Carve-out

Large groups are not built out of tidy pieces. One division uses the same computer systems as three others. Its salespeople sell two divisions' products. Its factory shares a site. Its contracts are signed by the parent company on behalf of everybody.

So when a group decides to sell one division, it discovers that the thing it wants to sell does not exist yet. It has to be built first.

That building is the carve-out. Systems have to be split or copied. Contracts have to be divided. People have to be assigned to one side or the other. And somebody has to work out what this division actually costs to run on its own — because until now the parent has been paying for the head office, the insurance and the accountants, and nobody counted.

The sale itself is an ordinary auction. It is just that it is run on a business that does not yet exist in the shape being sold, which is why bidders discount for the uncertainty and why a well-prepared separation is worth real money.

Carve-out accounts

Accounts have to be prepared for an entity that never filed any. Revenue is usually straightforward. Costs are not, and three categories cause almost all the argument:

• Allocated central costs. The head office, the group finance team, the group insurance. Whatever share the division was charged internally is rarely what it would pay alone.
• Standalone costs. The things it does not pay for now and will have to: its own finance function, its own auditors, its own treasury.
• Dis-synergies. Buying power lost, contracts that were group-wide, discounts that only existed because of the parent's size.

A seller presents a low standalone cost and a buyer models a higher one. The gap, multiplied by whatever multiple is being paid, is frequently larger than everything else being negotiated.

Separation planning

Done properly this starts before the sale, not after. The seller maps every shared dependency — systems, premises, people, contracts, licences, data — and decides for each one whether it is split, duplicated, or provided under a service agreement afterwards. A seller that arrives at an auction with that map has a shorter process and a better price, because bidders are pricing certainty as much as cash flow.

Transitional services

Whatever cannot be separated by completion is provided by the seller afterwards, at an agreed price for an agreed period. The seller ends up running the business it has just sold, and neither side enjoys it. Two rules learned the hard way: agree the exit criteria as well as the duration, and price extensions in advance so the renegotiation nobody wants is already written down.

Why groups do this at all

Three arguments, and they are genuinely different:

• Focus. Management attention is finite and a division that is nobody's priority is worth more to somebody for whom it is the whole business.
• The conglomerate discount. The market frequently values a group below the sum of its parts, because investors cannot buy the piece they want. See the sum of the parts, where the arithmetic is done.
• Capital. Sometimes the sale is the point, and the proceeds pay down debt or fund something else.

Whether any of these actually creates value in a given case is not a question this page decides. What is true is that the mechanism is real and the costs of it are usually underestimated by everybody involved.

Carve-out or spin-off

The separation work is nearly identical; the transaction on top of it is not. A carve-out sells the division for cash to one buyer. A spin-off gives it to the existing shareholders as a second listed company: no buyer, no price, no cash. The choice depends on whether the group needs the money, whether a buyer exists, and whether the tax treatment of a distribution is available.

Stranded costs

The half nobody models. When a division leaves, the head office it was helping to pay for does not shrink by the same amount. The costs it used to carry are stranded on the remaining group, and the reported profitability of what is left falls even though nothing about it changed. A seller that has not planned for that has sold a division and damaged the rest.

What a buyer should price

• The standalone cost base, not the allocated one.
• The one-off cost of separation — systems, rebranding, new licences — which lands in the first two years and is real cash.
• The transitional services bill, and what happens if it runs long.
• Key customer contracts that need consent to move, which are a veto in somebody else's hands.

The formulas above are standard textbook formulations, simplified for teaching. They explain the mechanism — they are not a valuation tool, and they will not reproduce a dealer’s price.