Gross Leverage and Net Leverage: One Subtraction Apart

Both readings divide a borrowing by a single year of earnings. The net one first takes the cash off the borrowing; the gross one does not. The subtraction is the entire difference, and it is a claim rather than a measurement: it says the cash is really there, really the borrower's, and really free to be handed over today. The claim is tested before the number is used.

Nobody argues about the arithmetic here. Eight hundred less three hundred is five hundred on any calculator ever built, and the division that follows is the same division on either route. The argument is about a sentence that never gets written down anywhere: that a rupee sitting in a bank account somewhere inside a business is worth exactly as much, to a lender, as a rupee of borrowing that was never taken. Sometimes that sentence is true. Often enough it is not, and the distance between the two readings then becomes the most quarrelled-over figure in a credit file.

Neither reading is the correct one and neither is the wrong one, and stating either would supply a verdict where a question is what the work needs. The work is narrow: put the two readings on one scale, write out what the subtraction asserts in plain sentences, test each assertion against the ways it fails, then reach for the reading that answers the question actually asked.

What separates the two readings, and how large is that separation here?

One declared set of inputs carries the work, and it belongs to no borrower at all. Earnings before interest, tax, depreciation and amortisation, measured across twelve months ending on one date, stand at Rs 200 crore. At that same date a borrowing of Rs 800 crore is outstanding, and Rs 300 crore of cash is being held against it.

The first reading takes the borrowing whole. Rs 800 crore set against Rs 200 crore of earnings gives 4.00 times a year's earnings. The second reading takes the cash off first. Rs 800 crore less Rs 300 crore is Rs 500 crore, and Rs 500 crore against the same Rs 200 crore of earnings gives 2.50 times those same earnings. Same borrowing, same earnings, same division, one subtraction, and the reading moves by 1.50 turns.

One row does all the work. Three of the four rows are identical down both columns. One line lifted off one balance sheet at one date produces the whole of the 1.50 turns. Before anything else is read into either figure, weigh the size of that second row. The row's size is the size of what somebody is asserting.

What is a reader claiming when the cash comes off the debt?

Here is the move that matters. Written out as a sentence, the subtraction stops looking like arithmetic. Taking Rs 300 crore off Rs 800 crore says, in one keystroke, four separate things about that Rs 300 crore: that the money exists in the place the accounts put it, that it is the borrower's own and not somebody else's sitting in the borrower's accounts, that nothing stops it being sent to whoever holds the borrowing, and that sending it would cost nothing and take no time.

Said aloud, those four are audibly not obviously true. The four are also not obviously false, and that is the difficulty. Each of the four is the kind of statement that is right most of the time in most businesses, so nobody develops the habit of checking, and the one occasion it matters is the occasion nobody checked.

The net reading is not a more careful version of the gross reading; it is the gross reading with a four part assertion bolted on. A more careful reading would be one that knew more. The net reading knows exactly the same three amounts and has simply decided what one of them means.

Try it on a household, where the sentences are easier to hear. A house has a home loan of Rs 30,00,000/- outstanding and Rs 5,00,000/- in a savings account. Netting them says the household is really only Rs 25,00,000/- in debt. Is it? The money is there, and it is theirs, so two of the four hold. But part of that Rs 5,00,000/- is the fund that carries a household through a hospital admission or six months without a salary. Handing it to the bank on a Tuesday morning would leave the loan smaller and the household one accident away from missing a payment on it. The arithmetic never had an opinion about that. The sentence underneath it did.

Which four things have to hold before the claim is safe?

Take the four one at a time. None of them is exotic. Each has a way of failing that is recognisable the moment it is described, and that is precisely the problem: they are recognisable in hindsight and invisible in a table.

Where the cash sits

A set of accounts prepared on a consolidationThe practice of adding up a parent business and the companies underneath it and presenting the total as though it were one entity. basis adds every bank balance in the group into one cash line. The borrowing, meanwhile, may have been taken by one specific company inside that group. Cash held two levels down is real cash, but moving it up to the company that signed the loan is somebody else's decision, taken by people with their own reasons, and possibly with a minority shareholder or a local lender to satisfy first. In the accounts, the cash and the debt sit in the same column. In life they sit in different rooms.

Whose money it is

Plenty of what shows up as cash is being looked after rather than held. A firm that collects payments for its clients has those receipts in its own bank account and owes every rupee of them onward. Money in an account pledged to a lender as security appears in the same line as the current account. A balance inside a joint arrangementA venture run together by two or more parties, where neither side can move the money in it without the other agreeing first. is half somebody else's, and neither half moves alone. Accounts often flag some of this as restricted cashMoney shown on the cash line that its holder is not free to spend, because it has been pledged, is being looked after for somebody else, or is locked by an agreement.. Where they do, a reader has been given a gift. Where they do not, the reader has been given a number.

Whether it is already committed

Cash that will pay next week's suppliers is spent money that has not yet left the building. So is the balance set aside for a tax instalment that falls due in three weeks. A balance sheet reports positions and not intentions, so neither is marked in any way at the date it is struck. Netting that money against the borrowing counts it twice: once as a reduction in the debt, and once as the payment owed to the people already waiting for it.

Whether it can actually move

The last condition surprises people. A cash balance can pass the other three and still fail this one. Cash sitting in a subsidiary in another country may be entirely the borrower's, entirely unspoken for, and entirely unavailable this quarter. Getting it home runs through an approval, a tax charge, or a queue. The same is true of a balance in an account that nobody may draw on without a signature from a party under no obligation to give one quickly. Money in that position is what people mean by trapped cashMoney its holder genuinely has, sitting somewhere it cannot be pulled out of on the day somebody wants it.. Ownership and reach are two different axes, and a cash balance can score full marks on one while scoring nothing on the other.

How much of a cash balance is machinery rather than surplus?

Now the condition most readers meet last and ought to meet first. Every business that trades holds a balance simply in order to keep trading. Wages leave on the first of the month, a big customer settles on the twentieth, rent goes out on the fifth, and the balance riding between those dates is not spare money at all. The balance is a working part, doing a job, in the same way a delivery van is doing a job when it is parked.

Stand outside a vegetable stall at six in the morning. The trader has a float in a tin: change for the first hour of customers, and enough to pay the wholesaler at the mandi before any of today's takings arrive. Ask that trader to hand over the float to settle a debt and the business closes at noon, not because it is unprofitable but because it has no change. The tin is a working balance, and every business has one. A cement plant's version has more zeroes and works exactly the same way.

The amount worth netting against a borrowing is only the amount left over after the working balance, and virtually nothing published anywhere tells a reader where the line between the two falls. A rule would be more welcome, and there is no rule. There is a cash figure, and inside it, somewhere, sits an amount the business cannot operate without, and the accounts do not draw the line.

Which means a reader who has not asked the question has already answered it. Netting the whole balance assumes the working balance is nil, the least likely of all the possible answers. A nil working balance is the assumption made by not making one, and that makes it the most common assumption in credit analysis.

Why is the removal a straight line rather than a switch?

Stop thinking of gross and net as two competing measures, and see them instead as the two ends of one line. Treating none of the cash as available gives the gross reading of 4.00 times a year's earnings. Treating all of it as available gives the net reading of 2.50 times the same earnings. Everything in between is a real position, and it is the position most real situations occupy.

The line is straight, and it is straight for a dull reason worth stating: the denominator never moves. Earnings stay at Rs 200 crore whatever is decided about the cash, so every Rs 20 crore treated as available removes exactly the same 0.10 turns. The first Rs 20 crore carries the reading from 4.00 times down to 3.90 times, and the last Rs 20 crore moves it by the same 0.10 turns at the far end of the run. No point on this line gives an extra rupee of cash more relief than the rupee before it. The reader has to supply a position rather than look for a natural one.

What share of the reading does the subtraction take, and does earnings size change it?

Here is a small piece of arithmetic that settles an argument people have without noticing. The gap between the two readings is 1.50 turns, and 1.50 turns out of a gross reading of 4.00 times is 37.50 per cent of the reading gone. Now compute something that looks unrelated: Rs 300 crore of cash against Rs 800 crore of borrowing is 37.50 per cent of the borrowing.

The same number twice, and it is forced rather than lucky. The gap is the cash divided by the earnings; the gross reading is the borrowing divided by the same earnings; divide one by the other and the earnings cancel out completely, leaving nothing but the cash over the borrowing. The proportion of a leverage reading that the subtraction removes is fixed by the cash and the borrowing alone, and no earnings figure anywhere can change it.

The absolute gap, unlike the proportion, does depend on earnings. Measured in turns, the distance between the two readings is the cash over the earnings, so a weaker year stretches it. Measured as a proportion of the reading, it never moves at all. Two true statements about the same subtraction, and a reader who confuses them will argue for the wrong one.

What does each reading settle, and what does neither of them touch?

The gross reading states the size of what has to be serviced and eventually repaid, and it does so making no assumption whatsoever. Making no assumption is its whole character. The gross reading is not cautious, it is not conservative, it is simply arithmetic on two published amounts with nothing added.

The net reading states the size of what would be left if every rupee of cash were sent to whoever holds the borrowing tomorrow morning. The thought experiment is a genuinely useful one, and a lender who never performs it is missing something real: a business with plenty of cash is in a different position from one with none, and a measure that cannot see the difference is blunt. The net reading is not a description of where things stand today. The figure describes a world in which one specific action has already been taken.

And now the longer list, the part both readings share. Neither of them states when the borrowing falls due, and a Rs 500 crore obligation maturing in eleven months is a different animal from the same amount maturing in eleven years. Neither states what the borrowing costs to carry. Neither states whether this year's Rs 200 crore of earnings is a normal year, a peak or a recovery. Neither states whether the cash will still be sitting there at the next reporting date, and cash is the fastest moving thing on a balance sheet. Neither states where a particular lender stands against the others if payment stops.

A reader holding both figures has one obligation described twice, not two independent facts, and the feeling of being better informed for having two numbers is the feeling of having counted an assumption as evidence.

Which reading answers which question?

So the practical output is not a preference but a branch, and the question that was asked decides which arm to take rather than which figure looks more respectable.

If a lending contract names a definition, that is the one to use and nothing else. A test written into an agreement is measured on the wording of that agreement, and an analyst's view about which reading is more sensible has no standing at all in that calculation. How much room a borrower has before such a test is failed, and how a change of definition eats that room, is covered separately. The room a lender thinks it has and the headroomThe distance between where a measured figure stands today and the limit that has been agreed for it in a contract. it actually has differ by exactly this subtraction.

If the question is whether a borrowing can be repaid out of what exists right now, the net reading is the one that speaks to it, and the four conditions are tested before it is written down. The net reading was built for exactly that case, and used with the conditions attached it does an honest job.

If the question is how much has to be found every year regardless of anything else, the gross reading is the one to take. Here is the reason, and it catches people out: interest is charged on what is owed, not on what is owed less whatever the borrower happens to be holding. A borrower with Rs 300 crore in the bank pays the same annual interest as a borrower with nothing in the bank, on the same Rs 800 crore of borrowing. Cash sitting in an account does not reduce a rupee of the debt serviceThe payments a borrowing demands as it runs: the interest charged each period, plus whatever principal has to be repaid in it. bill. Every rupee of cash reduces the net reading, and not one rupee of it reduces what has to be paid across next year.

Which of the two survives a move from one borrower to another?

Here the argument for the gross reading stops being a matter of taste. Businesses hold cash for entirely unrelated reasons. One collects money from its customers weeks before it delivers anything, so a fat balance is simply other people's timing showing up in its accounts. Another is holding a reserve against a payment it already knows is coming. A third has had a good run and has not yet spent what it made.

Now set two of those side by side and compare their net readings. The comparison sets two different sentences about two different cash balances side by side, dressed up as one measurement. The arithmetic is identical in both cases, and that is exactly what makes the comparison feel legitimate. The sentence underneath the arithmetic is different in each, and that is what makes it worthless.

The gross reading travels between borrowers because it assumes nothing, and that single property is the entire case for it. It states no more than the net reading. The gross reading states less, and it states the same amount of less about everybody. A comparison needs exactly that. A measure that carries a different hidden assumption for each subject is not a measure; it is a set of separate opinions written in the same font.

Household version, same shape. Two neighbours each hold Rs 5,00,000/- in savings against the same home loan. One is saving for a wedding eight months away. The other simply has not got round to spending it. Their net positions read the same on paper, and only one of them could hand the money over. Nothing in the arithmetic can tell the two apart, and nothing in the reported figure was ever meant to.

What does it cost when nobody asks where the cash is?

How do a lender, an analyst and a household each put this to work?

A lender writing a credit note does two things with the two readings, and they are different jobs. The gross reading goes into the sizing question: how big is this obligation next to what the business earns, and what has to be found every year to carry it. The net reading goes into the repayment question, and it goes in with its conditions attached. A lender who writes the net figure alone has put an unverified claim into a file that other people will rely on for years. The best notes list the four conditions underneath and mark each one confirmed, refused or unknown. An unknown is not a failure. An unknown that was never written down is.

An analyst comparing several businesses has a narrower job and a firmer answer. Line them up on the gross reading, the only one of the two that means the same thing in every column. Then bring the cash back as its own column, in rupees, next to a column saying where it sits. Two columns and a sentence beat one clever combined figure every time. The reader can see what was assumed instead of having to reconstruct it.

An investor buying the borrowing rather than lending it has a third angle again. The five year government SPOT rate marks where the pricing of a five year borrowing begins, and how far beyond it a particular borrower has to reach is a question covered separately. One thing matters: if the price was struck against a net reading, the price was struck against an assumption, and the assumption is worth reading before the price is.

And the household is the plainest case of the three. Before savings are used to pay down a loan, the four questions are asked in ordinary words. Is the money actually there. Does it belong to the household, or is it a parent's money parked in the account. Is it already promised to the school fee in April. And could it be got at on the day it was needed back. If any of those has a shaky answer, the loan the household thinks it has is not the loan it actually has.

Where would this be looked up for Palash Cements Limited, and what is there?

Neither figure here belongs to Palash Cements Limited. The name appears across the corpus as the borrower behind an invented five year bond carrying a 9.10 per cent annual coupon on Rs 1,000.00/- of face amount, and that is the entire extent of what exists about it. There is no cash balance. There is no schedule of what it has borrowed. There is no earnings figure for any twelve month period. There is no note explaining where a rupee of anything would sit, and that note is exactly what the four conditions demand.

So the two readings worked through above are readings of one declared set of inputs that describes no company at all, and the comparison between them is a comparison of two ways of reading four amounts rather than a finding about any borrower. Whether Palash Cements Limited is carrying too much borrowing cannot be answered without a cash balance, a borrowing schedule and an earnings figure, and none of the three exists for it.

Palash Cements Limited and the five year bond written against that name are invented.
Educational material. Not advice on any investment, tax, budget or market position.