Gross Debt vs Net Debt: Which One the Bridge Deducts

Gross debt is everything borrowed. Net debt deducts the cash. Sankalp Industrial Systems Limited, an invented manufacturer, owes Rs 6,00,00,00,000 and holds Rs 1,20,00,00,000. Net debt is therefore Rs 4,80,00,00,000, or 1.67 times earnings before interest, tax, depreciation and amortisation (EBITDA) against 2.08 times on gross debt. The walk from enterprise value to equity value deducts gross debt and adds the cash on its own line. Deducting net debt as well counts the same cash twice.

A household makes the awkward part of this distinction land faster than a balance sheet does. Suppose a household has a home loan of Rs 30,00,000 outstanding and Rs 4,00,000 sitting in a savings account. Asked what it owes, it has one answer: Rs 30,00,000. The bank's statement says Rs 30,00,000, the loan agreement says Rs 30,00,000, and Rs 30,00,000 is the balance the interest is worked out on every single month.

Now ask a slightly different question. Suppose the household emptied the savings account into the loan tomorrow morning: the balance left would be Rs 26,00,000. The emptied-account question is a completely reasonable thing to want to know, and it is a completely different question from the first. Nobody has emptied anything, so Rs 26,00,000 is not a number anybody can demand. Gross and net are not two attempts at the same figure; they are the answers to two different questions, and almost every mistake in this area comes from using one of them to answer the other one.

What exactly is gross debt, and what is it made of?

Gross debtThe total amount borrowed, across every tranche, before any deduction. is the total amount borrowed, added up across every borrowing the company has, before anything at all is taken off it. Gross debt is not an estimate and not a view. The total is a stack of contracts, each with an amount, a rate, a payment date and a set of conditions, and the stack can be read off the documents.

Sankalp Industrial Systems Limited is a listed manufacturer of industrial valves, precision castings and the aftermarket parts and service that go with them. At Year 0 it has borrowed in three separate places. There is a secured rupee term loan of Rs 3,00,00,00,000. There are listed unsecured non-convertible debentures of Rs 2,00,00,00,000. There is a working capital facility of Rs 1,00,00,00,000 drawn against receivables and inventory. Three hundred plus two hundred plus one hundred crore is Rs 6,00,00,00,000, and that is gross debt. The rate and the maturity each of those three carries are settled earlier in this subject.

The important property of that Rs 6,00,00,00,000 is that somebody can ask for it. Each of the three has a counterparty who can point at a document. The term loan has a repayment date written into it. The debentures are held by people who will expect their money back on a stated day. The working capital facility comes up for renewal every year, and a facility that is not renewed becomes repayable. None of that is affected in the slightest by whether the company happens to be holding cash on the day it is read.

What is net debt, and why can nobody demand it?

Net debtGross debt less cash and cash equivalents. is gross debt less cash and cash equivalentsMoney in the bank and near-money holdings, as disclosed in the accounts.. For this invented company, Rs 6,00,00,00,000 less Rs 1,20,00,00,000 leaves Rs 4,80,00,00,000. One subtraction, and the definition is finished.

The subtraction is worth saying out loud, and it is the part readers skate over. The subtraction imagines a transaction. Suppose the company took its whole bank balance and used it to pay down borrowings this afternoon: Rs 4,80,00,00,000 would be left owing. Knowing that figure is useful, and the usefulness is the reason the measure exists at all. A company sitting on a large balance is in a different position from an identical company with an empty account, and gross debt on its own does not show that difference.

But no lender has a claim on net debt, no document mentions it, and no payment is ever computed from it. Net debt is a figure somebody chose to work out. Choosing to work a figure out is not a criticism; a great deal of useful analysis is figures somebody chose to work out. The warning is about where the measure stops being safe to use, and the places where it stops are what follows.

Notice what did not change between the top bar and the bottom bar of that picture. The three tranches are still there. The rates are still what they were. The repayment dates have not shifted by a day, and nobody at any of the three lenders has been told anything. The only thing that happened is that somebody performed a subtraction on a sheet of paper.

Is all of the cash actually available to deduct?

No, and this is the first place the neat definition starts to fray. A business needs a working balance to exist. Wages go out on a date, suppliers get paid on their terms, and money arrives from customers when it arrives; the account has to be able to absorb the gap between those without the company having to ring anybody. Take the food stall outside an office building. The stall has to buy tomorrow's vegetables at five in the morning, before a single customer has paid it anything. Its float cannot run down to nothing at closing time.

The record for this invented company splits the balance for exactly that reason. Of the Rs 1,20,00,00,000 held at Year 0, Rs 40,00,00,000 is operating cashThe part of the cash balance the business needs to run day to day. It cannot be paid away without disturbing operations. the business needs to run, and Rs 80,00,00,000 is excess cashThe part of the cash balance that could be released without disturbing the running of the business.. How that line gets drawn in practice, and what a surplus balance does to a measured return, are covered separately in this subject, and the split is taken here as given.

Once the split is accepted, the tidy definition of net debt has a choice inside it that most readers never notice. If net debt is meant to be what would be left after using the cash to repay, then strictly only the cash that could actually be released is available. On that stricter reading net debt is Rs 6,00,00,00,000 less Rs 80,00,00,000, or Rs 5,20,00,00,000. Against EBITDA of Rs 2,88,00,00,000 that is 1.81 times rather than 1.67 times.

Be careful with that Rs 5,20,00,00,000. The stricter figure is not a correction of anything. The figure of record for this invented company is net debt of Rs 4,80,00,00,000 on the full cash balance, and that is the number every other treatment of it uses. The Rs 5,20,00,00,000 is a second convention, set out here so that a reader who meets it somewhere else knows what they are looking at rather than assuming somebody has made an arithmetic error. Two houses can compute net debt two ways, both defensibly, and land 0.14 turns of EBITDA apart on the same company on the same day.

Which cash convention is used here, and why must it be stated?

The whole Rs 1,20,00,00,000 is deducted and added above, and the convention is worth being explicit about. A sharp reader will suspect a double count inside it. The suspicion goes like this: if Rs 40,00,00,000 of that cash is genuinely operating, is it not already inside the operating business somewhere, and does adding all of it not count that part twice?

Here the answer is structural rather than a matter of taste. Net working capital in this invented forecast is receivables of Rs 2,16,00,00,000 plus inventory of Rs 1,44,00,00,000 less payables of Rs 1,80,00,00,000, and it contains no cash at all. The operating Rs 40,00,00,000 is nowhere inside the model that produced the operating value, so nothing whatever is counted twice by adding the whole balance. Every figure above uses the gross cash convention. The alternative moves the answer, so a write-up has to say which convention it used.

A house that treats the operating Rs 40,00,00,000 as part of the business and adds only the Rs 80,00,00,000 of excess lands Rs 40,00,00,000 lower. On 20,00,00,000 shares that is exactly Rs 2.00 a share. Neither house has made a mistake. A house that does not say which convention it used has made one. A reader cannot reconcile two answers when one of the two differences is invisible.

What do the two leverage ratios look like side by side?

A leverage ratioBorrowing measured against a flow, most often EBITDA. The size of the borrowing then reads as years of that flow. measures borrowing against a flow, and EBITDA is the usual flow. For this invented company at Year 0, EBITDA is Rs 2,88,00,00,000. Gross debt to EBITDA is Rs 6,00,00,00,000 over Rs 2,88,00,00,000, or 2.08 times. Net debt to EBITDA is Rs 4,80,00,00,000 over the same Rs 2,88,00,00,000, or 1.67 times. Both are computed from the same two disclosures on the same day.

The gap between the two is not a coincidence and it is not variable: it is the cash balance divided by EBITDA, and here that is Rs 1,20,00,00,000 over Rs 2,88,00,00,000, being 0.42 turns. Borrow another hundred crore and both measures rise by the same amount, so the gap is still 0.42 turns. Repay a hundred crore out of new equity and both fall together. Only the cash balance moves the gap. The distinction is really a fact about the cash rather than a fact about the borrowing.

The 0.42 turns is worth carrying around as a number rather than as an idea. If leverage of 1.67 times is quoted for this company where something over two was expected, the thing to ask about is now precise, and the answer to that question is a bank balance rather than a difference of opinion.

Which of the two figures is the interest actually charged on?

Gross, always, and this settles one argument outright rather than leaving it to preference. Interest for the year on this invented company's borrowings is Rs 48,00,00,000, being the blended 8.00 per cent that its three contracts work out to, charged on Rs 6,00,00,00,000. Interest is not charged on Rs 4,80,00,00,000. If it were, the bill would be Rs 38,40,00,000, and no lender in the arrangement has agreed to anything of the sort.

Think about the household again. The bank does not send a smaller demand because there is money in the savings account. The bank sends the demand for the loan it made. The savings account earns whatever it earns, separately, and the two facts sit side by side without touching. A lender charges on what it lent. Every ratio built out of the interest bill is therefore a gross measure, whether or not anybody says so.

Interest cover for this invented company is earnings before interest and tax (EBIT) of Rs 2,40,00,00,000 over interest of Rs 48,00,00,000, or exactly 5.00 times. One property of that ratio catches people out: it does not change at all when the view switches from a gross one to a net one. Both of its inputs are flows for the year, one an operating flow and one a contracted payment, and neither of them is a balance sitting on the balance sheet at a moment in time. The gross against net distinction simply has no purchase on it.

The immunity of interest cover marks where the distinction bites and where it does not, and is worth holding on to. The distinction bites on any measure built from a stock: debt to EBITDA, debt to capital, debt to assets. A measure built from two flows is untouched. A reader who knows which of the two is in front of them has already avoided about half the confusion available in this area.

Where does the distinction reach a valuation at all?

At one line of the walk from enterprise valueThe value of the operating business, before deciding who has a claim on it. to equity valueWhat is left for shareholders once every other claim on the business has been settled., often called the bridgeThe line by line walk from the value of the business to the value of its equity., and that one line is what the rest of this guide is about. The walk itself is covered separately in this subject, so the setting is worth stating in two sentences and no more.

A discounted cash flow values the operating business. For this invented company the answer is Rs 21,28,13,79,094, being five years of forecast cash flow and a terminal value discounted at a weighted average cost of capital of exactly 12.00 per cent, with year-end discounting. How that 12.00 per cent was assembled and how the model was built are settled elsewhere and are restated here rather than rebuilt. The one thing that matters here is what that figure is a value of: the operating business, and nothing else.

Which is exactly why a walk is needed at all. The cash sitting in the bank produced none of the operating cash flow that was forecast, so it is not inside the Rs 21,28,13,79,094 and has to be added. A surplus land parcel and a holding in an associate company produce none of it either, so they are added too, at Rs 1,00,00,00,000 between them. The borrowings are somebody else's claim on what the operating business is worth, so they come out. A minority interestThe share of a consolidated subsidiary that belongs to somebody outside the group. of Rs 60,00,00,000 is the quarter of a consolidated subsidiary that belongs to somebody else, and it comes out as well. Why each of those lines exists is settled separately, and only the arithmetic and the sign on two of them matter here.

Why does the walk add cash on one line and deduct gross debt on another?

Because those two lines are doing two separate jobs, and each of them has to be done exactly once. Adding the cash is the answer to a question about assets: there is money in the bank, it produced none of the forecast operating flow, so it belongs to the holders of the business and has to be brought in. Deducting the borrowings is the answer to a question about claims: Rs 6,00,00,00,000 is owed to three sets of lenders, and whatever the operating business plus the cash plus the surplus assets are worth, those lenders come first.

Written out with the signs: Rs 21,28,13,79,094 plus Rs 1,20,00,00,000 plus Rs 1,00,00,00,000 less Rs 6,00,00,00,000 less Rs 60,00,00,000, giving equity value of Rs 16,88,13,79,094. On 20,00,00,000 shares that is a value per shareEquity value divided by the number of shares in issue. of Rs 84.41.

The rule that keeps this honest is simple enough to write on one line: the cash appears on exactly one line of the walk. There are two correct ways to build it and they give the same answer. Either add the cash and deduct gross debt, as the walk above does, or deduct net debt and add no cash at all: Rs 21,28,13,79,094 plus Rs 1,00,00,00,000 less Rs 4,80,00,00,000 less Rs 60,00,00,000 lands on the same Rs 16,88,13,79,094. Both are right. A mixture of the two must never happen.

What happens if net debt is deducted and the cash line is left in?

The cash is counted twice, and the answer comes out too high by exactly the cash balance. The same walk with one line changed: Rs 21,28,13,79,094 plus Rs 1,20,00,00,000 plus Rs 1,00,00,00,000 less Rs 4,80,00,00,000 less Rs 60,00,00,000 is Rs 18,08,13,79,094. Against the correct Rs 16,88,13,79,094 that is Rs 1,20,00,00,000 too high, and Rs 1,20,00,00,000 is the cash balance to the rupee.

The reason it lands on exactly the cash balance, rather than on something approximately like it, is visible in the algebra. Net debt is defined as gross debt less cash, so deducting net debt is deducting gross debt and then adding the cash back. So a walk that deducts net debt has already added the cash once, inside that line, well before the line that says add cash is reached. Adding it again on its own line adds it a second time. The error is always precisely the cash balance, always in the same direction, and it therefore grows with the balance rather than staying a fixed nuisance.

The scaling is the point worth carrying. A company holding Rs 3,00,00,00,000 rather than Rs 1,20,00,00,000 would be overstated by Rs 3,00,00,00,000 by the same slip. The error is not a rounding, it is not a matter of judgement, and it does not partly cancel against anything. The overstatement is one number, counted twice.

Why does almost nobody catch this particular error?

Because of where the wrong answer lands. The traded price of this invented company is Rs 90.00 a share. The double countIncluding the same amount twice in one calculation, usually where two separately correct steps overlap. produces Rs 90.41, a mere 41 paise away from it. The correctly built figure is Rs 84.41, Rs 5.59 away, or 6.21 per cent below.

Sit with what that does to a person building the model. An analyst who makes the slip opens a file that says the business is worth almost exactly what it trades at. Nothing needs explaining. Nobody asks a question. The output has the shape of a model that has been done properly, and the closeness feels like a result rather than a coincidence. The analyst who does it correctly opens a file six per cent below the market and has to go and account for the difference in front of somebody.

The error validates itself and the correct answer does not. The arrangement is precisely backwards from what a reviewer's instincts expect. The sense check that normally catches arithmetic slips is the comparison against something known, and here the comparison against something known is the thing hiding the slip. The agreement gets treated as the check, so an error that moves a model towards the market is the hardest kind there is to find.

One rule is worth taking away from this even if nothing else is. Agreement between a model and a market price is not evidence that the model was built correctly. A wrong model can agree and a right one can disagree, and the invented case above demonstrates both at once. The only check that catches this is mechanical: the walk written out with a sign and a reason against every line, and written confirmation that the cash appears on exactly one of them.

One honesty note about the 41 paise. The wrong answer landing so close to the traded price is a coincidence of this invented case, checkable by anybody from the figures above. The closeness is not a general property of the error. On another company the double count might land far from the price, or well below it, and nothing about that would make it less of an error. The shape of the trap generalises; the distance does not.

What does the error look like at every other cash balance?

The error looks like a gap that opens in exact step with the bank account, and watching the gap open is the quickest way to stop treating the double count as an abstraction. Hold everything about this invented company still except the cash: gross debt stays at Rs 6,00,00,00,000, the enterprise value stays at Rs 21,28,13,79,094, the non-operating assets stay at Rs 1,00,00,00,000 and the minority interest at Rs 60,00,00,000. Then move the balance.

Two things are worth watching for as the control moves. The first is that the shaded space between the two lines is always exactly the cash balance expressed per share, so at a nil balance the two readings are identical and there is no error to make. The second is that the wrong line crosses the traded price at a cash balance of Rs 1,15,93,10,453, a little below the balance this company actually holds. Nothing about that crossing means anything; it is simply where two straight lines meet, and it is a reminder that the appearance of agreement is generated by arithmetic rather than by insight.

Which measure does a company quote about itself?

The friendlier one, almost always, and it is worth being unbothered about that. Net debt to EBITDA of 1.67 times is a better-looking figure than gross debt to EBITDA of 2.08 times, so 1.67 times is the one that tends to appear on a presentation slide. The accounts, meanwhile, disclose borrowings and cash as two separate lines and usually leave the ratio to the reader. The gross figure is the one a reader computes for themselves.

Nobody has done anything wrong in that picture, and treating it as evidence of anything is a mistake in the other direction. Both figures are correct. Choosing the flattering one for a slide is what every organisation everywhere does with every number it has a choice about. The obligation that does exist is narrow and it is easy to meet: say which measure is being shown.

The reason to care is comparison. Reading 1.67 times for one company and 2.20 times for another and concluding that the second is the more heavily borrowed may be nothing more than reading a net figure against a gross one. Two companies compared on two different measures have not been compared at all. One habit fixes the problem: ask every single time which measure is in front of the reader, and quote both in any write-up.

What travels beyond this company, and what does not?

The distinction and the rule travel. Gross debt is a set of contracts and answers what is owed. Net debt is a subtraction and answers what would be left if the cash were used. Anything built on a stock of debt moves when the measure switches between them, by the cash divided by whatever the denominator is, and anything built on two flows does not move at all. In a walk from the value of a business to the value of its equity, the cash appears on exactly one line, and mixing the two correct methods counts it twice.

None of the numbers travels, and the coincidence that made the failure so vivid least of all. The Rs 84.41, the Rs 90.41, the Rs 90.00 and the 41 paise between the last two belong to one invented company on one invented day. The wrong answer landing nearer the traded price than the right one is a property of these figures and not a law about this error. The refusal carries away, not the arithmetic: agreement with a market price is never, on its own, evidence that a walk was built correctly.

And the last thing that does not travel is any conclusion about the company. The correctly built Rs 84.41 sitting below the traded Rs 90.00 is not evidence that this invented company is cheap, dear, undervalued or overvalued. The Rs 5.59 is a difference between a model output and a price, and a difference between two numbers is not a verdict about either of them. The whole point of the comparison here is the arithmetic error, not the gap.