180,000 that is $1,177,200 of lending value instead of $1,438,800, 18.2 percent less — while the balance has grown by seventeen weeks of interest, $26,100.67 on a $940,000 balance at 8.4 percent.