How a Bank Reads a Candidate

A bank that lends puts its money behind a business on incomplete information, and lives with the result for years. A bank that hires puts a seat, a salary, its training and its name behind a person, on less information still. It is the same institution making the same kind of decision, and it brings the same habits to both.

The habit is older than any interview format. In 1912 the counsel to a congressional committee asked J. P. Morgan whether commercial credit was not based primarily on money or property. “No, sir,” Morgan said; “the first thing is character.” The modern regulator says the same thing from the other end: collateral “should be considered the institution's second way out.” Character first, then the capacity to repay, security last. Read that order as a hiring rule and nothing in it needs changing.

So the useful question before an interview is not what they will ask. It is what a lender would ask: what paper is this panel writing, and what will it say?

Before anyone meets you, the bank has done what it does with a small loan: scored it. The online tests and the CV screen are a scorecard — quick, consistent, and built to say no cheaply. They are not a judgment of you. They decide whether a person will make one. Every round is in the Exam Hall to practise on, because practice is what a scorecard rewards.

The interview is the part of the decision a scorecard cannot do. The panel has your CV the way an approver has three years of accounts: as the appendix. What they are there to form is the paper — the reading of those facts that says what they mean and whether to back them. Either you bring that reading, or it is written without you.

A credit paper answers nine questions about a borrower, in an order approvers rely on. Each has a twin in the interview room, and each goes wrong in the room the way it goes wrong on the page.

Two things follow from the order. The recommendation comes first on the page and last in the preparation: you cannot write it until the other eight are done, and once it is written every other answer has somewhere to go. And the CV is section nine. It is the appendix, the place the facts are kept in full so that the body can speak in sentences. Reading it back to the panel is handing an approver the appendices and no paper.

A bank reads a borrower's cash before it believes its profit, because profit is partly an opinion and cash is not. A CV has the same two layers. The title, the society, the “led a team of five” are the profit line: true as stated, and partly an opinion. What converts is specific — what you were handed, what you decided, what changed and by how much, and who could confirm it.

A panel that hears a claim asks for the conversion, exactly as an analyst does. Do it yourself before you are asked. For each line you intend to rely on, write one sentence with a decision or a number in it. Three such lines are worth more than a full page of titles. It is what the interview page means when it says to prepare the behavioural questions as evidence, not as stories.

Security matters. It decides whether many loans can be written at all, and a degree, a grade average and a referee decide whether many applications are read. But no competent lender advances money because the security is good. Security is what the bank falls back on when the first way out has failed.

An application that leads with its security — the university's name, the grade, the scholarship — is making the case a borrower makes when the cash flow will not carry the loan. Lead with what repays: what you can already do, shown. Let the security sit where it sits in a paper, in section six, stated once and exactly as it is. A failed paper stated plainly costs less than one the panel finds for itself.

A paper that lists fifteen risks has not decided which ones matter, and a paper with no residual risk anywhere has not finished thinking. Both failures have exact twins in the room. The candidate who offers a strength in disguise has written a risk section with nothing left over, and the panel reads it as an approver would: this writer either has not looked or will not say. The candidate who lists everything has not decided.

The form that works is the paper's. Name the two or three gaps that could actually cost the bank something if it hired you: no banking internship, a failed paper, a habit of going quiet in a group. Say what reduces each, with something that has already happened rather than an intention. Then say what is left.

The residual is the part a panel listens to most closely, because it is the risk they are choosing to take. Naming it yourself is the strongest evidence of the thing they are hiring for, which the book this site works from calls intellectual honesty — one of its Four Signals.

Many declines in lending are declines of a request rather than of a borrower. Issue 01 on this site ends in exactly that — “No, as framed” — with a different conversation on offer. A rejection after an interview is usually the same kind of no: this seat, this intake, this evidence.

A borrower who is declined asks what would change the answer, and comes back with it. Ask the same question, in those words, wherever the bank offers feedback, and treat what you hear as a condition to meet rather than a verdict to absorb. The borrower's side of that conversation is in When the Bank Says No, and almost all of it transfers.

Three places, and they matter.

A lender's return is capped at its interest, so it reads a file almost wholly for what could go wrong. An employer keeps the upside. It will sometimes back promise that no lender would, so show the ceiling as well as the floor.

A lender can verify. It has audited accounts, a valuation, a register of security. A panel has an hour and your word, so it verifies the only way it can: with the follow-up question. The answer to the second question is the evidence for the first.

And a person is not a balance sheet. The oldest rule in lending puts character ahead of every number, and none of this is a way of reducing yourself to ratios. It is a way of knowing which question you are being asked.

Not technical preparation. If the seat has a technical round — a coding exercise, a modelling test, investment-banking technicals — that is a different genre, and nothing here helps with it.

Not a script. The one-page paper is for writing before the interview, not for reciting in it. A memorised answer is usually detectable within two sentences.

Not a guarantee, and not advice. Panels differ, markets differ, and this is written from commercial banking practice in New Zealand and Australia. The candidate on the filled-in page is a fictional composite.

The one-page version comes free when you subscribe, at the foot of this page: the nine sections, blank, and the same page filled in for a composite candidate. Write it before the first interview and revise it after each one.

Then do the other half of the preparation, which is the half the panel is hiring: read a business you have never seen, and reach a view. Issue 04 is a clean file where the disciplined answer is yes. Issue 01 looks the same on the surface and is not. The question that decides most interviews — “Walk me through this borrower” — is section 07 of the interview page.