Net Stable Funding Ratio: Matching Funding to Asset Life

The net stable funding ratio asks a one year structural question: is stable funding at least as large as what the asset book requires. At Vindhya Commercial Bank Limited, invented, available stable funding of Rs 62,400 crore over required stable funding of Rs 57,600 crore is 108.3 per cent. The net stable funding ratio is not a longer coverage ratio.

There is a question every lender in the world eventually has to answer about itself, and it has nothing to do with whether it survives a bad month. The question is this: given how long the money a lender has lent out is going to stay lent out, is the money it borrowed to do the lending going to stay borrowed for anything like as long? A lender that funds a fifteen year loan with a fortnight of institutional money has not made a mistake this morning. The lender has made a mistake that arrives on a schedule, over and over, every fortnight for fifteen years, and every one of those mornings is a morning it has to persuade somebody to lend to it again.

The net stable funding ratioA measure of whether an institution's funding is stable enough, over a one year horizon, for the assets it is funding. is the measure built to ask exactly that. Its two sides are made out of judgements rather than out of contract dates, and the reason it is not the thirty day measure stretched out takes the longest to explain. Every figure here belongs to Vindhya Commercial Bank Limited, and every weighting behind those figures is that bank's own.

What does the net stable funding ratio actually ask?

Start with the everyday version. The mechanism is identical at any scale. A trader buys a small cold storage unit for a vegetable business. The unit will earn for ten years. She pays for it with money borrowed from a lender who wants it all back in four months. Nothing is wrong with the unit and nothing is wrong with the loan. The join between them is what is wrong. The earning is spread over ten years and the repayment lands in four months, so she must find new money roughly thirty times before the unit has paid for itself, and every one of those thirty times is a moment when somebody can say no. She is not short of cash today. She is short of certainty about a hundred and twenty tomorrows.

The join is what this ratio measures, and the ratio measures the join as a shape rather than as an event. The ratio is a structuralConcerned with the lasting shape of the balance sheet rather than with what happens in a particular stressed month. measure, and a structural measure asks about the ordinary standing arrangement of the balance sheet rather than about a bad week. The question is not whether the institution survives a shock, it is whether the funding shape it is running in calm conditions can go on running at all. A measure that only ever looked at stress would miss the mismatch entirely. The mismatch is fully present, fully visible and fully dangerous on a perfectly quiet Tuesday.

The horizon chosen for the question is one year. The one year horizon is not arbitrary and it is not a compromise. A year is long enough that overnight money, a two week interbank line and a three month certificate all fall clearly on the unstable side, and long enough that a ten year loan and a five year loan both fall clearly on the demanding side. A year is the horizon at which the shape shows up, and showing the shape is the entire point of picking a year.

So the measure has to do two things at once. The measure has to say how much of the funding can reasonably be expected still to be there in a year, and how much funding the assets are going to keep needing over that same year. Then it puts the first over the second. Stable fundingFunding an institution can reasonably expect to still hold in a year, which is a judgement expressed as a weight rather than a contract. is the name for the first of those, and the whole of the difficulty lives in deciding what counts.

What makes a rupee of funding stable?

Not its contract. Stability and contractual maturity are different things, and confusing them is the commonest way to misread the measure. A savings account is contractually repayable this afternoon on demand, and a three month certificate is contractually locked for three months, and yet the savings balance is treated as far more stable than the certificate. The treatment is not sloppiness. The weighting is a judgement about behaviour, and behaviour is what actually determines whether the money is there next year.

Think about who is on the other end. A person with a savings account keeps it because their salary lands in it, their rent goes out of it and moving it is a nuisance. A rival offering a quarter point more will not move them. In a bad week they will mostly not move the balance at all, for the simple reason that they are not watching. An institution holding a three month certificate is watching. The institution has a screen, a mandate and a colleague whose performance is judged on the rate obtained. When the certificate matures, it will go wherever the rate is best, and if the name has been in the news it will not come back at any rate. The stability of a rupee of funding is a statement about how closely the person holding it is paying attention, and that is why it is expressed as a weight rather than read off a maturity date.

Available stable fundingA weighted measure of funding, in which funding likely to stay for a year counts more heavily than funding that could leave tomorrow. is the total that comes out of applying those judgements. Every item on the funding side of the balance sheet is multiplied by a weight for how likely it is still to be there in a year, and the products are added up. Equity sits at the top. Equity does not leave, and nobody can ask for it back. Long dated borrowing comes next. Its return date is beyond the horizon anyway. Small, slow moving deposits come next, for the reason just given. Shorter deposits come below them. Overnight institutional money sits at the bottom, and in the extreme it counts for nothing at all. Money that can leave tomorrow morning cannot honestly be described as funding a loan that runs to the end of the decade.

What makes an asset require stable funding?

The other side runs on the mirror image of the same idea, and it trips people up because the direction reverses. On the funding side a heavy weight is good news. On the asset side a heavy weight is a demand. Required stable fundingA weighted measure of assets, in which a long or illiquid asset requires more stable funding than a short or liquid one. is the total that comes from asking, of every asset, how much funding it is going to keep tying up over the year ahead.

Two things drive it, and they are separate. The first is how long the asset is going to sit there. A loan with seven years left to run will still be there next year, so whatever is funding it must still be there too. A loan repaid in four months releases its own funding and stops asking. The second is how easily somebody else would take it off the institution's hands. A government security can be turned into cash quickly and at a visible price, so it can be funded with shorter money without the join snapping. A loan against a factory in a small town has no such buyer waiting, and the moment it has to be sold at short notice its price turns out to be a matter of opinion.

An asset requires stable funding to the extent that it is both long lived and hard to sell, and the two conditions compound rather than merely adding. A short liquid asset asks for almost nothing. A long illiquid asset asks for close to all of itself. Vindhya Commercial Bank Limited holds a funded exposure of Rs 1,680 crore to Nirjhar Industries Limited, invented, with a weighted average remaining life of 3.4 years and security taken over inventory and receivables. The exposure is a long asset with no queue of buyers, and it is exactly the kind of asset this side of the measure is built to charge for. The record does not say what the exposure actually contributes to the required side.

How is the ratio built, and what does it read at this bank?

The construction is the least interesting part, a relief after all the judgement in the weights. Add up the weighted funding side. Add up the weighted asset side. Divide the first by the second. Both sides are measured on the same one year view, both are stated in the same unit, and the result is a percentage that is above a hundred when the stable funding is at least as large as what the assets require, and below a hundred when it is not.

At Vindhya Commercial Bank Limited, invented, on a balance sheet of Rs 96,000 crore at the month 12 reporting date, the two totals are Rs 62,400 crore of available stable funding and Rs 57,600 crore of required stable funding. Dividing 62,400 by 57,600 gives 108.3 per cent. The two sides are Rs 4,800 crore apart, and that gap, rather than the percentage, is the number a treasurer actually manages.

Two pieces of context make the totals readable, and both are worth holding. Available stable funding of Rs 62,400 crore is 67.2 per cent of the Rs 92,880 crore this bank carries in deposits, borrowings and equity together. Required stable funding of Rs 57,600 crore is 60.0 per cent of total assets of Rs 96,000 crore. So roughly two thirds of the funding survives the weighting and roughly three fifths of the assets make a claim on it, and the ratio is what falls out of those two survivals meeting.

The trap sitting in plain sight in those totals

Required stable funding at this bank is Rs 57,600 crore. Net advances at this bank are also Rs 57,600 crore. The two figures are the same number and not the same thing, and there is no relationship between them whatsoever. One is a weighted measure computed across the entire asset side, taking something from cash, something from the investment book, something from every loan. The other is a single line on the balance sheet, being gross advances of Rs 58,800 crore less provisions of Rs 1,200 crore. Two different objects sharing a figure is one of the commonest ways a reconciliation gets quietly lost, and the rule here is absolute: name the object every time, and never let a reader infer a relationship from a coincidence.

How much room is there, and on which side?

The gap is Rs 4,800 crore, and that single gap gives two different answers depending on which side of the ratio it is read from. Available stable funding could fall by Rs 4,800 crore before the ratio reached 100.0 per cent, and Rs 4,800 crore is 7.7 per cent of Rs 62,400 crore. Required stable funding could rise by Rs 4,800 crore for the same result, and Rs 4,800 crore is 8.3 per cent of Rs 57,600 crore. The same rupee gap reads as two different percentages because the two denominators differ, so headroomThe distance between the current position and the point at which a measure reaches its threshold, which reads as a different percentage on each side. quoted without naming its side has left the reader to guess. The difference is about six tenths of a percentage point. Six tenths sounds small until it is the sentence a committee acts on.

Why is this not a longer version of the thirty day measure?

Vindhya Commercial Bank Limited also runs a thirty day measure, putting a stock of assets it could turn into cash against a modelled stressed outflow over the same month, and at month 12 that measure reads 125.0 per cent. The thirty day coverage ratio is covered separately on its own terms. Standing 125.0 next to 108.3 invites a comparison that has to be refused.

Because the comparison will be made. A reader who meets 125.0 and then 108.3 concludes that the bank is comfortable over a month and a little less comfortable over a year, and that the second number is the first one stretched. The conclusion is wrong in four separate ways at once, and each of the four is a different kind of wrongness.

Take the numerator difference first. Nobody ever sees that one. The thirty day measure counts a stock of things: assets that could be sold or pledged before lunch. The net stable funding ratio counts nothing of the sort. Its numerator is a statement about the institution's liabilities, about how long other people are likely to leave their money with it. One numerator is a pile of assets and the other is a judgement about depositors. The two numerators are not the same kind of object, and no amount of lengthening the horizon turns one into the other.

The denominators diverge just as sharply. The thirty day measure puts a modelled stressed outflow underneath, a description of a bad month that has not happened and may never happen. The structural measure puts underneath it what the assets require in ordinary conditions, with nothing going wrong at all. One denominator is a hypothesis and the other is a description of the balance sheet as it stands. And so the questions differ: survival against sustainabilityWhether a funding structure can continue without repeated refinancing risk, which is what this ratio is built to test.. Getting through a month is a question with an end date. Whether an arrangement can go on indefinitely is a question with no end date at all.

What moves the ratio, and what moves it fastest?

Only two things move it, and they move it in opposite directions. Lengthen the funding, by replacing money that leaves quickly with money that stays, and the numerator rises. Grow the asset book, particularly at the long and illiquid end, and the denominator rises. Everything a treasury actually does to this ratio is one of those two moves or a combination of them.

The arithmetic of the first move is unusually friendly, and it is the one genuinely operational thing this measure hands a treasurer. Hold the asset side still and the ratio is 62,400 plus x over 57,600, where x is what has been added to the stable funding side. Solved for round targets, it becomes a price list. Rs 960 crore takes the ratio to exactly 110.0 per cent. Rs 3,840 crore takes it to exactly 115.0. Rs 6,720 crore to exactly 120.0. Rs 9,600 crore to exactly 125.0. Going the other way, Rs 2,400 crore lost from the stable side takes it to 104.2 per cent, and Rs 4,800 crore lost takes it to exactly 100.0. While the asset side stands still, each percentage point costs a stated and constant Rs 576 crore of stable funding. A constant price per point turns a governance measure into something a treasurer can actually plan against.

The second move is not symmetric with the first, and the asymmetry catches people out. Growing what the assets require by Rs 4,800 crore takes the ratio to 100.0 per cent, and Rs 4,800 crore is 8.3 per cent growth in the required side. Losing Rs 4,800 crore of stable funding does the same, and that is 7.7 per cent of the available side. The rupees are identical and the percentages are not, for exactly the reason set out earlier. A further difference shows up in practice. A treasury can add term funding in weeks if the market will take its paper. The asset side moves on the timetable of the lending business and the borrowers, and that timetable is quarters rather than weeks.

What does this case not settle about the ratio?

The case has a limit, and it is worth stating exactly. Vindhya Commercial Bank Limited reports two totals and nothing underneath them. There is no schedule of which balance sheet items produced the Rs 62,400 crore, no schedule of which produced the Rs 57,600 crore, and no weighting factorThe percentage applied to a balance before it is counted on either side, which in this case is not given for any line and is therefore never stated. for any category on either side. Nothing is given to decompose either total with, so neither total can be decomposed.

The temptation at this point is to build a component table that looks complete, and that would teach a fiction that falls apart the first time the reader meets a real balance sheet. An invented split would carry an invented weight for savings balances, an invented weight for the loan book, an invented weight for the securities, and a reader would leave believing they had seen how the sides are built when they had seen a decoration. The mechanism is teachable without it. The arithmetic of the ratio is teachable without it. The one thing that is not teachable without it is any statement of the form this line contributes that much.

What that absence actually costs

An absence is only worth naming if it prevents something, so here is exactly what it prevents. Vindhya Commercial Bank Limited is running a live breach on its own internal cap on how much of its funding may come from institutional sources. Wholesale term deposits of Rs 11,520 crore plus borrowings of Rs 8,400 crore make Rs 19,920 crore against total liabilities of Rs 88,320 crore, and Rs 19,920 crore is 22.6 per cent against a cap of 20.0 per cent that the bank set for itself. Curing the breach would take Rs 2,256 crore of wholesale money replaced by retail deposits. Subtracting 2,256 from 19,920 leaves 17,664, and 17,664 over 88,320 is exactly 20.0 per cent.

Now ask the obvious follow up question: what would the cure do to this ratio? Retail deposits are more stable than institutional money, so the available side would be weighted more kindly, so the ratio would move up. The direction is all that can honestly be said. Without a weighting factor for either category the size of the move cannot be computed at all, and the correct professional answer is to name the direction, state that the size is not computable from this record, and refuse to produce a figure. A figure produced anyway would be a guess wearing the clothes of a calculation, and somebody downstream would put it in a paper.

Who actually reads this ratio, and what do they do with it?

Three people read it, and they read it for three different things. Three readers with three different uses is the usual sign that a measure is doing real work rather than filling a report.

Devendra Achar, who runs treasury at this invented bank, reads it as a price list. He is the person who has to find the funding, and the straight line above is his planning tool: he knows that while the lending book is flat, every point of ratio costs him Rs 576 crore of money that stays for more than a year, and he can compare that cost against what the same rupees would cost him raised short. He also reads the gap rather than the percentage. Rs 4,800 crore is something he can go and buy, and 8.3 per cent is not.

An analyst at another institution, looking at this bank as a counterparty rather than as an employer, reads it as a warning about the next twelve months of behaviour. A structurally short funded lender has to come back to the market constantly. Its appetite for a counterparty's business is hostage to its own refinancing calendar, and when conditions turn it will be bidding for money at the same moment everybody else is. The ratio tells an outsider how often this institution is going to need somebody to say yes. How often a counterparty needs a yes is more useful to know than almost any single figure on its balance sheet.

And a lending officer inside the bank reads it as a constraint they will feel indirectly. Every rupee of long dated lending drags in a demand for funding that stays, so a tight structural position makes long dated lending more expensive internally. The internal price is how a treasury measure ends up changing what gets lent, to whom, and for how long, without a single person in the branch ever seeing the ratio itself.

Take the household version once more. The mechanism does not change with the size of the balance sheet. A person paying for a twenty year home loan out of a salary is structurally sound: the funding, being the salary, keeps arriving for as long as the asset needs it. The same person paying the same loan by rolling a personal borrowing every six months is structurally unsound even if the payments are current. The arrangement itself has to be renewed forty times, and each renewal is somebody else's decision. Nothing about the house has changed. The join has.

Where does the standard come from, and what binds an Indian bank?

The mechanism above is jurisdiction free. Two weighted totals, one over the other, on a one year horizon, is a piece of arithmetic that would work anywhere. The numbers that make it bite are not jurisdiction free: which category gets which weight, what the ratio has to reach, how often it has to be reported and from when. Since the weights are the entire content of this measure, the implementation is not a detail sitting beside the standard, it is most of the standard as it actually applies to any given bank.