A bank earns on the gap between two prices, using money belonging to other people, in a business where confidence is part of the machinery rather than a sentiment about it.
Banking gets described in complicated language, but the core is plain. A bank takes in money at one rate, lends it at a higher one, and lives on the difference. The sector's distinctive risks all grow out of that arrangement and out of how the funding is assembled.
The gap between what a bank pays for funds and what it earns on lending is the margin the whole business runs on. That margin responds to the general level of interest rates and to the relationship between short-term and long-term ones, since banks characteristically take money in on short terms and lend it out on long ones. Shift that relationship and bank profitability moves directly, which is why the sector reacts to rate conditions in a way most sectors simply do not.
Leverage is the second defining feature, and it is not a choice. A bank holds assets worth a large multiple of its own capital because its business consists of deploying funds belonging to others. No aggressive management decision produces this; it is what banking is. The arithmetic, though, is unforgiving. A fairly modest deterioration in asset values can wipe out a large share of the capital standing behind them. That is why banks can fail so quickly, and why they are regulated as closely as they are.
Third, the funding rests on confidence. Much of what a bank holds is repayable on demand or at short notice, while much of what it has lent cannot be called back quickly. The mismatch is not a defect. It is the function, converting short-term money into longer-term credit, and it works because the providers of funds do not all ask for them at the same time. They normally do not. When confidence wavers, the mismatch becomes acute with very little warning.
The risk inside the lending itself is that borrowers do not repay, and the timing of that risk complicates any reading of bank results. Loans get made in good conditions, when borrowers look sound and lending feels safe. The losses surface later, once conditions have turned. A bank's profits during an expansion can therefore look excellent precisely because the loans destined to go wrong have not yet done so. Current profitability is a poor guide to the quality of the lending that produced it.
That lag redirects the useful questions. What standards were applied when the lending was done, visible only in hindsight, and how much capital stands against the possibility of loss. A bank running thin capital alongside rapid loan growth during favourable conditions is accumulating a risk that will not appear in its results until the weather changes, and by then the lending is already on the books.
The financial sector also gathers quite different businesses under one heading. An insurer collects premiums against future claims and invests the money in between. An asset manager earns fees on funds it manages for others and carries almost none of the balance sheet risk a bank does. The similarities are thinner than the label suggests.
Banking's risks are structural rather than incidental, and reported profitability in this sector is unusually bad at revealing what is being accumulated beneath it. That is the ground on which questions about any particular institution have to be asked, offered here as ground and not as an assessment of one.