Demerger, Spin-Off and Carve-Out: What Separates Them

A demerger, a spin-off and a carve-out all take part of a business away from the rest, and they differ on two things only: who holds the separated part afterwards, and whether any cash arrives. A carve-out brings in an outside buyer, so it is the only one where a rupee enters. For Meghdoot Coated Products Limited, that single difference decides whether a separation touches the Rs 900 crore at all.

Underneath all three words sits one exercise, and it is an unglamorous one. Somebody sits with a list of everything the business holds and everything it owes, and writes A or B beside each line. Writing A or B beside each line is the whole of a separation. The three words differ only in what happens after the list is finished: who is left holding the B column, and whether anybody paid for it. The business the line gets drawn through below is Meghdoot Coated Products Limited, an invented manufacturer carrying Rs 900 crore of borrowings against Rs 60 crore of annual operating earnings.

What does separating a business actually involve?

Every asset, every liability, every contract, every employee, every permission and every rupee of borrowing has to end up on one side of the line or the other. There is no third column. A machine cannot be half on each side; a supply agreement cannot be partly with one company and partly with another; a loan cannot be owed by a company that no longer exists in the form the lender wrote its agreement against. So the work of a separation is the work of allocating, line by line, until the list is exhausted.

Nothing in that exercise creates a rupee, an order, a customer or a unit of output. The plant makes the same product on the day after the line is drawn as it did on the day before. The customers place the same orders. The employees do the same work. The lenders are owed exactly what they were owed. Drawing a line across a business is an act of description, and description does not change the thing described.

Almost every mistaken argument about a separation is a version of forgetting that description changes nothing, so the thought is worth holding firmly. The line looks decisive on a slide. The line is a strong black rule down the middle of an organisation chart, and the eye reads a strong black rule as an event. The line is not an event. The line is a labelling exercise whose consequences arrive later, through completely different doors: through what somebody is then willing to pay for one of the two labels, or through what somebody is then willing to lend against one of them.

Three of the rows are harder than the rest, and they are worth naming now because they come back later. Contracts do not always move because somebody has decided they should: many of them need the party on the other side to agree. Moving a contract is then a novationMoving a contract from one party to another with the agreement of the party on the far side of it. Without that agreement, the contract stays where it is. question rather than an allocation question. Permissions do not always move either: a licenceA permission from an authority to carry on a particular activity, issued to a named holder. It does not always travel just because the machinery it covers does. is issued to a named holder and may have to be applied for afresh by whichever company ends up running the activity. And borrowings almost never move quietly. A loan agreement typically carries a change of controlA term written into a loan or a contract that gives the other side rights the moment the party it deals with passes into different hands. term, and a group structure that has been redrawn is exactly the circumstance those terms were written for.

What is a demerger, and who holds the separated business afterwards?

In a demerger the company itself divides. One legal entity becomes two, and the people who held shares in the original hold shares in both of the resulting companies, in some stated proportion. Nobody has bought anything. Nobody has sold anything. A shareholder who held one certificate in the morning holds two certificates in the evening, and between them those two certificates describe the same underlying business the single certificate described.

Two consequences follow from that, and both matter more than they look: no outside party paid anything, and the register of holders on both sides starts out identical. The first is why a demerger cannot, by itself, put a rupee anywhere near a lender. The second is why a demerger is often the easiest of the three to get agreed among shareholders. Everybody receives the same proportion of the separated side that they held of the whole, so nobody is being diluted relative to anybody else.

A demerger is genuinely for separation of fate. Two businesses that were tied together in one entity stop being tied. Each side can be looked at, financed, judged and eventually sold on its own record rather than on the record of the combination. Separated fate is a real change and it is worth having. Separated fate is not a change to any number on either side of the line the moment the line is drawn.

What is a spin-off, and how is it different in practice?

In a spin-off, the business being separated is already sitting in its own entity, held by the parent. Nothing has to be divided. A distribution happens instead: the parent hands the shares of that entity out to its own shareholders, who then hold the parent and the separated company side by side rather than one inside the other.

No outside party paid anything here either, and the practical difference from a demerger is whether the company has to be divided at all or only a shareholding distributed. That difference is mostly a question of how the group was assembled in the first place. If the business to be separated was bought as a company and kept as a company, the entity already exists and only has to be handed out. If it grew up inside the main trading company, sharing its plant, its staff and its bank account, the second entity has to be built before anything can be handed out, and that building is the allocation exercise in the figure above.

A lender, though, sees the two as the same, and that is worth saying plainly. In both cases the shares of the separated business move from one set of hands to another set of hands that overlaps entirely. No cash came in. If the separated business had borrowings, it still has them; if the parent had borrowings, it still has them. Shares have been redistributed among people who already held them. A lender cannot be paid in that.

What is a carve-out, and why is it the only one where a rupee enters?

A carve-out separates a business and then sells a stake in it to somebody outside the group. Selling a stake to an outsider is the whole of the difference. An outside buyer commits considerationWhat a buyer actually hands over. Cash, shares, a promise to pay later, or some combination of the three, and the word covers all of them. and, where that consideration is cash, cash arrives that was not in the group the day before.

A carve-out is the only one of the three where a rupee enters from outside, and everything this subject cares about follows from that one fact. A demerger rearranges holders. A spin-off rearranges holders. A carve-out brings in a party who was not a holder, and who pays to become one. Whether that money then reaches the people who are owed money is a separate question with its own answer, but at least there is money to ask the question about.

Notice that the carve-out has a second effect the other two do not: it introduces somebody with an independent view of what the separated business is worth. Nothing is bought in a demerger or a spin-off, so neither of them ever prices anything. A carve-out produces a price, struck by a party spending real money, on a defined stake. For a business whose lenders are arguing about how much of the borrowing the earnings can support, an outside price is evidence that did not exist before, and evidence is scarce in these conversations.

Divestiture vs Carve-Out: what does the seller still hold the next morning?

A divestiture sells a business outright. The seller hands over the whole of it and keeps nothing: no shareholding, no board seat, no exposure to how it does next year. A carve-out separates a business and sells part of it, so the seller walks away from the table still holding a stake in the thing it has just sold part of.

For a business that owes money, the practical difference is this: a divestiture converts the whole of that unit into cash once, and a carve-out converts part of it into cash and leaves the rest as an asset whose value nobody has yet tested. The two outcomes differ in what is left behind, not only in how much cash arrives. The divestiture is finished. Whatever the buyer paid is what the unit was worth to the seller, and there is nothing further to argue about. The carve-out is not finished, and the part that is not finished is being carried on the seller side of the line at a number that has never met a bid.

There is a fair argument for each. A whole unit handed over at once, to one buyer, under time pressure, tends to fetch a price that reflects the pressure. A partial exit keeps the seller exposed to a recovery it believes in, and lets a later disposal of the rest happen at a price set by whatever the business does in the meantime. Against that, holding a residual stake means the money problem is only partly solved and the party just sold to is now sitting across the table on every decision.

The honest way to describe the difference is not that one is better, but that they end in different places. After a divestiture, the seller has cash and a closed file. After a carve-out, the seller has less cash and an open position. Which of those two a business at Rs 900 crore of borrowings should prefer is a judgement about its own circumstances.

How to analyse a Demerger Structure, and in what order?

There are six things to establish, and the order is not arbitrary. Each one narrows the next. Worked out of order, the earlier ones have to be done again.

The sixth is the one that decides whether the structure is workable, and it is the one most often settled last. That ordering is not accidental, and it is not laziness either. Step six genuinely depends on step five: how much borrowing a side should carry cannot sensibly be stated until what that side earns on its own is known. The trouble is that a structure can be designed, drawn, socialised and half agreed on steps one to five, and then step six arrives and says the whole shape does not work. Everything before it was about the business. Step six is about the lenders, and the lenders were not in the room for steps one to five.

Step five deserves a warning of its own, and the quiet error lives there. Giving each side its share of the group earnings figure and moving on is tempting. Sharing out a group figure is not what standing alone means. A group carries certain costs once and uses them twice: one finance team, one set of systems, one head office, one insurance programme. Costs like those are shared costsA cost paid once and used by two businesses at the same time. When the two part company, one of them starts paying it alone, or both start paying for their own version of it., and after a separation somebody has to pay for each of them alone. So standalone earningsWhat a business earns once it stops leaning on anything the rest of the group was paying for. Usually a smaller figure than its share of the group total. for the two sides added together are normally lower than the group figure they came out of, and treating them as a clean division of one number quietly assumes a cost saving nobody has found.

Where do the borrowings go, and what does that do to each side?

Here the description of structures gives way to arithmetic. Meghdoot Coated Products Limited carries Rs 900 crore of borrowings against Rs 60 crore of operating earnings. The reading is 15.0 times, and 15.0 times does not become a smaller number because somebody has drawn a line through the business. The Rs 900 crore is still owed, and the Rs 60 crore is still what the whole of it earns.

Working the allocation takes a split of the earnings. The record for Meghdoot Coated Products Limited carries one earnings figure and no division of the business at all. So the split below is an assumption made for this illustration and nothing else: a larger side taking three quarters of the earnings, Rs 45 crore, and a smaller side taking one quarter, Rs 15 crore. There is no plant, no product line and no division behind those figures. The two figures exist so the arithmetic can be seen.

Take the third row slowly. The third row is the finding. Rs 675 crore spreads across Rs 45 crore of earnings and lands on 15.0 times. Rs 225 crore spreads across Rs 15 crore of earnings and lands on 15.0 times. Two companies, two balance sheets, two sets of lenders, two boards, and the identical reading the single business had before anybody drew anything.

The separation moved nothing, and that outcome is forced arithmetic rather than a coincidence: a fixed total divided in exactly the same proportion as a fixed total cannot change the quotient. The smaller side took 25 per cent of the earnings and 25 per cent of the borrowings. Twenty five per cent of a numerator over twenty five per cent of a denominator is the original fraction. The answer was predictable without touching a calculator, and predictability is the point: the allocation never had the power to move the ratio.

How does the carve-out look on the same figures? The mechanism is the same, but no number can be put on it, and the reason is worth stating. If an outside buyer paid for a stake in the smaller side, cash would arrive from outside the group, and only then would anything reach the Rs 900 crore at all. But the record for Meghdoot Coated Products Limited carries no bid, no offer and no valuation of any part of the business, so any figure written here would be a figure invented twice over. The shape of the consequence can still be stated: whatever arrived would then face exactly the same ranking question as every other rupee available to a business that cannot pay everybody, and that ranking question is settled separately.

What happens at the two extremes of the same allocation?

The proportional case is the tidy one. Now push the allocation to its limits. The limits are what make the finding unarguable.

Put the whole Rs 900 crore on the larger side and that side stands at 20.0 times: Rs 900 crore over Rs 45 crore is 20.0. The smaller side stands at nil, carrying no borrowing at all. The smaller side looks like the winner until the larger one is looked at. The larger side has gone from 15.0 times to 20.0 times, worse than the position the whole business was in before anybody separated anything.

Now push the other way. Put the whole Rs 900 crore on the smaller side and that side stands at 60.0 times: Rs 900 crore over Rs 15 crore is 60.0. The larger side is at nil. One side is now completely clean and the other is carrying four times the ratio the combined business carried.

When the line was drawn the Rs 900 crore stayed exactly as large as it was, and so did the Rs 60 crore underneath it, so no allocation exists in which both sides come out below where the whole started. Every rupee taken off one side lands on the other. Every turn one side improves is a turn the other side gives up, and because the smaller side has a third of the earnings of the larger one, a turn moved off the larger side costs three turns on the smaller. The trade is the whole mechanism, and it does not have an exception hidden in it anywhere.

Two shops under one roof

Strip the vocabulary away and the mechanism is a familiar one. Two shops trade under one roof, opened years ago on one loan taken for both, and the loan is now heavy against what the two of them take between them. The stronger shop, on the street side, does most of the trade. The weaker one at the back does less.

So a wall goes up. Two doors, two shutters, two counters, two sets of books. On the day the wall is finished, ask the only question that matters: how much less is owed to the bank? The answer is not a rupee less. The bank is owed exactly what it was owed the day before, and the two shops between them take exactly what they took the day before. The wall changed who runs which counter. The wall did not change the loan.

Then somebody says the front shop should carry less of the loan. The back shop, after all, was always the weaker one. Fine, but say the second half of the sentence out loud: the back shop now carries more of a loan on takings that were smaller to begin with. The wall did not create a place to put debt where nobody has to service it. Only a buyer for one of the shops changes what is owed, and a buyer is a different transaction from a wall.

So what is a separation genuinely good for?

Arithmetic like that reads as an argument that separations are pointless, and that is the wrong conclusion drawn from the right numbers. A separation is genuinely good for something other than the ratio.

A separation delivers one thing: each side can be financed, governed, run, judged and eventually sold on its own merits, rather than inside a group whose position it does not share. A perfectly healthy business sitting inside a stretched group is treated as part of a stretched group by everybody who looks at it: by lenders who see the consolidated position, by customers who read about the group, by good people deciding whether to join it. Separated, it can go and refinance on its own strength and be assessed for what it is. The gain is real, and it has nothing to do with the leverage arithmetic.

There is a second gain and it is the one that eventually matters most for a business that cannot pay. A separated business is sellable in a way an entangled one is not. A buyer can be shown a defined perimeter, a defined set of contracts and a defined set of people, and can price them. Nobody can price a business that is inseparably tangled with another one. So much of the work in this area is really the work of making something saleable.

A separation can make a sale possible, and it is the sale that brings the cash, not the separation. The two events get compressed into one story afterwards, and the story gives the credit to the structure. The structure did the enabling. The buyer did the paying.

How does a lender, an analyst or a household actually use this?

A lender being asked to consent to a separation reads it in one direction only: which side is my borrowing going to sit on when this is done, and what does that side earn standing alone? Nothing else on the proposal changes the lender position. A lender to the group whose loan is being allocated to the smaller of the two sides has just had the earnings under its loan cut, whatever the covering note says about strategic clarity. A separation involving borrowings is also rarely a decision the company takes by itself: the loan documents usually give the lenders a say, and a cross guaranteeA promise by one part of a group to answer for another part of it borrowing. It is why a separation can run into a lender who never lent to the side being separated. given by one part of the group for another part is exactly the kind of thread that has to be unpicked before the two sides can genuinely stand apart.

An analyst reads a separation proposal by looking straight past the strategic language for two numbers per side: what does it earn standing alone, and what borrowing is it being asked to carry. If the proposal gives the first without the second, the analyst has been handed half a structure. The single most useful habit here is to recompute both sides yourself and check whether they add back to the group position. If they do not, something in the proposal is being assumed rather than allocated. In this case they do add back: Rs 675 crore plus Rs 225 crore is the Rs 900 crore, and Rs 45 crore plus Rs 15 crore is the Rs 60 crore.

An investor holding shares in a business being demerged has a smaller job than they might think, at least on day one. The investor holds the same underlying business in two envelopes instead of one. Two things are worth watching: whether the standalone earnings on each side turn out to be what the proposal said, once shared costs stop being shared, and where the borrowings ended up. The two answers decide what each envelope is actually worth, and neither of them is settled on the day the separation is announced.

And a household reads the same mechanism whenever it thinks about splitting a joint position. Two brothers running one business on one loan taken jointly can divide the shop, the stock and the customer book between them cleanly enough. The loan is the row that resists. The loan was taken by both, it is owed by both, and dividing the business does not divide it. Somebody has to actually agree to release one of them, and the party who has to agree is the one who lent the money.