How Efficiency Ratios Read a Lender's Operating Costs

An efficiency ratio measures what a lender spends to run itself against something it produced. Operating expenses over total income give one ratio, over assets another, and each answers a different question. Operating leverage is the relationship behind all of them: when income grows faster than costs, every efficiency ratio improves at once without anything else changing.

Operating leverage is where the trouble lives. The relationship lets a ratio improve while nothing improved. A reader who has not met the idea sees a cost ratio falling and reaches for the obvious explanation. Somebody controlled costs. A lender's costs can rise and its cost ratio fall in the same year, to the rupee, and the pairing is hard to overlook in anybody's figures once it has been seen.

What is an efficiency ratio in a lender?

The shape is always the same. Operating expensesWhat it costs to run the institution for the year: premises, systems, salaries and the rest. Counted before anything is set aside against bad loans and before tax. go on top. Something the lender produced goes underneath. The division gives a reading of what running the place costs against what running it achieved. The division is the whole of the machinery, and it takes ten seconds to learn.

The choice of what goes underneath is not a detail of presentation but the question itself, and that takes longer to learn. Three different denominators are three different questions, not three versions of one question. The answers they produce are not roughly similar. The three answers are not even the same order of magnitude. One expense figure at one lender in one year can read 2.10 per cent or 51.22 per cent depending on nothing but which line it was divided by, and both readings are correct.

Here is the everyday version, and it costs nothing to picture. Ten shops sit in one mall and pay roughly the same rent. Measured against each shop's sales, that rent gives one ranking. Measured against floor space, it gives a second. Measured against the number of customers who walked through the door, it gives a third. The rent never changed. Nothing about any shop changed between the three rankings. The question changed, and the question was hiding in the denominator the whole time.

Which efficiency ratios are worth building, and what is each one's base?

Three. A name that leaves the base out is the beginning of every argument on this subject, so each one is named with its base attached. Take Suvarna Commercial Bank Limited, invented, for one stated year. Its operating expenses for the year are Rs 5,040 crore. The expense figure of Rs 5,040 crore stays put through all three ratios.

The first ratio puts those expenses over total incomeNet interest income plus other income, added together. Total income is one of several lines a lender could divide its costs by, and not a self-evident choice.. At Suvarna Commercial Bank Limited that line is net interest incomeInterest earned less interest paid out, settled earlier in this material. Net interest income is what a lender makes on lending itself, before anything it charges for services. of Rs 7,440 crore together with other incomeWhat a lender earns that is not interest: fees, charges and commissions. Some lenders have a great deal of it and some have none at all. of Rs 2,400 crore, so it comes to Rs 9,840 crore for the year. Divide: Rs 5,040 crore against Rs 9,840 crore gives 51.22 per cent. The 51.22 per cent asks how much of everything that came in was spent on running the place.

The second puts the same expenses over total assets. Divide again: Rs 5,040 crore against Rs 2,40,000 crore gives 2.10 per cent, and that reading asks a completely different question: what does it cost to run one rupee of balance sheet for a year? The third narrows the base to earning assetsAdvances and investments taken together, which is what a lender holds that pays it interest. Part of any balance sheet earns nothing, so earning assets are narrower than total assets.. Earning assets are advances of Rs 1,44,000 crore together with investments of Rs 60,000 crore, adding to Rs 2,04,000 crore. The same expenses against that base give 2.47 per cent, and it asks the same question of a narrower base.

All three are correct and not one of them is the efficiency ratio. There is no definite article available here, and reaching for one is how a reader ends up comparing a figure struck on income with a figure struck on assets and concluding something about a lender. Suvarna Commercial Bank Limited is a listed commercial bank, and all three readings belong to it in one stated year.

The answers sit far apart because the bases sit far apart. Total assets of Rs 2,40,000 crore and total income of Rs 9,840 crore are not two versions of the same quantity. One is a stock of things the lender holds and the other is a flow the lender earned in a year, and the stock is enormously larger than the flow.

Why does the same expense figure produce four different answers?

There is a fourth reading, and it is the one most often met quoted without a base attached. Net interest income alone is Rs 7,440 crore, and the same Rs 5,040 crore over that base reads 67.74 per cent. The fourth reading is a legitimate measure. The measure asks what running the place costs against the income the lender makes on lending itself, leaving fees and charges out entirely.

So one unchanged expense figure, at one institution, in one year, reads 2.10, 2.47, 51.22 and 67.74 per cent, and a figure quoted without its base carries no information about which of the four it is. Sit with the spread for a second. The gap between the smallest and the largest of those readings is not a rounding argument or a matter of judgement. The gap is a factor of more than thirty, and every one of the four readings is the correct answer to the question it was asked.

The spread is why the base belongs in the name of the ratio and in the same sentence as the number, every time, without exception and without the reader having to ask. A colleague who says the cost ratio is sixty-eight has said nothing that can be acted on. A colleague who says operating expenses are 67.74 per cent of net interest income has said something that can be checked, argued with and put beside another lender's figure.

What is operating leverage in a lender?

Operating leverage is the plain fact that costs and income grow at different speeds, and that every ratio built out of the two therefore moves with the difference between those speeds rather than with either of them on its own. The difference between those speeds is the entire idea. Operating leverage is not a technique, a policy or a lever anybody pulls.

Operating leverage describes a cost structure rather than a decision taken each year, so crediting or blaming anybody for what it does is the standard mistake. Most of what it costs to run a lender is a committed costAn expense that is already in place before the year starts and does not move with how much business gets done. A lease signed last year is one; a licence fee is another.: premises leased, systems bought and running, staff hired, licences held. Premises, systems, staff and licences are all in place before the year opens. The income line is not settled in advance in the same way at all, so the year's income growth is what decides which direction the ratio travels.

The household version is exact enough to be useful. A household paying rent, school fees and a loan instalment has most of next year's outgoings settled before next year begins. If the salary rises ten per cent, the share of income going to those fixed items falls, and nobody in that household became more disciplined. If the salary falls ten per cent, the share rises, and nobody became careless. The committed items simply sat there while the other number moved.

What happens when income grows faster than costs?

Work it. Hold the operating expenses of Suvarna Commercial Bank Limited growing at 4.0 per cent for the year. Growth of 4.0 per cent takes them from Rs 5,040 crore to Rs 5,241.60 crore. Now suppose total income grows 10.0 per cent, from Rs 9,840 crore to Rs 10,824 crore. Divide: Rs 5,241.60 crore over Rs 10,824 crore is 48.43 per cent, against 51.22 per cent in the base year. Operating profitTotal income less operating expenses, taken before anything is set aside against bad loans and before tax. Provisions and tax are covered separately. goes from Rs 4,800 crore to Rs 5,582.40 crore, which is 16.30 per cent more.

Costs went up 4.0 per cent and the efficiency ratio improved, and both of those sentences describe the same year. Nothing was economised. Not one rupee of the Rs 201.60 crore of extra spending was avoided. The number underneath the ratio grew faster than the number on top, so the ratio fell. A ratio only ever reports on a relationship between two lines and never on either line by itself, so it has no way of showing which line moved.

The same arithmetic pushed one step further makes it obvious. Suppose income grows at exactly 4.0 per cent too, the same rate as costs, taking it to Rs 10,233.60 crore. Rs 5,241.60 crore divided by Rs 10,233.60 crore is 51.22 per cent, the base year reading again, unchanged to the decimal. Both figures are larger than they were. Operating profit is Rs 4,992 crore against Rs 4,800 crore, a rise of 4.00 per cent. The institution is bigger and the ratio has not moved a hundredth.

There is one setting of the income growth at which the ratio does not move at all, and it is worth naming because it is the cleanest proof that the ratio measures a gap rather than a level. The setting is the rate at which the costs themselves are growing. Grow income at 4.0 per cent alongside costs at 4.0 per cent and the ratio is frozen while everything it is made of is larger.

What happens when income falls and the costs do not?

Run the same year the other way and keep everything else identical. Operating expenses are still Rs 5,241.60 crore. A fall in income does not reach back and cancel a lease already signed. Total income falls 10.0 per cent instead, from Rs 9,840 crore to Rs 8,856 crore. The ratio reads Rs 5,241.60 crore over Rs 8,856 crore, or 59.19 per cent against 51.22 per cent in the base year. Operating profit goes from Rs 4,800 crore to Rs 3,614.40 crore, a fall of 24.70 per cent.

The same 10.0 per cent income move produced 16.30 per cent of operating profit going up and took 24.70 per cent away going down, off one identical cost base. Operating leverage is not a benefit and it is not a risk. Operating leverage is an amplifier, and it amplifies whichever way the year happens to go. A reader who has learned it as the good news half has learned half of an amplifier, and half an amplifier is worse than none.

Why is the fall bigger? Because the arithmetic is not symmetric around the base. Operating profit is what is left after a committed cost has been taken out of a moving income line, and a proportional fall in income takes its whole self out of that remainder while the cost stands still. The remainder is the smaller of the two numbers, so the same rupee movement is a larger share of it. The asymmetry needs no behaviour from anybody to explain it.

Why does a lender carry more of this than most businesses?

Two lists explain it, and neither is complicated. A lender's income moves with interest rates and with the size of its book, and both of those can change inside a single year. Rates are set outside the institution entirely. A book can grow or shrink faster than premises can be opened or closed. So the income line has several fast-moving inputs, none of which anybody inside the institution fully controls.

The cost base has the opposite character. Premises are leased for years. Core systems are bought, installed and lived with. Compliance work is staffed to a standard rather than to a volume. People are hired and trained on horizons measured in years. Every one of those moves slowly, and most of them move upward when they move at all.

The gap between how fast the income line can move and how slowly the cost base can move is the whole of the effect, and it is why a lender's cost ratios can look transformed in a year in which nothing about how it is run changed at all. That is the practical takeaway. When a shop's sales halve, its stock purchases halve with them, so a shop with mostly variable costs does not have this problem in anything like the same degree. A lender's premises do not halve.

Which efficiency measure can be carried across two lenders?

One of the three, and only one. Operating expenses over assets uses a denominator that means the same thing at both institutions. At Suvarna Commercial Bank Limited, Rs 5,040 crore over Rs 2,40,000 crore is 2.10 per cent. Rukmini Finance Limited, invented, lends but takes no deposits. Its Rs 540 crore over assets under managementThe book of loans a finance company is running. Its equivalent of a bank's total assets for a measure struck on the size of the balance sheet. of Rs 18,000 crore is 3.00 per cent, for the same stated year.

The income-based measures do not travel between these two at all, and the reason is worth seeing rather than accepting. Rukmini Finance Limited's operating expenses of Rs 540 crore over its net interest income of Rs 1,386 crore read 38.96 per cent. Put that beside Suvarna Commercial Bank Limited's 51.22 per cent and the finance company looks like the leaner of the two. But the bank's 51.22 per cent is struck on total income, and net interest income is 75.61 per cent of that. The finance company has no other income at all, so its income line and its net interest income line are one and the same. The two figures are not measuring the same width of income.

Change the base and the ordering of the two lenders reverses, so neither ordering settles anything. On an income base, the finance company reads lower. On an asset base, the bank reads lower. Nothing about either institution changed between those two sentences. Both readings are arithmetically correct and the two of them together rank the pair in opposite directions.

So the measure that travels is worth taking for exactly what it is. Operating expenses over assets at 2.10 per cent against 3.00 per cent shows what each spends to run a rupee of assets, and it stops there. The asset-based measure does not say which is better run. The two are lending to different borrowers, in different sizes, with different work attached to each rupee, and a finance company that does more work per rupee lent would show a higher figure whether it were run well or badly. The arithmetic supports a view rather than settling one.

Where do these figures stop, and what does that make each one?

Three things a reader would want are not in these figures, and naming each one is more useful than filling it in. Each absence is classified by what can still be derived from it: a ceiling, a floor, or nothing at all.

The first. The Rs 5,040 crore of operating expenses is never split between what was committed before the year opened and what moves with the size of the book. Most of a lender's cost base is committed, and that is a statement about how lenders are built rather than a reading off these figures. The committed share of that Rs 5,040 crore at Suvarna Commercial Bank Limited is unknowable, and the share is neither a ceiling nor a floor, since nothing divides the line at all.

The second. Each institution carries one stated year and no series whatever: no quarter, no earlier year, no month by month path. So the 48.43 per cent and the 59.19 per cent worked above are not readings of a second year. Both readings are constructions from two growth rates chosen to make the arithmetic legible, and what this institution's ratio would actually read next year is unknowable as well. Nobody can forecast that ratio without next year's income, and a lender's income moves with rates nobody inside the institution sets.

The third absence has an actual answer, and that separates it from the other two. Operating expenses over total assets read 2.10 per cent. Earning assets are a part of total assets rather than a different quantity, and so is any narrower base that might be chosen. Dividing an unchanged numerator by a smaller number can only produce a larger answer. So 2.10 per cent is a floor, and every cost-to-assets reading struck on any narrower base at this institution is at least that. The 2.47 per cent on earning assets is one such reading, and it sits above the floor exactly as it must.