How to Read a Bank's Results

A bank's income statement reads like any company's, with two differences. Its revenue is mostly the gap between two interest rates rather than a price for goods. And one of its biggest costs — the charge for loans that will not be repaid in full — is an estimate that moves with the economy.

Here is the year of a bank constructed for this page, the same one the other pages in this order use: US$10.0bn of assets, US$8.0bn of them loans, funded by US$7.0bn of customer deposits, US$2.0bn borrowed from investors and US$1.0bn of shareholders' equity.

Every line above net profit is a decision or an estimate someone in the bank made. Read them in order, because each tells you something the next one cannot.

The margin. Net interest income is the bank's engine — US$250m here, five-sixths of its income. Against the US$9.5bn of loans and liquid assets that earn it, it is a net interest margin of 2.63%. The margin is read against the bank's own history and its own model: a bank lending to small businesses against cheap deposits runs a wider one than a bank lending to large companies against wholesale money, and both can be well run. What moves it, and which side of the balance sheet moves first, is the subject of Net Interest Margin and the Repricing Gap.

Other income. Fees, payments, foreign exchange and the like — US$50m, a sixth of the total. It matters less for its size than for its steadiness: a fee that comes back every year is worth more than a gain booked once.

Costs. Operating expenses of US$140m against US$300m of income give a cost-to-income ratio of 46.7%. The ratio rewards a bank for cutting costs and for growing income alike, and cannot tell you which happened, so read the two lines beneath it as well as the ratio. A bank that holds the ratio down by spending less on its systems or its credit staff may be borrowing from later years.

The impairment charge. US$32m, or 0.40% of the US$8.0bn of loans: the amount the bank added to its provisions this year for losses it now expects. Under the expected-credit-loss accounting of IFRS 9 the charge is forward-looking — a loan whose credit risk has not risen significantly carries an allowance for twelve months of expected loss, and one whose credit risk has risen significantly carries the loss expected over its whole life — so it rises with the bank's view of its borrowers before any of them has defaulted. It is a charge, not cash: the loans have not been written off. And it is the line that moves most from a good year to a bad one, for reasons section 04 shows.

Profit, and the return on it. Before the charge the bank earned US$160m; after it, and tax at 28%, US$92.2m — a return on equity of 9.2% on the shareholders' US$1.0bn. Whether that is good depends on what the shareholders require, which no annual report prints. This page assumes a cost of equity of 10% — an assumption, not a measured fact. On it, the bank earned US$7.8m less than its owners asked for: profitable, and still short.

A bank earns its profit on borrowed money — US$9.0bn of it here, against US$1.0bn of equity — so its results print a second set of numbers beside the profit, about the balance sheet the profit rests on.

Capital. The bank's risk-weighted assets are US$6.0bn: the US$10.0bn balance sheet weighted by how risky the rules treat each asset. Against them it holds US$780m of common equity tier 1 capital — the US$1.0bn of equity, less US$220m of deductions for intangible assets and the dividend it has declared. US$120m of subordinated bonds that count as tier 2 capital bring total capital to US$900m.

The Basel Committee's floor is common equity of 4.5% of risk-weighted assets, with a capital conservation buffer of 2.5% on top, and a regulator can add buffers for the credit cycle and for banks large enough to matter to the whole system. What the 13.0% is compared with is what this bank's own regulator requires of it — and what a reader watches is the direction of the ratio from one year to the next.

Problem loans. US$120m of non-performing loans — on the common definition, loans more than 90 days overdue, or unlikely to be repaid in full without the bank selling the security — is 1.5% of the book. Against the book the bank holds US$100m of provisions: US$60m against those loans specifically and US$40m across the performing loans. Provisions over non-performing loans is the coverage ratio, 83% here; the uncovered remainder is what security and recoveries must bring in.

Funding and liquidity. Loans of US$8.0bn against deposits of US$7.0bn is a loans-to-deposits ratio of 114%: the bank lends a billion dollars more than its customers deposit, and borrows the difference from investors, who can ask for more or stop lending sooner than most depositors do. Liquid assets of US$1.5bn are 21% of customer deposits. Where the regulator requires them, the results also print two Basel ratios: the liquidity coverage ratio, which asks whether the bank holds enough high-quality liquid assets to survive a stress lasting 30 days, and the net stable funding ratio, which compares the funding the bank can rely on over a year with the funding its assets require.

CAPITAL AS THE BUFFER, THE NUMBER → · COVERAGE, THE NUMBER →

Now hold everything still except one line. In a downturn the bank's borrowers struggle, its view of the losses ahead worsens, and the impairment charge goes from 0.40% of loans to 1.20%.

Income and costs have not moved, and profit has halved. The charge rose by US$64m; after tax, profit falls by US$46.1m, and the return on equity drops from 9.2% to 4.6% — less than half the assumed cost of equity. In a real downturn the other lines move too: loans that have stopped paying stop earning, lending slows, fees fall. But the margin and the costs are set by contracts and by staff, and they move slowly; the charge is an estimate of the losses across US$8.0bn of loans, and a change of less than one point of the book is US$64m. That is why it is the line that decides a bank's year — and why, in a good year, a reader asks what is keeping it low.

The bad year is still a profit. The US$160m earned before impairment is the first thing that absorbs a bad loan: the charge would have to pass 2.0% of loans before the bank made a loss. Only then would losses come out of the US$780m of common equity — the second buffer, and the reason the capital ratio is printed beside the profit. What happens inside the bank as the charge rises — arrears, provisions, write-offs, and who absorbs each — is What Happens Inside a Bank When a Loan Goes Bad.

STAGING, THE FIRST QUESTION → · THE WATCH LIST BEFORE THE LOSSES →

The profit above is statutory profit: the figure the accounting standards produce. Some banks also report an “underlying” or “cash” profit beside it, which leaves out items management regards as outside the run of the business. Such a figure can be useful, and it can flatter. Read the reconciliation between the two, ask whether the items left out recur — a cost that is one-off every year is a cost of the business — and notice which figure the commentary leads with. For a company reporting under IFRS Accounting Standards, IFRS 18 applies to annual periods beginning on or after 1 January 2027, and requires such management-defined performance measures to be disclosed in a single note and reconciled to the most directly comparable subtotal the standards define.

The second document is the regulatory one. The Basel Committee's disclosure standard — Pillar 3 — exists to promote market discipline: it asks a bank to publish key information on its regulatory capital and its risk exposures in a standalone report, which may be appended to the financial report but must be easy to find, and to publish it at the same time as the financial report for the same period. Its tables follow common templates, which makes it the place to compare one bank's capital, risk-weighted assets and liquidity with another's on the same basis. The Committee's standards have no legal force of their own; they reach a bank through its own country's rules, so the document's name and detail differ from one country to the next, and who checks them does too.

None of these numbers, read alone, says whether a bank is good. Read together and across several years, they say how it makes its money, how much of that it keeps, and what stands between a bad year and its depositors.