The Liquidity Adjustment Facility: How the RBI Manages Daily Liquidity
The Liquidity Adjustment Facility is the arrangement through which the Reserve Bank of India (RBI) lends to the system when the system is short and absorbs from it when the system is in surplus. The facility exists so the overnight rate stays inside the corridor, and it works in both directions because a system can hold too much as easily as too little.
Three things already established do the work in that sentence, and each is covered separately. Start with the corridor: a floor, a policy rateThe single interest rate a central bank decides on and announces. It is the reference the rest of short-dated borrowing prices off, and it is changed by a decision rather than by anything happening in the market on a given day. at its middle and a ceiling, all three settled by a decision taken elsewhere and not on any day the rate happens to move. Next comes the position: the system taken as a whole is either holding more funds than it needs or fewer, and that gap is a quantity a figure can be put against. Last comes the market where short-dated funds actually change hands, and that is where the overnight rateThe interest rate on money borrowed today and repaid on the next working day. It is the shortest borrowing there is, which is why it responds to a shortage or a surplus of funds faster than anything else. comes from in the first place.
Put those three together and a gap appears that nobody has yet filled. A corridor is a pair of numbers. A position is a quantity. Neither of them, on its own, does anything. Something has to stand between the position and the corridor and make the second hold while the first moves around. The Liquidity Adjustment Facility is what stands between the position and the corridor.
The Reserve Bank of India and the Liquidity Adjustment Facility are named as themselves. Saying what a facility of this kind is designed to do is a different act from describing what it did. Every figure below belongs to Sankhya, an invented country whose corridor and position were built small enough to be checked by hand.
What kind of thing is the Liquidity Adjustment Facility?
One distinction decides everything that follows, and it is clearest taken from somewhere ordinary first.
A co-operative in a market town keeps a counter open every working day. The counter has two windows. At one, a member who is short can draw funds against what they have deposited. At the other, a member who has more than they need for the day can place funds and take them back tomorrow. Now ask what the counter amounts to. The counter is not the drawing and it is not the placing. The counter is the fact that both windows are there, at known terms, whether or not a single member walks in today. On a quiet day nobody uses it and it has not gone anywhere. On a busy day a queue forms at one window and nothing about the counter has changed except the queue.
A counter of that sort is a standing arrangement, and the Liquidity Adjustment Facility is a standing arrangement of exactly the same kind. The Liquidity Adjustment Facility is an arrangement rather than an event. The facility has a side facing each direction: the system can borrow through it when it is short, and it can place funds through it when it holds more than it needs. Its existence is a fact about the arrangement. Its use is a fact about the day.
The Liquidity Adjustment Facility is defined by what is available through it, not by what passed through it, and confusing those two is where most misreadings of the subject begin. A reader who thinks of the facility as something that happens will look at a day with little activity and conclude the arrangement has been withdrawn. Concluding that is like concluding the co-operative has closed because nobody queued on a wet Tuesday.
India, and only so far as the naming goes
The Liquidity Adjustment Facility is an Indian arrangement operated by the Reserve Bank of India, and both are named here as themselves. The corridor of 5.75, 6.00 and 6.25 per cent used throughout belongs to Sankhya and to nowhere else. The corridor actually in force, and the terms on which the real facility stands, are published by the Reserve Bank of India and move only when a decision moves them.
A reader looks at a stretch in which very little passed through the Liquidity Adjustment Facility in either direction, and wants to know what kind of thing they are looking at. What is the correct reading?
What problem does the Liquidity Adjustment Facility solve?
Go back to the market town, and this time watch a price rather than a counter.
Suppose the market committee announces that onions will trade between Rs 18/- and Rs 22/- a kilo. If the committee does nothing else, that announcement is a description of what it hopes will happen. On a day when three trucks fail to arrive, buyers who need onions will pay Rs 30/-, and the announcement will have been powerless to stop them. A number painted on a board has no way of stopping a person who needs a thing from paying what it takes to get it. Now change one detail. The committee keeps its own stock, and states that it will sell to anyone at Rs 22/- and buy from anyone at Rs 18/-. Nothing about the announced range has changed. Everything about what it does has. Rs 22/- is available, so nobody now pays Rs 30/-. Rs 18/- is available, so nobody now sells at Rs 15/-.
The onion committee's problem is the problem the Liquidity Adjustment Facility solves, transposed from onions to funds. A corridor with nothing standing at its ends is a description of where the overnight rate is expected to sit. Holders of funds with nowhere better to put them will accept less and less, so a system in surplus pushes the rate down toward the floor and then through it. A borrower who must have funds today will pay what it takes, so a system short of funds pushes the rate up toward the ceiling and then through it. Without something available at each end, a corridor is a forecast rather than a boundary, and the Liquidity Adjustment Facility is what converts the one into the other.
Notice what the conversion costs. The conversion costs the willingness to actually transact at the stated ends, in whatever size turns up. A committee that says it will sell at Rs 22/- and then runs out of onions has a ceiling until the third truck fails and no ceiling afterwards. The same logic applies to the ends of a corridor, and that is why a facility of this kind is described as standing rather than as occasional.
The Sankhya corridor is stated as 5.75, 6.00 and 6.25 per cent. What is it about the Liquidity Adjustment Facility that makes those three numbers behave like a boundary rather than an expectation?
Why does the Liquidity Adjustment Facility have to work in both directions?
Two-sidedness is the part most readers skip, and skipping it makes the rest of the subject look like an accident of design rather than a consequence of the problem.
Take the market committee again and give it only half the arrangement. The committee keeps a stock of onions and will sell at Rs 22/-, but it will not buy from anyone at any price. Two kinds of day follow. On short days the price stops at Rs 22/-, exactly as intended. On glut days, when four extra trucks arrive and every seller wants to clear stock before it spoils, the price falls to Rs 14/-, to Rs 11/-, to whatever a buyer feels like offering. The announced range has one live end and one dead one. Reverse the arrangement, letting the committee buy at Rs 18/- but never sell, and the failure simply changes ends: gluts are handled and shortages are not.
The money marketThe market in short-dated borrowing and lending, where funds change hands for a day or for a few weeks at a time. It is covered separately, and appears here only as the place the overnight rate is set. has both kinds of day, and they are not rare exceptions to a normal state. A holder of funds with no better use will lend at a lower and lower rate rather than hold them idle, so a system carrying more funds than it needs pushes the overnight rate down. A system short of funds pushes the rate up, because a borrower who must settleTo complete a payment that has fallen due, so the funds actually leave one account and arrive in another on the day they are owed rather than at some later point. today will pay more and more. Surplus and shortage are mirror images, and neither pressure is unusual.
So the two-sidedness is not a courtesy extended to both kinds of participant. A facility with only one side would hold one end of the corridor and leave the other end open, so the corridor would be a boundary in one direction and a description in the other. Lending to the system when it is short is what keeps the overnight rate off the ceiling. Absorbing from the system when it is in surplus is what keeps the rate off the floor. Remove either half and the whole structure stops being a corridor and becomes a line with a wall on one side of it.
Suppose an arrangement of this kind could only lend to a system that is short, and could never absorb from a system in surplus. What would follow for the Sankhya corridor of 5.75, 6.00 and 6.25 per cent?
Why does the overnight rate stop moving once the shortage gets large enough?
Here is where Sankhya earns its keep.
Sankhya's corridor is fixed at a floor of 5.75 per cent, a policy rate of 6.00 per cent and a ceiling of 6.25 per cent. Sankhya also runs a plain and entirely made-up rule for how the position pushes the overnight rate about: every Rs 1,00,000 crore of net surplus or shortage shifts the rate by a quarter of a percentage point, and the corridor then holds whatever comes out of that. The rule is a teaching device rather than a description of any market. The rule is a straight line because a straight line can be checked by hand, and the point being made survives whatever shape is put in its place.
Work two shortages. With Sankhya short by Rs 1,00,000 crore, the rule gives 6.00 plus 0.25, landing on 6.25 per cent, the ceiling exactly. Now put the shortage at twice that size. The rule on its own gives 6.00 plus 0.50, or 6.50 per cent. The rate observed is 6.25 per cent. The shortage doubled. The rate stood exactly where it was.
The obvious reading of that is wrong, and the wrong reading is worth naming because almost everybody reaches for it first. The wrong reading is that 6.25 per cent is a limit the market is obeying, as though participants had agreed not to deal above it. Nothing of the sort is happening. The rate stops at 6.25 per cent because funds can be had at 6.25 per cent, and a borrower who can obtain funds at that price has no reason on earth to pay 6.30. The ceiling is not a rule the market is following. The ceiling is a price at which funds are available, and availability at a price is what stops a price, in this market and in every other.
Reading the ceiling as a price rather than a rule shows what would break the ceiling and what would not. A larger shortage would not. A bigger queue at a counter that will serve everyone does not raise the counter's price. The ceiling breaks only when the funds are not actually available at that price to somebody who turns up for them. Such a failure is a statement about the arrangement rather than about the size of the shortage.
Sankhya's shortage doubles from Rs 1,00,000 crore to Rs 2,00,000 crore and the overnight rate stays at 6.25 per cent. What is the correct explanation?
What does the whole Sankhya position look like when it is run across the facility?
The five cases below are the whole of the Sankhya worked instance, and the set carries something no single row does, so they are worth reading together rather than one at a time.
Read the rate column downward and it climbs: 5.88, 6.00, 6.10, 6.25, and then 6.25 again. Read the position column downward and it swings from surplus through balance into a shortage that keeps growing. Now read the fourth column, the one that matters most and the one most readers never look at. The policy rate is 6.00 per cent in every single row, so nothing in this table is a decision, and every difference between the rows comes from the position rather than from anybody changing their mind.
| The Sankhya net position | Which side of the facility the system would be using | Overnight rate | Policy rate |
|---|---|---|---|
| Surplus of Rs 50,000 crore | The system places funds | 5.88 per cent | 6.00 per cent |
| Balanced | Neither side in use | 6.00 per cent | 6.00 per cent |
| Shortage of Rs 40,000 crore | The system borrows | 6.10 per cent | 6.00 per cent |
| Shortage of Rs 1,00,000 crore | The system borrows | 6.25 per cent | 6.00 per cent |
| Shortage of Rs 2,00,000 crore | The system borrows | 6.25 per cent | 6.00 per cent |
Work the sums yourself, because they are small and because a reader who has worked them argues with the conclusion rather than with the figures. Take the surplus row first. Rs 50,000 crore is one half of the Rs 1,00,000 crore step, so the shift is one half of 0.25, giving 0.125, and 6.00 less 0.125 comes to 5.875, printed as 5.88 per cent. The Rs 40,000 crore row is 0.4 of a step, so the shift is 0.4 times 0.25, giving 0.10, added this time rather than taken off because the system is short: 6.10 per cent. The Rs 1,00,000 crore row is one whole step, so 6.00 and 0.25 give 6.25 per cent, sitting on the ceiling. The Rs 2,00,000 crore row is two whole steps and would come to 6.50 per cent. The corridor does not allow that rate, so 6.25 per cent stands. Four short sums produce every figure in that table.
Move the Sankhya position and watch where the overnight rate can and cannot go.
The panel opens on the balanced case: neither side of the facility in use and the overnight rate at 6.00 per cent, exactly the second row of the table above. Dragging toward either end moves two markers. The solid one is the overnight rate. The hollow one is where the illustrative rule on its own would have put it, and once the two separate the corridor is doing its work. Pushed far past the point where the solid marker stops, the solid marker refuses to move while the hollow one keeps going.
The panel has moved from a surplus of Rs 50,000 crore to a shortage of Rs 1,00,000 crore, and the side of the facility in use has flipped from placing funds to borrowing. Has policy changed?
Why is a day's operation not a policy decision?
The distinction between a decision and an operation is worth being slow about.
Consider a household that has agreed a monthly budget. The budget is a decision: somebody sat down, looked at what was coming in and what was going out, and settled on a figure. Then there is the tin on the shelf. Some days money goes into the tin because the vegetable seller charged less than expected. Some days money comes out because a school fee fell due early. An observer watching only the tin would see constant movement in both directions and might conclude the household was forever revising its plans. The observer's conclusion would be wrong. The plan has not moved at all. The tin is how the household gets through the gap between when money arrives and when it is needed, and getting through that gap is a different activity from deciding how much to spend.
The corridor is the budget and the facility is the tin. The corridor is set by a decision, taken deliberately, and moving it is an act of intent. Keeping the overnight rate inside the corridor is an operation, carried out as the position requires, and the position moves for reasons that have nothing to do with anybody's intent. An operation is not a signal, and reading each day's operation as a change of intent is the single most common error made about the facility.
The test is mechanical and can be applied without knowing anything else. The question is whether the corridor moved. If the floor, the policy rate and the ceiling are where they were, then what is on view is the position being managed inside an unchanged structure, however energetic it looks. If the corridor moved, that is a decision, and it will have been taken and announced as one. The two things live at different levels and are not versions of each other. Where the band comes from, how it gets settled and what moving it is meant to achieve all have their own treatment, alongside the rate that sits at its middle.
Two things could change on a given stretch: where the corridor sits, and where the overnight rate sits inside it. Which of the two is settled by a decision rather than by the position?
The reading that turns an operation into news
Here is the mistake in the form it actually arrives in. A reader looks at a stretch of Sankhya in which the system was short throughout and used one side of the facility heavily and repeatedly. The reader writes that Sankhya's central authority has shifted its stance and is now supplying funds more readily than before.
Now put that reading against the worked instance above. Across all five cases the policy rate is 6.00 per cent. The floor is 5.75 per cent and the ceiling is 6.25 per cent in every one of them. Not one number describing the corridor differs between the first row and the last. The position differs, swinging from a surplus of Rs 50,000 crore to a shortage of Rs 2,00,000 crore, and the overnight rate differs with it, following the position to 6.25 per cent and then stopping at the ceiling. The reader has taken the tin for the budget, and the cost of that mistake is a stated change in intent where no intent moved at all.
The fix is one question asked before any other. Did the corridor move? If the floor, the policy rate and the ceiling are unchanged, then the thing being observed is the position being handled inside a structure nobody touched, and it carries no information about intent whatsoever. Repetition does not change this. A shortage that runs for a long stretch is a long-running shortage, and a long-running shortage handled through an unchanged corridor is still an unchanged corridor.
A reader reports that heavy and repeated use of one side of the facility, across a stretch in which the corridor never moved, amounts to a change in policy. What has gone wrong in that reading?
What can the Liquidity Adjustment Facility not do?
Every mechanism worth learning has a boundary, and this one's boundary is sharper than most because the thing it cannot do looks so much like the thing it can.
Return once more to the market committee. The committee can hold the price of onions between Rs 18/- and Rs 22/- for as long as it is willing to buy and sell at those prices. The committee cannot make the shopkeeper at the end of the lane extend three weeks of credit to a customer whose last two payments were late. The committee has a price. The committee has no view on that customer, no way of forming one, and no means of acting on it if it had.
The Liquidity Adjustment Facility sits in exactly that position. The facility manages where the overnight rate sits at the short endThe very shortest stretch of the borrowing scale, where funds are lent for a day or a handful of days rather than for years. Where that borrowing happens is covered separately., and it does so by making funds available at the ends of the corridor. The facility does not decide whether a loan is extended to any particular borrower, it does not reach anybody's creditworthinessA lender's judgement of whether a particular borrower will repay. It is formed from the borrower's own record, income and security, and it is covered under credit and lending rather than here., and it does not move the corridor. The corridor is settled elsewhere and by a different act. The three limits are not gaps somebody forgot to fill, but the ground a facility of this kind is not built to touch.
The case where this boundary bites hardest is a credit crunchA stretch in which lending falls sharply even though the price of borrowing has not risen and may have fallen. A crunch is covered separately, and named here as the case this facility cannot reach., in which lending contracts even while the price of borrowing is coming down, and the facility can do nothing about it whatsoever. The reason is worth stating plainly: the facility works on the availability and the price of funds at the short end, and a crunch is a situation in which the price of funds no longer decides whether credit is extended. A tool aimed at price cannot fix a problem that is no longer about price. A crunch is set out on its own, under lending that stops regardless of rates.
Sankhya's lending contracts sharply over a stretch in which the price of borrowing is falling rather than rising. What can the Liquidity Adjustment Facility do about that?
What does an analyst actually read off the facility?
Two questions, in this order
Someone covering short-dated conditions for a living does not read the facility for drama. Analysts read it for two things, and the order matters because the second is useless without the first.
The first question is direction. Which side of the arrangement is the system using? If funds are being placed, the system has been carrying more than it needs. If funds are being taken, it has been short. Direction is a fact about which way the position has been running, and it is the closest a reader gets to a direct read on the position from the short end. In the Sankhya table above, direction alone separates the first row from the last three.
The second question is persistence. Has the system been on the same side for one stretch or for many consecutive ones? Direction shows which way the position has been running, and persistence separates a passing mismatch from a standing condition, and neither figure means anything without the other. A single stretch on one side is ordinary: money moves about for reasons that have nothing to do with anything structural. The same side, over and over, is a different observation. A mismatch that keeps recurring is describing a condition rather than an accident.
An analyst never treats either figure as intent. Both are readings of the position. Neither is a reading of what anybody has decided, and the moment a note starts inferring the second from the first it has left what the data can support. The composite read on how tight or loose conditions actually are is a separate exercise that draws on several series at once, and it is covered on its own.
Where to read about the facility itself
The Liquidity Adjustment Facility and the Reserve Bank of India are real, and the terms on which the facility stands are set out by the body that operates it. The sites below are where those terms are published, and a rate or an amount taken from anywhere else is a copy of a record that has since moved on.
Why a rate read off an arrangement stops being true. A rate in force, a position outstanding, an amount that went through an arrangement: each of those is a record somebody has to keep up. Copied into a sentence, it is accurate on the day it is copied. Then it decays without ever announcing that it has, and a confident-looking sentence sits in a reference long after it stopped being true. The institution running the arrangement keeps that record itself and stamps it, and a stamped record is the only version worth quoting.
Checked against
| Source | Document | Site |
|---|---|---|
| Reserve Bank of India | The facility set out in the words of the body that operates it | rbi.org.in |
| Reserve Bank of India | Material on how the rate band is operated in practice | rbi.org.in |
| Bank for International Settlements | Published material on central bank operating frameworks and standing arrangements | bis.org |
Sankhya, the market town with its onion committee, and the household with the tin on the shelf are invented.
Educational material. Not advice on any investment, tax, budget or market position.