The Private Credit Stack: Senior to Subordinated

Private credit is a loan made by a fund rather than by a bank, held to maturity rather than traded. Its stack runs from senior secured debt at the top, through unitranche and mezzanine, down to subordinated and unsecured claims at the bottom. Two facts fix any position in that order: what it ranks ahead of, and what security it can actually claim.

Everything that follows rests on one idea that a reader arriving from equity does not yet have, and it is worth sitting with before any arithmetic turns up. A shareholder holds a residual: whatever is left when everybody else has been paid. A lender holds a claim of a fixed size that is paid in a fixed order, and the order is written down long before anything goes wrong. So the interesting question about a loan is never how much it pays, it is where it stands on the day there is not enough to go round. A rate is only ever the price attached to a place in that order. Read on its own it says nothing about the place, and this guide ends with the exact arithmetic of a reader who tried.

What is private credit, and how is it different from a bond anybody can buy?

Something small enough to hold in mind makes the point. A cousin asks a relative for Rs 2,00,000 to fit out a tea stall and says she will pay it back over two years with something on top. There are three ways to say yes. The money can be lent against nothing at all, on her word. The money can be lent against the equipment. If the stall closes, the ovens are the lender's to sell before anybody else touches them. Or it can be lent on the written understanding that the bank which already financed her scooter is paid first. Same cousin, same stall, same Rs 2,00,000, three completely different positions. Nothing about her changed between them. The document changed, and nothing else did.

Private creditA loan made by a fund rather than by a bank, held rather than traded. is that choice made at scale by a pooled vehicle that lends other people's money and writes the paper itself. The lender is not a bank and the loan is not a listed instrument. Nobody quotes a price for it in the morning paper. There is no screen, no last traded figure and no stranger standing ready to take it off the fund's hands at four in the afternoon. The fund negotiates the amount, the rate, the security and the promises directly with the borrower, signs, and then sits with the position until it matures or something happens to it.

The whole difference from a bond a reader could buy lies in that one fact. A listed bond is an instrument somebody else drafted, sold to a crowd, and left to be traded between strangers. A private loan is a document two parties wrote for each other. A private loan is held rather than traded, and its terms are changed by the lenders who signed it rather than by whoever happens to hold it that week.

The fund worked throughout is Nilgiri Direct Lending Fund I, invented, managed by Nilgiri Alternatives Advisors Private Limited, with Nilgiri Trusteeship Services Private Limited as its trustee. The fund is registered as a Category II Alternative Investment Fund with the Securities and Exchange Board of India. Investors committed Rs 3,00,00,00,000 to it. At its record date it has lent Rs 2,40,00,00,000 across eight positions, and every one of those eight sits inside a different borrower's capital structure. Every rate quoted here is this one invented fund's own contracted rate, and not one of them is a statement about what private credit costs in India or anywhere else. The record does not name the eight borrowers.

Where does a credit fund's cash come from, and why does it arrive every quarter?

How Private Credit Funds Work

A fund that buys companies has one cash event per holding, and it lands on a day nobody can name in advance: somebody eventually buys the company, and until they do, nothing comes back. A fund that lends is built the other way round. The document says how much and it says when. Interest falls due on dates written into the agreement, principal falls due at maturity, and both of those are contracted amounts on contracted dates rather than a price somebody first has to be found for.

The circuit is short. Investors commit, the manager calls that money in pieces as it finds loans to make, the money goes out as facilities, coupon comes back on the timetable in each document, principal comes back at maturity or on an early repayment, and the fund passes its receipts on to its investors. How the fund then splits those receipts between its investors and its manager, what its fee is charged on, and what its preferred return and its carried interest are, all belong to a separate treatment and are used here without being reworked. This vehicle's fee is charged on drawn capital rather than on commitments, its preferred return is 7.0 per cent and its carried interest 15.0 per cent, and all three are this invented fund's own contracted terms.

The one structural consequence worth carrying forward is that a credit fund's cash comes back in many small dated pieces rather than in one undated lump, and that changes the shape of what can go wrong with it. An equity holding disappoints when a buyer will not pay what the manager hoped. A loan disappoints on a specific date when a specific payment does not arrive, or earlier than that, when a promise inside the document is broken while the payments are still arriving perfectly well. A missed payment and a broken promise are two different events with two different consequences, and keeping them apart is most of the work of everything that follows.

What actually fixes a lender's place in the order of claims?

Two facts, and nothing else. The first is rankThe place a claim holds in the order in which claims are paid.: the claims this one is paid before, and the claims it must wait behind. Rank is set by contract. A lender can agree in writing to be paid after another lender, and once it has, no amount of arguing on the day changes the order. The second is security: what identified property this claim can actually take hold of if the borrower fails. A chargeA right over identified property that lets a lender claim it if the borrower fails. is the name for that right. A first charge over the machines means that if the machines are sold, the holder of that charge is paid out of the sale proceeds before anybody else sees a rupee of them. A second charge over the same machines means exactly what it sounds like: its holder takes whatever is left of that specific property after the first charge has been paid in full, and if nothing is left, it takes nothing.

Rank and security together place any lender in any document. Rank gives the queue. Security gives whether the queue is for a specific pile of property or for whatever general pool remains. A rate is not a third fact, it is a price the borrower agreed to pay for the particular combination of rank and security the lender ended up with. The eight positions below are therefore drawn with the rate sitting beside the layer rather than in a column of its own.

One thing must be said plainly before the picture. The eight positions are loans to eight different borrowers. Rank only ever exists inside one borrower's capital structure. Position 1 and position 5 are claims on entirely different businesses, so neither one ranks ahead of the other. The drawing shows which layer each of the eight occupies inside its own borrower's structure. Reading it as a queue among the eight themselves is the first mistake available on this subject, and it is worth refusing at the start.

What sits at the top of the order, and what does a first charge actually buy?

Position 1 is a senior secured term loan of Rs 50,00,00,000 carrying a cash coupon of 13.5 per cent, this invented fund's own contracted rate. Position 1 holds a first charge over the borrower's fixed assets and a first charge over its receivables. Two separate charges over two separate pools of property, both first in line. If that borrower were ever sold up, the proceeds of the machines and the proceeds of the money its customers owe it would both be applied to this claim before any other creditor of that business saw a rupee from either pool.

The document also carries a leverage test. At drawdown the borrower's net debt was 3.0 times its earnings before interest, tax, depreciation and amortisation, and the agreement sets a maintenance level of 3.75 times, tested every quarter. How that test is sized, what a borrower can do to move it, and what happens to the arithmetic when earnings are adjusted, are each taken properly further on. The leverage test runs on a fixed timetable rather than at the lender's convenience, so it produces a dated moment four times a year at which somebody has to look.

Now the first of three places in this guide where a denominator has to be named out loud. Position 1 is 20.8 per cent of the Rs 2,40,00,00,000 this fund has actually lent, and 16.7 per cent of the Rs 3,00,00,00,000 its investors committed, and both sentences describe the same Rs 50,00,00,000. Neither figure is wrong. The reader will supply the other denominator, so a sentence that quotes one of them without saying which it used has stated something false.

What are the layers below the top, and how do they actually differ?

Three of the eight positions sit below a plain senior secured loan, and they are the three instruments a reader will meet by name most often. Each of them differs from the others on exactly three things: what it ranks behind, what property it can take hold of, and how the coupon is paid. Everything else about them is decoration.

Unitranche

A unitrancheOne facility priced at a single blended rate covering what would otherwise be a senior and a subordinated piece. facility is one instrument doing the work of two. Rather than a borrower signing a senior loan with one lender and a junior loan with another, and then everybody arguing about the interaction between the two documents, the borrower signs a single facility at a single blended rate. Position 2 is exactly that: Rs 45,00,00,000 at a blended 15.0 per cent, with a first charge over all assets. From the borrower's side there is one lender to talk to, one rate, one set of promises and one signature.

The senior and junior pieces have not disappeared, they have moved. Behind the single facility sits an agreement between the fund and a bank that divides recoveries between them if the borrower fails, and the borrower is generally not a party to how that division works. The instrument the borrower signed is simple. The arrangement behind it is not, and how that split is actually written is taken further on in its own treatment.

Mezzanine Debt

Mezzanine debtDebt ranking behind senior debt, usually with a second charge and part of its coupon paid in more debt. sits below the senior layer and takes a second charge. Position 3 is Rs 30,00,00,000 with a second charge over the borrower's assets, paying 11.0 per cent in cash plus 5.0 per cent paid not in money but in more debt, so the amount owed grows rather than cash leaving the business. Both figures are this invented fund's own contracted rates.

The second charge is the part readers get wrong, so it is worth being blunt. A second charge is not half a first charge and it is not a weaker version of one. A second charge is a claim on whatever is left of a specific property after the first charge over that same property has been paid in full. If the property realises less than the first charge is owed, the second charge reaches nothing at all. Not a reduced amount. Nothing. How interest paid in more debt compounds, and what it does to the cash a lender actually receives against the income it books, are worked properly further on.

Subordinated Debt

Subordinated debtDebt that has agreed by contract to be paid after another debt. is the plainest of the three and the hardest to feel. Subordinated debt is debt that has agreed, in writing, to be paid after another debt. Position 5 is Rs 20,00,00,000 at 16.0 per cent, unsecured, ranking behind a bank. Read that sentence again slowly. Two separate disadvantages sit inside it, and they are often confused into one. Unsecured means there is no charge over any identified property, so this claim is queueing for whatever general pool is left rather than for a particular pile. Ranking behind a bank means that even in that general pool, it waits until the bank has been paid in full.

The 16.0 per cent on position 5 is not a bonus for being clever, it is the price the borrower agreed to pay for a lender willing to stand last with nothing to take hold of. A rate as the price of a place in the order is the spine of the whole subject, and the arithmetic that proves it arrives two sections further on.

What changes when the same obligation is listed and traded instead?

Private Credit against Public Credit

Imagine the same borrower doing two things on the same morning. The borrower issues a listed bond on one set of terms and signs a private loan on identical terms. Same amount, same rate, same maturity, same promises. By the afternoon the two holders are in genuinely different situations, and the difference has nothing to do with the terms.

Public credit is priced by a market. Somebody quotes the bond, somebody trades it, and a price exists whether or not it is a sensible one. A private loan has no such price, so it is marked on a timetable using a stated method rather than discovered by a transaction. Public credit can be sold to anybody who will buy it, so the holder has a way out that does not involve the borrower at all. A private loan is generally held to maturity, and the honest description of the holder's exit is that there usually is not one. And public credit is changed by a process designed for a crowd of holders who have never met. A private loan is changed by the small set of lenders who signed it, sitting in a room, on terms they negotiate directly.

The terms are identical and the position is not, and the whole of the difference is who sets the price, who can leave, and who can change the document. A discount exists at all because of those three differences, and position 8 in this book is exactly such a purchase.

Position 8 in this book is that discount made concrete. Position 8 is an obligation with a face valueThe amount an obligation says it will repay, as distinct from what somebody paid for it. of Rs 30,00,00,000, bought in the secondary market for Rs 15,00,00,000, being 50 paise in the rupee. Two denominators again, and this is the second of the three places in this guide where they have to be named. The Rs 15,00,00,000 price is 50 paise in the rupee of the Rs 30,00,00,000 face value, and it is 6.25 per cent of the Rs 2,40,00,00,000 this fund has lent, exactly. Fifteen over 240 lands on a half, so the figure is stated unrounded rather than pushed either way. Moyer, Distressed Debt Analysis, 2005, is where the frame used here comes from: a position bought below its face value is read by where it stands in the order of claims rather than by the discount on the price tag. Exactly what a purchase below face value buys, and what it does not, is taken properly further on.

What does a covenant do, and what does it not do?

How Covenants Affect Private Credit Risk

A covenantA promise in a loan document, tested on a fixed timetable, that gives the lender a right if it is broken. is a promise inside the loan document, tested on a fixed timetable, that hands the lender a right if it is broken. The definition is that one sentence, and every word of it is doing work. A covenant is a promise, so it lives in the document rather than in anybody's good intentions. The promise is tested on a timetable, so a breach is discovered on a date rather than whenever somebody happens to look. And what a covenant hands over is a right. A right is not the same object as a payment.

Position 3 is where the worked case shows it happening. At the Year 2 Q3 test the borrower came in at 4.4 times against a 3.75 times maintenance level. The borrower had not missed a payment. Nothing had failed to arrive, no cash was short, and a reader watching only the bank statement would have seen nothing at all. The test produced a dated moment at which the lender and the borrower had to sit down, and the document was reopened and reset on new terms. The position is performing at the record date. How that reset was actually constructed, term by term, is covered separately.

A covenant does not make anybody pay; it makes somebody talk, on a date fixed in advance and while there is still something left to talk about. A covenant does exactly that much. Equally worth stating is what a covenant does not do: it does not repay the loan, it does not crystallise a loss, it does not create security where none was taken, and it does not move the lender up the order of claims by even one place. A lender who breaks a covenant is exactly as senior on the day after as on the day before.

Position 5 is the other event entirely. Time on this fund runs from its own drawdown, so Year 3 Q1 is the first quarter of its third year rather than a calendar date. In Year 3 Q1 the borrower missed a coupon and did not cure it. No test was tripped, no conversation was scheduled, and nothing about the document changed. A contracted payment did not arrive, and from that moment the position stopped being a stream of dated receipts and became a place in a queue.

What happens when a borrower defaults, and what decides the recovery?

How Default and Recovery Affect Private Credit Outcomes

A default turns every abstract sentence so far into a number. The enterprise behind position 5 was sold in a distressed sale for Rs 68,00,00,000. The Rs 68,00,00,000 belongs to the invented case and is not what any enterprise realises in a distressed sale anywhere. Costs of the sale are ignored throughout so the arithmetic stays visible.

Three claims stood against that Rs 68,00,00,000, and they were paid in a written order. Claim 1 was a bank's senior secured term loan of Rs 60,00,00,000 with a first charge over the fixed assets and the receivables. Claim 1 took Rs 60,00,00,000, being 100 paise in the rupee of what it was owed. Claim 2 was this fund's subordinated unsecured Rs 20,00,00,000. Claim 2 took the residual Rs 8,00,00,000, being 40 paise in the rupee. Claim 3 was the borrower's own ordinary shares. The money stopped at claim 2, so the shares took nothing at all. Check the arithmetic in the other direction: 60 plus 8 is 68, and nothing is unaccounted for.

Notice what did not appear anywhere in that paragraph. Nobody asked whether the business was a good one, whether its accounts were well kept, or why it failed. The recoveryWhat a claim actually receives, expressed in paise in the rupee of what it was owed. was decided by the size of the claim above it and the amount realised, and by nothing else. Moyer, Distressed Debt Analysis, 2005, is the frame in use here: a position in trouble is read by where it stands in the order of claims rather than by any judgement about the business, and that discipline is what makes the next table possible.

So now run the counterfactual. Nothing about it happened, so it must be labelled a counterfactual every time it is used. Counterfactual 1, the rank counterfactual: suppose the same Rs 20,00,00,000 had sat in the first-charge class alongside the bank's Rs 60,00,00,000 instead of behind it. The class would then claim Rs 80,00,00,000 against Rs 68,00,00,000 realised, and the members of one class share what there is in proportion to what they are owed. The bank takes 60 over 80 of Rs 68,00,00,000, being Rs 51,00,00,000. The fund takes 20 over 80 of Rs 68,00,00,000, being Rs 17,00,00,000. Check: 51 plus 17 is 68.

Rs 17,00,00,000 against Rs 8,00,00,000 is a difference of Rs 9,00,00,000 on the same borrower, the same enterprise and the same day, and not one fact about that business moved between the two columns. The borrower was equally troubled, the sale raised exactly the same amount, and the calendar did not turn. The only difference was a line in a document signed years earlier about where this lender stood and what it could take hold of. Rank and security decided the whole of it, and that is why a rate can never be read on its own. Counterfactual 1 is arithmetic on an alternative that did not occur, not a statement about what any position recovers.

What does one default do to a book of eight loans?

How Private Credit Funds Manage Loan Concentration

Think about a household running entirely on one salary. Nothing about that salary is wrong. The trouble is arithmetic: if it stops, one hundred per cent of the money stops, and no amount of care about which employer it comes from changes that. Put two salaries in the same house and a stoppage costs half. Concentration is not a judgement about any borrower, it is a statement about how much of the whole sits behind any one of them.

The fund has lent Rs 2,40,00,00,000 across eight positions. As shares of that cost, they run 20.8, 18.8, 12.5, 14.6, 8.3, 10.4, 8.3 and 6.25 per cent, and on the unrounded figures they sum to exactly 100.0. Position 8 lands on a half at 15 over 240, being 6.25 per cent exactly, so it is stated unrounded rather than pushed to 6.2 or 6.3 by a convention nobody agreed. The four largest, being positions 1, 2, 4 and 3, are Rs 1,60,00,00,000 between them, or 66.7 per cent of cost. Two thirds of this book sits behind four borrowers.

Position 7 is where the third and last denominator in this guide has to be named. Position 7 is a participation of Rs 20,00,00,000 in a facility of Rs 4,00,00,00,000 shared across six lenders. The participation is 5.0 per cent of the Rs 4,00,00,00,000 facility, and that same 5.0 per cent is the fund's share of the vote inside it. The same Rs 20,00,00,000 is 8.3 per cent of the Rs 2,40,00,00,000 the fund has lent. Two true sentences, two different denominators, and each one answers a different question: the first is about how much say the fund has in that borrower's document, the second about how much of its own book is exposed to that borrower. Confusing the two is the commonest error on this subject and it is not a rounding error, it is a change of subject.

Then the arithmetic everybody actually wants. One position of the eight defaulted. Position 5 was 8.3 per cent of the Rs 2,40,00,00,000 lent. Position 5 cost Rs 20,00,00,000 and recovered Rs 8,00,00,000, so Rs 12,00,00,000 did not come back, and Rs 12,00,00,000 is 5.0 per cent of the Rs 2,40,00,00,000. Eight positions, one default, and what it did to the whole book is a single line of division. Having eight rather than one means exactly that line of division, and it is the only reason the number of positions is worth stating at all.

Who does this arithmetic actually land on?

Three people meet it in an ordinary week, and it means something different to each of them. The first is whoever marks this fund's book at the end of a quarter. There is no screen price for any of the eight, so the mark on a position that has stopped paying is not an opinion about the business, it is an estimate of what stands above the claim and what the property behind it would fetch. When position 5 was written to Rs 8,00,00,000, the working was the Rs 60,00,00,000 sitting ahead of it. Everything the mark rests on is in the documents rather than in a view.

The second is whoever runs finance at a borrower. A covenant test is a date in their calendar, not a threat, and a maintenance level is the number they have to be able to explain four times a year while nothing has yet gone wrong. The maintenance level is also why the borrower under position 3 was in a room in Year 2 Q3 with every payment made on time. The document produced the meeting.

The third is somebody reading a quarterly report from a credit fund and trying to work out what the numbers mean. Two questions do all the work: what does each position rank behind, and what property can it actually take hold of. A report that gives eight rates and no ranks has handed over the price of eight places without saying what any of the places are. Rank and security say where a claim stands in the order of payment. Whether standing there was worth 16.0 per cent is a judgement those two facts cannot make for anybody.