Direct Finance and Intermediation: Who Holds the Claim
Direct finance leaves the saver holding a claim issued by the borrower. One claim exists, the saver judged the borrower, and the saver carries the failure. Intermediation puts an institution between them, holding two separate claims: one it owes the saver, one the borrower owes it. The saver has no claim on the borrower at all, and the institution's own capital takes the loss first.
Money that is not being spent sits on one side. Spending that has not been paid for sits on the other. Economists call those two ends the surplus side and the deficit side, and getting the money across the gap between them is the only job anybody in this account is doing. There are exactly two ways to do it, and a reader who can tell them apart can read almost any arrangement they will ever meet.
Both routes have to settle the same four jobs, and neither route makes any of them disappear. Somebody has to size the promise, decide what it is worth, hold it until it comes good, and swallow the shortfall if it does not. Direct finance leaves all four jobs sitting with the saver; intermediation sells all four to an institution, and the price of that sale is the difference between the rate the borrower pays and the rate the saver receives. Every other difference between the two routes comes out of that sale, one criterion at a time.
What is direct finance, exactly?
A saver hands money to the party that is going to spend it, and receives a piece of paper from that same party saying what is owed and when. The exchange is the whole of it. The money goes one way, a promise comes back the other way, and the promise was written by the party now holding the money.
Consider a cousin opening a second shop. He needs eighty thousand rupees for stock. The saver has it, gives it to him, and writes on a sheet of paper that he owes eighty thousand rupees, repayable in a year, with something on top. Both sign it, and the saver puts the copy in a steel almirah. Nothing else happened. No third party looked at his shop, checked his supplier, formed a view on whether the second location gets any footfall, or stood behind the promise in any way. The saver did all of that alone, or did not do it at all, and either way the outcome is now the saver's.
On the direct route exactly one claim exists in the world, it names the borrower as the party who owes, and the saver is the party who holds it. If the cousin's second shop does well, the money comes back to the saver and the amount on top is entirely the saver's. If the second shop is empty for eleven months, the paper in the almirah is still worth exactly whatever he can pay, and there is no one else to ask.
What is financial intermediation, exactly?
With an institution in the middle, what it actually does is not what most people picture. The institution does not take a saver's money and pass it along to a borrower on that saver's behalf. The institution does two separate things that never touch each other outside its own books.
First, it takes money from savers on its own promise: money in, and the institution owes a stated amount back on stated terms. The promise is the institution's, signed by the institution, and it stands whatever happens anywhere else. Second, and entirely separately, it lends money on its own account, taking claims from borrowers made out to itself. The two sides meet nowhere except inside the institution's own balance sheet.
The consequence is the single fact most readers have never had put to them plainly: the saver holds a claim on the institution and no claim whatsoever on the borrower, so if the borrower fails, the saver's claim does not change by one rupee. Everything else follows from it.
The everyday version. A customer leaves two thousand rupees with the shopkeeper down the lane against goods to be collected over the month. He uses that float to pay the farmer who supplies him. If the farmer's crop fails, the customer has not lost two thousand rupees. The shopkeeper still owes the goods. Whatever the cash physically did, the customer's money was never lent to the farmer. The promise was always from the shopkeeper, and his problem with the farmer is his problem. The shopkeeper's float is the whole architecture, and a bank differs from the shopkeeper in scale, in regulation and in what it can be made to prove, but not in that shape.
How many claims does each route create, and who holds each?
One on the direct route, held by the saver against the borrower. Two on the intermediated route, one held by the saver against the institution and one held by the institution against the borrower. So far this is just counting. The part worth slowing down for is what those two claims look like next to each other.
The second claim is not a copy of the first with a different name on it: the two have different amounts, different lengths and different terms, and the entire business of an intermediary lives in the space between them. Savers want their money back soon, in small amounts, with certainty. Borrowers want money for years, in large amounts, with room to breathe. Nobody makes those two wishes agree by finding the rare pair who happen to match. An institution takes both wishes as they are, promises the saver what the saver wants, promises the borrower what the borrower wants, and carries the mismatch itself.
Ten thousand savers each leave a small amount they can take back quickly. One borrower takes a large amount for seven years. The institution has changed the size, the length and the certainty of what it holds on one side into something quite different on the other side, and the reason it is allowed to do that is that it put its own capitalThe owners' money inside an institution, the part that belongs to nobody who has to be repaid. It is what stands in front of every other claimant when something goes wrong. behind the difference.
A saver puts money into an institution. The institution lends it out. How many claims now exist, and what does the saver hold?
Who decides the borrower is worth lending to?
On the direct route, the saver does it. Or the saver pays somebody to do it. Or, far more often than anybody admits, the saver does not do it at all and calls the resulting feeling confidence. On the intermediated route the institution does it once, properly, and does it on behalf of every saver who ever hands it money, none of whom will ever see the file.
Here is the economics that decides everything downstream: finding out whether a borrower will actually repay costs roughly the same whether the amount lent is one rupee or one thousand crore. Somebody still has to look at the accounts, visit the premises, understand the trade, check who else has a claim, and form a view. The work of finding out does not shrink when the amount shrinks.
So the arithmetic is brutal and it is nobody's opinion. If investigating a borrower costs a fixed amount of effort, and the amount lent is small, the effort eats the entire return before the lending starts. If the amount is large, the same effort is trivial against what is at stake. The fixed cost of finding out sorts borrowers into two heaps, and everything about who can use which route follows from it.
A lender's borrowers fail on a small share of what they owe. Whose money falls first?
Who carries the loss when the borrower does not pay?
This is the one place where the two routes stop being different arrangements of the same thing and become genuinely different experiences.
On the direct route the arithmetic is one line long. Whatever is not recovered comes off what the saver holds, immediately, in full, and there is nobody in front of them. The cousin repays sixty thousand of the eighty thousand and the saver is down twenty thousand rupees. There is no queue. A queue needs at least two people in it, and here there is only the saver.
On the intermediated route there is a queue, and the saver is not at the front of it. The institution's owners are. Their capital is money that belongs to nobody who has to be repaid, and it is positioned to be used up first. Until it is gone, the saver's claim on the institution has not moved by one rupee, and the institution goes on owing exactly what it always owed.
Say the true thing and refuse the false one in the same breath: intermediation moves the loss into a queue with the institution's owners at the front of it, and it does not make the loss disappear. Somebody is still short by exactly the amount the borrower did not pay. Intermediation changes only the order in which people find out. The cover a depositor has if an institution does fail is set by the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in.
What does Suvarna Commercial Bank's own capital actually absorb?
The queue can be measured. Suvarna Commercial Bank Limited, an invented bank, puts its capital at Rs 24,000 crore and its advances at Rs 1,44,000 crore. Dividing the first by the second gives 16.67 per cent of advances. 16.67 per cent of advances is the loss rate on the loan book at which the owners' capital is entirely used up, and it is the number the whole comparison turns on.
Note what that percentage is struck on. Mistaking the base is the commonest way to get this wrong. The 16.67 per cent is struck on advances. Put that identical capital over the whole balance sheet of Rs 2,40,000 crore instead and it reads 10.0 per cent. The second ratio answers a different question with a different figure underneath it. Two ratios, one numerator, and only one of them is about the loan book.
Taken to an extreme, and it is an extreme: the bank reports gross advancesThe full amount lent, counted before anything is taken off for the part expected not to come back. The word gross is doing real work: a net figure is a smaller number answering a different question. that are non-performing of Rs 6,480 crore, a figure that comes to 4.50 per cent when it is struck on gross advances. Suppose every rupee of that were written offRecognising formally that an amount owed is not coming back, and taking it out of what the books say is owed. It is an admission on paper rather than a payment to anybody. with nothing recovered on any of it. Rs 6,480 crore set beside capital of Rs 24,000 crore is 27.0 per cent of that capital; set beside the whole Rs 2,40,000 crore balance sheet it is 2.70 per cent of assets. The depositor's claim does not move at all.
Set the identical book beside the direct route. A saver who had put that same Rs 1,44,000 crore out directly, recovering nothing on the same 4.50 per cent, is down 4.50 per cent of their money on the day it is written off. In full. With nobody in front of them. Same borrowers, same failures, same rupees, and two completely different mornings.
Two things about that arithmetic have to be said out loud.
The first is that writing the whole book off is a deliberate extreme and is not what the bank reports. The bank holds provisionsAmounts already set aside against lending that is not expected to come back in full. The money is recognised as gone in the accounts before anybody knows exactly how much of it really is. against 70.0 per cent of that non-performing figure, gross, so Rs 4,536 crore is already held and Rs 1,944 crore is left net. Against net advancesGross advances after the provisions held against them are taken off. Because the top and the bottom of the ratio both change, a net figure is never simply a smaller version of the gross one. of Rs 1,39,464 crore, that Rs 1,944 crore works out at 1.39 per cent. The bottom of the fraction changes along the way: the 4.50 per cent is struck on gross advances, the 1.39 per cent is struck on net advances, and anybody who parks the two over one denominator has just made one of them false.
The second is an absence. The record carries no issuance figure at all, so no rupee can be traced down the direct route into a borrower's hands. The record carries instead a year of trading at Kaveri Stock Exchange Limited, Rs 48,00,000 crore of turnoverAdd up what one holder paid another for paper that already existed, across a whole year. Not a rupee of that total is new money arriving anywhere. in the stated year, and that turnover is not money that reached a borrower. Paper moving between people who already own some raises nothing for anybody.
Suvarna Commercial Bank Limited has capital of Rs 24,000 crore and advances of Rs 1,44,000 crore. At what loss rate on the loan book is that capital used up?
What do the two exposures look like as shapes?
As the loss rate on the loan book climbs, what happens to each saver is not a big version and a small version of one picture. The two are different shapes, and the difference between the shapes is the whole argument.
The direct saver's holding falls in a straight line from the very first rupee not recovered. The intermediated saver's claim stays perfectly flat and then breaks. Flat, flat, flat, and then it begins to fall. One route gives a slope from zero. The other gives a floor and then an edge, and the whole skill of reading an institution is knowing roughly where that edge sits.
Move the loss rate on the loan book, and watch the same loss land in two completely different places.
One control moves: the share of Suvarna Commercial Bank Limited's Rs 1,44,000 crore of advances that is lost, anywhere from nothing to 20.0 per cent of advances. Everything else is held exactly where the record puts it. As the control moves, the loss bar grows across the queue, the two pools underneath give up whatever has reached them, and the bar at the foot shows what the very same loss does to a saver who had lent that identical book directly instead. The red marker never moves: it stands at 16.67 per cent of advances, where the owners' capital runs out. Setting the control below that marker, on it and past it shows which bars move in each case. The panel opens at 4.50 per cent of advances, the bank's own reported gross non-performing figure.
Every rupee of the Rs 6,480 crore this bank reports as non-performing, gross, is written off with nothing recovered. What happens to the depositor's claim?
What does the saver hold, and how do they get out early?
On the direct route the saver holds a claim on a named party. Getting out before it matures means finding somebody willing to take that claim off their hands, and taking whatever price that person offers on the day the saver needs to leave. If a secondary marketA place where people who already hold a claim sell it to other people. No new money reaches the borrower there; the paper simply changes hands, and how those places work is covered separately. exists for the thing held, that is easier. If it does not, the exit is a conversation with whoever can be found.
On the intermediated route the saver holds a claim on the institution, repayable on the terms it stated. Getting out means asking the institution, not hunting for a buyer. The exit feels effortless.
The asymmetry is worth stating plainly: one route needs the saver to find a buyer and the other needs the institution to have the cash on the day the saver asks, and those are two different things to worry about rather than one of them being worry-free. The institution promised to pay on demand while its money is out on advances for years. The institution manages that gap every single day. When it manages the gap well nobody notices, and liquidityHaving the cash on the day somebody asks for it. An institution can be entirely solvent on paper and still fail this test, because owing less than one owns says nothing about what is in the drawer this morning. catches people out for exactly that reason: the failure mode is invisible right up until the moment it is the only thing in the room.
Two savers both want their money back next week. One holds a claim on a borrower, the other holds a deposit. What does each of them actually need?
Investigating a borrower costs roughly the same whatever the amount lent. Which route does that arithmetic favour for a very small borrower?
Which borrowers can use each route at all?
Direct finance needs three things at once. A borrower somebody is willing to find out about. An amount large enough to justify the finding out. And information a stranger can actually check. A saver who cannot verify anything is not investing, they are hoping.
Intermediation reaches everybody below that line, and it does it by spreading the cost of judging across everything the institution lends. The same credit team, the same processes and the same accumulated knowledge of a trade get applied to ten thousand small borrowers, at a cost per borrower that no individual saver could ever match on their own.
Here is the consequence worth carrying away: the stall outside the office building is not shut out of direct finance by anybody's decision or anybody's judgement of it, it is shut out by the arithmetic of investigation cost, and the intermediary exists precisely to beat that arithmetic. No one sat in a room and excluded the stall. The stall needs forty thousand rupees for a bigger griddle, and the work of checking its takings costs more than the whole of what forty thousand rupees can ever pay anybody. An institution that has already built the machinery for judging small borrowers does that check for a fraction of the cost, and the stall gets its griddle.
Disclosure before a security may be offered to the public, and who may be offered one privately and in what minimum amount, are set by the Securities and Exchange Board of India (SEBI) at sebi.gov.in.
How is each route paid for?
On the direct route the saver keeps the whole of what the borrower pays, less whatever they paid anybody to arrange it. Simple, and it is the reason the direct route looks cheaper at a glance.
On the intermediated route the institution keeps the difference between what it charges the borrower and what it pays the saver. Work it on the record. Suvarna Commercial Bank Limited reports interest earned of Rs 18,600 crore across everything it lends and holds, and interest expended of Rs 11,160 crore across everything it owes. The difference is Rs 7,440 crore, and that difference is not profit. Operating expenses of Rs 5,040 crore come out of it, and provisions of Rs 1,800 crore come out of it, leaving Rs 600 crore. Other income of Rs 2,400 crore is what carries the reported profit before tax to Rs 3,000 crore.
One line on that Rs 7,440 crore. The figure is the commonest place a reader gets misled, and the discipline is easy. Set it against earning assets of Rs 2,04,000 crore, being advances of Rs 1,44,000 crore plus investments of Rs 60,000 crore, and it is 3.65 per cent. Set that identical Rs 7,440 crore beside the whole Rs 2,40,000 crore balance sheet and the very same income reads 3.10 per cent. Nothing about the bank changed between those two sentences. Only the denominator did. A margin quoted without its base is not a fact. How a bank earns that difference, and how a finance company that takes no deposits funds itself instead, are both worked in full separately.
The difference is not a fee bolted onto the same product: it is the whole of what the saver gave up in exchange for not doing the four jobs, and a reader who sees it as a charge has misread what was bought. The saver did not buy a worse version of a direct claim. The saver bought a claim on an institution instead of a claim on a borrower, with somebody else doing the sizing, the pricing, the carrying and the losing.
A saver moving from a direct claim to a deposit is paid less than the borrower is charged. What did they buy with the difference?
What do the two routes look like side by side?
Every criterion above changed the answer to a question a saver would actually ask. Here they are in one place, in the order they were built, so the whole comparison can be read at a glance rather than reassembled from memory.
Seven values, and the authorities that set them
Every requirement touched on here is set by an authority and every one of them moves. Each row below carries its authority, and the value in force is published at the site named in the row.
| Reserve Bank of India | The capital an intermediary must hold against what it lends, and the base it is struck on | rbi.org.in |
| Reserve Bank of India | How much may be lent to one borrower, or to one group of borrowers | rbi.org.in |
| Reserve Bank of India | When an advance stops being treated as performing, and what must be provided against it | rbi.org.in |
| SEBI | What an issuer must disclose before it may offer a security to the public | sebi.gov.in |
| SEBI | Who may be offered a security privately, and in what minimum amount | sebi.gov.in |
| SEBI | The margin a buyer must put up before dealing on an exchange | sebi.gov.in |
| Deposit Insurance and Credit Guarantee Corporation | The cover a depositor has if a bank fails, what it applies to and what it does not | dicgc.org.in |
How does anybody actually use this on a Tuesday?
Four people use this distinction constantly, and none of them would describe it in these words.
A household deciding where to put money it will need in eighteen months is choosing between an institution's promise and a borrower's promise, and the honest question is not which pays more. The question is which failure they are able to see coming and which one they would rather live with. Neither answer is wrong; picking without knowing they were picking is.
An analyst reading a lender does not read the lender's borrowers. There are ten thousand of them and no list. The analyst reads the capital against the book, the pace at which advances that stopped performing are appearing, and whether the funding is the kind that stays put when people get nervous. Every one of those questions is a question about the queue, and the queue is the reason they are the right questions.
A person buying a security directly is doing all four jobs themselves whether they realise it or not, and the cheapness of the route is precisely the bill for the work they have taken on. An institution's own treasury sits on both sides at once, and that arrangement is taken up below.
What goes wrong: believing the risk went away
A saver stops lending directly and puts the money into an institution. Somebody tells them, correctly, that the borrower's failure no longer touches them. And they draw the conclusion that nothing can touch them now. The step is a small one and almost everybody takes it.
The queue is real and the arithmetic above is exact. Suvarna Commercial Bank Limited's Rs 24,000 crore of capital stands in front of the depositor and takes the first losses, all the way to 16.67 per cent of advances. Failure did not go away. The failure that matters changed. The depositor is no longer exposed to any one borrower, and is now exposed to the institution: to whether its capital is enough for the losses it actually takes, and to whether it has the cash on the day people ask.
Two versions of the error are worth naming. The saver who researches a lender's borrowers when they should be reading the lender. And the saver who compares two institutions on what each of them pays and looks at nothing else at all. Comparing the price of two promises is not reading either promise.
The arrangement is built to make the borrower invisible, and it works, so nobody who made this error was careless: on the intermediated route, read the institution. The fix is one line long. And for deposit cover, go to the Deposit Insurance and Credit Guarantee Corporation at dicgc.org.in, the authority that sets it.
A saver holding a deposit wants to know how safe it is and starts researching the lender's borrowers. What should they be reading instead?
So does a saver pick one route or the other?
No, and a reader who thinks otherwise has been taught a false choice.
Savers use both routes at once, usually without ever having framed it that way. The same household holds a deposit and holds a security directly, on the same afternoon, out of the same salary. Neither holding knows the other exists. And the institution taking that household's deposit holds securities on its own balance sheet as well. An intermediary doing that is doing direct finance with somebody else's money in its hands.
Sharpen it with a second institution from the same record. Rukmini Finance Limited, also invented, lends but takes no deposits at all. Rukmini Finance funds itself in the market instead, with borrowings of Rs 14,400 crore against net worth of Rs 3,600 crore and assets under management of Rs 18,000 crore. So the people who lent it that Rs 14,400 crore are holding claims on Rukmini Finance directly, on the direct route. Rukmini Finance is at the same time the intermediary for every borrower on its own book. One institution, sitting on the deficit side of one transaction and in the middle of another, on the same morning.
The same division applies to it, and it takes four seconds. Rs 3,600 crore of capital against Rs 18,000 crore of assets under management is 20.0 per cent of assets under management, against the bank's 16.67 per cent of advances. Same queue, same order, different depth, and neither number is better than the other.
A route is a property of a transaction, not a camp anybody belongs to. The question is asked fresh of each thing held: who issued this promise, and who stands in front of the holder if it goes wrong. Two questions, and they are the two that carry.
A household holds a deposit and also holds a security directly. Which route is that household on?
Where each neighbouring thread is picked up. Whether a whole system is built mostly on institutions or mostly on markets is a different question, covered separately. A claim seen from each of its two ends, where one starts and what finishes it off, is covered separately. How a bank actually earns the difference between what it charges and what it pays, and how a finance company that takes no deposits funds itself instead, are worked in full separately. How a security is issued, priced, sold and settled is covered separately, and so is which securities are worth holding. Deposit cover, issuer disclosure, who may be offered a security privately, how much may be lent to one borrower and the margin a buyer puts up each sit with whichever authority is printed beside them in the rows above, and their names and sites stand there where a value would otherwise be.
Which authority fills the blanks left above?
Seven values in this comparison are set by authorities. The rows below name who sets each one and where it is published. Every rupee figure belongs to Suvarna Commercial Bank Limited, whose balance sheet balances, and each percentage is worked from two figures printed beside it, so any of them can be redone on paper.
| Authority | What it sets | Site | Confirmed |
|---|---|---|---|
| Reserve Bank of India | The capital an intermediary must hold against what it lends and the base it is struck on; how much may be lent to one borrower or one group; when an advance stops being treated as performing and what must be provided against it | rbi.org.in | 23 August 2026 |
| SEBI | What an issuer must disclose before it may offer a security to the public; who may be offered one privately and in what minimum amount; the margin a buyer puts up before dealing on an exchange | sebi.gov.in | 23 August 2026 |
| Deposit Insurance and Credit Guarantee Corporation | What a deposit is covered for, what that cover applies to and what it does not | dicgc.org.in | 23 August 2026 |
Suvarna Commercial Bank Limited, Rukmini Finance Limited and Kaveri Stock Exchange Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.