Deposit Insurance: What It Covers and What It Does Not

Deposit insurance is a promise from a separate body, not from the institution itself, that if the institution cannot hand a balance back, the body will, up to a stated amount and on stated terms. The amount, the unit it is counted per, the accounts that qualify and the institutions inside the scheme are all set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and every one of them moves.

No figure describing the scheme appears below. Not the amount a depositor is covered for, not the premium anybody pays, not the time a settlement takes, not a single number describing the scheme in any form. Every one of those figures is set by an authority, and every one of them moves. A printed figure that has quietly changed still reads as an answer, and the reasoning behind leaving each one blank is set out in full further down.

The mechanism does not move at all, and the mechanism is what comes next. Why the promise has to come from somewhere other than the institution. Why the scheme is not really about paying anybody. Why a rarely asked question, about the unit the cover is counted per, decides more of the answer than the amount does. And where every blank left below can be filled in, in ten minutes, from the source.

What is the promise, and who is making it?

A deposit is where everything else hangs off, so it comes first. Money put into an institution is not a box left in a locker with the depositor's name on it. The money has been swapped for a promise, and the promise is that the money comes back when the depositor asks. A promise of that kind is a demand liabilityAn amount an institution owes that the person owed it can ask for at any time, without waiting for a date to arrive. What the institution did with the money in the meantime is a separate question.: an amount the institution owes and that the depositor can call for at any moment, without notice and without giving a reason.

A promise to repay on demand is worth exactly what the promisor's ability to keep it is worth on the morning everybody asks. And here is the awkward part. Asked what happens when it cannot pay a depositor, the only answer the institution can give from its own resources is nothing at all. A promise by an institution to pay a depositor if it cannot pay that depositor is not a promise, it is a sentence. The sentence has the grammar of a guarantee and none of the substance.

So the promise has to come from somewhere else. Consider a borrower who offers, as reassurance, that if they run out of money they will repay the lender anyway. No lender would accept that, and refusing it would not be unkind. Now suppose a different person, with their own money and their own income, signs alongside and says that if the borrower cannot pay, they will. A second signature is a different thing entirely, and the difference is not the wording. The difference is whose pocket the second promise comes out of.

Deposit insurance is that second signature, at the scale of a whole system. In India the body doing the signing is the Deposit Insurance and Credit Guarantee Corporation, at dicgc.org.in. The arrangement works only because the second promisor is genuinely separate: its ability to keep its promise does not move when the institution's balance sheet moves. Take that separation away and what remains is not a weaker version of deposit insurance. Nothing at all remains.

Who pays for the cover, and does a depositor pay anything?

Almost everybody arrives at this question with the same assumption, and it is wrong in a specific and useful way. Because no depositor has ever been handed a bill for it, the cover is assumed to be free. Nothing appears on a statement. Nobody signs up. Nobody is offered the choice of going without and paying less.

The cover is not free. The institutions inside the scheme pay a premium into it, and the cost of that premium sits inside the price of everything those institutions do. The premium is in the rate offered on balances, in the rate charged on lending, in the charges on services. A depositor pays for it the way a resident pays for the lift in a building: no cheque ever leaves anybody's hand with the word lift written on it, and the lift is unquestionably being paid for.

Why does that matter beyond the satisfaction of getting the fact right? Because free things are worth learning about only out of curiosity, and things already being paid for are worth learning about whether anybody is curious or not. A depositor is buying this cover indirectly whether they know it or not. Knowing what cover reaches and what it does not reach is therefore rather less optional than it looks. Being already paid for is the whole justification for everything below.

Who pays, on what base and at what rate is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in. The row for the premium is drawn below with the value left to the source, and the reason it stays that way comes further down.

What is the cover for, if it is not compensation?

Here is the sentence that takes some getting used to. Paying people back is the work of the day the scheme is used. The scheme is not built for that day.

The scheme exists so that nobody has a reason to ask. A depositor who knows their balance is good whatever happens to the institution has no reason to be early, no reason to be at the front, no reason to act on a rumour. And a scheme that has produced that state of mind across a whole system has never had to hand over a single rupee. The outcome it was built to prevent did not happen.

The same thing is familiar in another setting. Consider a fire escape in a building. Its purpose is not to carry people out of a fire. Its purpose is that the building is one people are willing to sleep in. A fire escape that has never been used has not been wasted; it has been doing its job every single night. Judging it by how many people it has carried out would be judging it by the number of fires, and the number of fires is the one number a fire escape exists to keep at zero.

The scheme is set out below as a shape and a route rather than as a settlement anybody can point at. A settled claim shows what one scheme did once, at one institution, under rules that have since moved. The shape shows the work done on every night the scheme is not used.

Why does the same arrangement support two completely different outcomes?

Seeing why the reassurance is the product takes one structural fact about the institution, and only one. The institution holds money that is repayable on demand and lends that money out for longer. Holding on demand and lending for longer is maturity transformationTaking in money that can be asked for at short notice and putting it to work for a longer period. It is the reason an institution can be entirely sound and still be unable to pay everybody at once., covered separately and worked through in full there.

Now hold that fact still and ask what a sensible depositor should do. The honest answer is that it depends entirely on what everybody else does, and that is not a dodge, it is the finding.

If each depositor expects the others to leave their balances where they are, then leaving one's own alone is the sensible thing to do. The institution is not called on to produce more than it can produce, everything settles normally, and everybody's expectation is confirmed. If instead each depositor expects the others to ask early, then asking early is the sensible thing to do. The rule pays people in the order they arrive, and being late is the one position nobody wants. Enough people acting on that expectation is a runWhat happens when many depositors ask for their balances at the same time. What can leave, how fast it can leave, and what selling quickly costs are worked through separately., and that expectation is confirmed too.

Both of those outcomes are stable, and nothing inside the institution decides which one happens. Not the quality of what it has lent, not how carefully it is run, not its net worthWhat is left of what an institution has after everything it owes is taken away. It belongs to the people who put up the institution's own capital, and it is a separate test from whether the institution can pay today.. The same balance sheet, unchanged in every particular, supports both. The structure of that argument belongs to Diamond and Dybvig, Bank Runs, Deposit Insurance and Liquidity, 1983.

There is a wedding buffet version of this, and it is exact. One counter, plenty of food, no shortage of anything. If everybody believes everybody else will wait their turn, standing near the back costs nothing and everybody waits. If word goes round that people are rushing, then rushing is the right move for each guest as well, and the fact that there was always enough food makes no difference to anybody who ends up at the back. Nothing about the food changed. Only what each guest expected of the other guests changed.

Somebody who asks for their money early is reading a rule correctly. The rule pays in the order people arrive. Reading a rule correctly and acting on it is not panic and it is not a character failing. Any account of this that treats early askers as the problem has quietly turned a structural fact into a complaint about people.

Why can an announcement do the work that a payment would do?

Put together, the last two parts yield something rather elegant. There are two outcomes. The one nobody wants is held up entirely by what each person expects of everybody else. So removing the reason for that expectation does not manage the bad outcome or make it less likely. Removing the reason takes the bad outcome out of the set.

Taking the bad outcome out of the set is what the second promise does. Only one outcome is left, so a depositor who knows the balance is good either way is no longer choosing between two. There is nothing to be early for. The rule about who gets paid in what order stops mattering to that depositor. Their position in the queue no longer changes what they end up with.

Notice what has not happened here. No money has moved. Nobody has been handed anything. The institution's balance sheet is exactly where it was. The only thing that changed is what a depositor knows about what happens if the worst happens. The bad outcome was never made of anything more solid than expectations in the first place, so knowing turns out to be enough.

Which gives the measurement rule that most people get backwards: a scheme that is never used has done its entire job. Counting how often a protection scheme has been called on is counting the number of times the outcome it exists to remove happened anyway. The fire escape is the same case. Low usage is the result, not the shortfall.

What sits inside the scheme, and where is that settled?

Four questions decide whether any particular balance is inside the scheme, and all four appear below with nothing written in the answer column. An empty answer column is generosity rather than evasion, and the reason it is the honest version comes further down.

The first question is which kinds of account qualify. Not every arrangement an institution offers is treated as a deposit for the scheme's own purposes, and the scheme's definition is the one that counts, not the label on the product. The second is which kinds of institution are inside the scheme at all. Being licensed to take money from the public and being inside a protection scheme are two different conditions, and neither implies the other. The third is the unit the cover is counted per. Because the unit decides more than the amount does, it gets a section of its own below. And the fourth is the amount itself.

The answers move and the questions do not, so four questions with an address beside each one make a better working document than any set of answers. Every one of the four is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and every one of them is answerable in a few minutes at that address. Two of the four questions have answers that depend on which institution it is, so the important refinement is to answer them for the specific institution actually dealt with rather than in general.

What sits outside it, and why are the surprises there?

Everything genuinely surprising lives on the outside list, and the outside list is almost never printed beside the inside one, so it deserves a section of its own. Four shapes sit outside, and all four can be named without naming a single threshold.

The first is a balance beyond whatever the amount turns out to be. Everybody already knows this one exists, even if nobody knows where the line is. The second is a claim on an institution that is not inside the scheme. Being inside is a question about the institution rather than about the depositor or the money.

The third and fourth are the interesting ones. The third is a financial claimAn amount one party can require from another. What a claim looks like from each of its two sides is settled separately. that is not a deposit at all, even though it was bought at the same counter, from the same person, in the same fifteen minutes. The fourth is an arrangement that looks like a deposit in every respect a customer can actually observe, and is not one in the way the scheme defines it.

Neither the third nor the fourth is visible from anything a person can see, and nobody who did not distinguish them has failed at anything. There is no cue. The counter is the same counter, the paperwork looks like paperwork, and the difference lives in a definition held in one place and displayed in none. Being careful would not have helped, so none of that is a failure of care.

A route helps instead. Which claims and which institutions are inside the scheme and which are not is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and those definitions move. The question worth asking is about the specific arrangement at hand rather than about the category it appears to belong to.

What is the cover counted per, and why is that the whole answer?

The counting unit is the question people get wrong most often, and it is the one that costs the most to get wrong. The cover is counted per something, and the whole of the answer depends on what that something is. The sentence in bold does not give the answer. The sentence gives the shape of the question, and the shape is the part that never changes.

There are at least four candidates, and they are not close together in their consequences. Counted per account, several balances at one institution each carry their own cover. Counted per depositor at an institution, all the balances that person holds there are added together first and one cover applies to the total. Counted per depositor across the whole scheme, balances at different institutions are added together too. Counted per institution, the question is about the institution rather than about any depositor at all.

One identical set of balances put through those four rules gives four different answers. Not four slightly different answers. Four answers that can differ by a multiple, from exactly the same money sitting in exactly the same places. The counting unit is not a detail of the result, it is the result.

The unit itself, and the rule for how several balances held in different ways are added up inside it, are set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and both of them move. So the row carries that name inside it and no value.

The point to leave with is short. The arithmetic is the easy half and the counting rule is the hard one, so the counting rule has to be read at the source before any arithmetic is done. Anybody can divide a balance by an amount. Nobody can work out from a statement what gets divided.

Start the control at one account and every rule agrees: one cover unit inside the scheme, or 16.67 per cent of a holding drawn at six cover units. With a single balance the unit genuinely does not matter, and that agreement is why the question stays invisible for most people. Move to four accounts and the rules separate. Counted per account the answer is 4.00 cover units and 66.67 per cent of the holding. The other three sit exactly where they were, at 1.00 cover unit and 16.67 per cent. Three of the four rules do not react to splitting at all, one of them reacts completely, and a statement shows neither.

What share of one bank's deposits sits inside the scheme?

Now the calculation everybody actually wants, worked honestly. Working it honestly means naming both of its inputs and leaving both of them blank, and two named blanks teach more than a made up figure would.

Suvarna Commercial Bank Limited, an invented bank, is used here only because it has a balance sheet that stands still. The bank holds deposits of Rs 1,92,000 crore against total assets of Rs 2,40,000 crore, so deposits are 80.0 per cent of assets. Of those deposits, current and savings accountsBalances that can be moved or withdrawn without waiting, as against money placed for a fixed period. What they cost an institution and why the mix matters to it are settled separately. are 42.0 per cent of deposits, or Rs 80,640 crore. Both percentages are short enough to check with a pen: Rs 1,92,000 crore divided by Rs 2,40,000 crore gives the 80.0 per cent, and 42.0 per cent of Rs 1,92,000 crore gives the Rs 80,640 crore. Every one of those figures is arithmetic on the bank's own reported numbers, and none of them says anything about the scheme.

So: what share of that deposit book sits inside the scheme? Two inputs are needed, neither of them is available, and the two are unavailable for completely different reasons. The difference between those two reasons is the actual teaching.

The first input is the amount and the counting unit. The amount and the unit are set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in and they move, so no value for either is printed. Both can be had from that address in about ten minutes. The first blank is a routing decision, and it can be closed today.

The second input is the spread of balances across the people who hold them: how many depositors there are, and how the Rs 1,92,000 crore is distributed among them. The spread of balances is not a published figure. A depositor count, a distribution of balance sizes and a split of the deposit book by holder are all things an institution knows and a reader outside it does not. The second blank is therefore not a routing decision and it cannot be closed in ten minutes. The institution itself has to disclose it.

Knowing which of those two blanks is which is the thing to take away. Rs 1,92,000 crore on its own answers nothing, and the reason it answers nothing is that a total gives the size of a book and never the shape of it. Rs 1,92,000 crore held by a very large number of small holders and Rs 1,92,000 crore held by a small number of very large holders are the same total and completely different answers to this question. Anybody quoting a covered share without saying where the distribution came from has either been given it or has made it up.

What does cover not do, even where it applies?

Three things, and all three are worth stating bluntly. The gap between what people think cover does and what it does is where the disappointment lives.

Cover does not keep the institution alive. Cover sits underneath the depositors; what happens to the institution itself is a separate process with its own decisions and its own order, and resolutionThe process for dealing with an institution that has reached the point of being wound down or restructured. What has to be decided, and in what order, is worked through separately. is worked through separately rather than here. A covered depositor and a surviving institution are two different outcomes and neither one implies the other.

Cover does not make a balance available at the same instant it stops being available. A settlement takes whatever time a settlement takes, and in the interval the money is not in the depositor's hands. The settlement time is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in and the row for it is left empty, but the shape of the point survives without the number: cover is about whether the balance arrives, not about when.

And the third one matters more than either for everything that comes after. Cover does not reach the people who funded the institution without being depositors. An institution is not funded only by deposits. At Suvarna Commercial Bank Limited deposits of Rs 1,92,000 crore are 80.0 per cent of total assets of Rs 2,40,000 crore, so about a fifth of what stands behind those assets is something other than deposits. Market borrowingMoney an institution raises from other institutions and investors rather than by taking deposits. It is the whole difference between a lender that takes deposits and one that does not, and it is settled separately. is the shape it usually takes at a lender that raises money without taking deposits, and Rukmini Finance Limited, also invented, funds itself that way entirely.

Why does that third point belong here rather than later? Because of what an uncovered claim is. A claim that is not covered is a claim whose holder still has every reason to act, and that claim is the connection along which trouble carries on. Cover changes who is connected to the institution. Cover does not change whether anybody is, and that is the sentence that carries forward.

Does the premium move with the risk the institution takes?

One question, and it is the hinge into everything that comes after. If every institution inside the scheme pays the same rate for the cover, then an institution taking more risk is being covered on exactly the same terms as one taking less, and the difference in what that cover is worth to each of them is being paid for by somebody. If instead the rate moves with the risk, the cost of taking more of it lands where it arose.

Notice the shape of what has just been said. Neither half is a claim about what any scheme anywhere does. Both halves state what follows from each of two possible arrangements, and both are true whichever arrangement is actually running.

Whether the premium varies with risk is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and the row here is empty like all the others, so which arrangement is in force anywhere is a question for that source. Why the question is worth asking at all is what matters, and the answer carries straight into what protection does to the behaviour of the people it protects, covered next.

How does a household actually use any of this?

Why is every figure describing the scheme left to its source?

Not one figure describing the scheme has appeared so far. Leaving every one of them out was a decision rather than an oversight, and the reason is worth setting out.

The amount, the counting unit, the accounts that qualify, the institutions inside the scheme, the premium and the settlement time are all set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, and every one of them moves.

A guide that printed any of them would not be out of date on the morning it changed, it would be wrong, and a reader relying on it would be relying on something false rather than something old. The distinction between old and wrong is the whole argument. Stale information announces itself eventually. A confidently printed figure that has quietly changed does not announce anything at all, and it is trusted precisely because it looks like an answer.

So the questions appear instead, with the address beside each one. A reader who spends ten minutes at that address has better information than any printed figure could have given them, and knows how to refresh it. Knowing how to refresh it is the part a figure could never have supplied. A route is not a consolation prize for a missing number. A route is the better outcome, and it is available today.

Who sets the values left blank here?

Seven rows, two authorities, and not one value filled in. Taken to the addresses inside it, the table becomes a working document; left as it is, it stays right, and staying right is more than a filled version of it could promise.