The File You Keep

Look at what the ledger measures, and what it cannot. The approvals will eventually grade themselves: Highview performs or it doesn't, Meridian's fitout amortises or it doesn't, and in three years the numbers will have opinions about your year that neither you nor the GM can argue with. But run your finger down the other entries. The Kowhai increment you didn't fund — if this Christmas disappoints, no report anywhere will credit the loss that didn't occur. The Basalt Ridge courtship you re-sequenced — the developer went elsewhere, and if his project fails at another bank, your escape will look, forever, like a deal you lost. Avoided losses have an accounting problem: they book nowhere, accrue to no one, and cannot be distinguished — from the outside — from timidity. This is the hidden curriculum's hardest lesson, and nobody teaches it because it cannot be examined.

That is why the readers' favourite, C, is the attractive answer and not the deepest one. The Pacific Premium conversation was the season's best visible work — a client restructured while his story could still be told as growth, angry in month one and stable by month nine. Choose it and the GM nods, because it is a story with a witness: the client himself. A is the same comfort in a different key — the clean yes that performs is what the book is built from, and it is also the entry every other banker in the building can show. D is the connoisseur's pick, and Aoraki deserves the pride. But all three share a property: they are the files where something happened, and judgment's full ledger includes the files where, because of you, nothing did.

The senior banker's answer is B, and the reasoning is practical, not sentimental. The declines are the only entries on the ledger with no witness, no performance data, and no one else keeping score — which means they are the only entries that will vanish entirely unless you record them. So the discipline: a private record of the deals you didn't do, revisited every year-end against what actually happened. The Kowhai increment — did Christmas vindicate the merchant or the bank? The Basalt project — how did it fare at the other bank? Torrent Bay's competitor structure — did anyone ever test that 65% advance? Most entries will resolve boring, and boring is the point: it calibrates. The occasional entry resolves catastrophic-elsewhere, and that is the only performance review the discipline of walking away will ever receive. Answer B in the meeting and the GM may look unconvinced. Answer B every year for a decade, with the record to show for it, and you become the person whose judgment the institution actually consults — which is a different job title from the one on your card, and better.

A last word for the season, because this file is also the Lab's. Eleven drills ago the question was whether a working capital increase should be approved. The questions since have gotten quieter and harder: what a silence discloses, what a covenant cannot do, when reporting is theatre, which signals move faster than numbers, how truth travels through a committee. If a residue is worth keeping from the whole season, keep this one: careers in credit are defined by the deals you don't do — but only if you write them down. The file you keep on yourself is the one that compounds. Season Two will open the corridor: the money starts crossing water. Bring the ledger.