Skip to content
← Resources / Articles
Carve-outs5 min read

A carve-out is two transactions wearing one signature page

You are buying a business, and you are also building an entity that has never existed. The second transaction has no purchase agreement, no closing date and no banker, which is precisely why it is where carve-out returns are won or quietly lost.

All M&A Research

Carve-outs and separations

Updated
Sunlight falling through the high windows of a factory building.
Sunlight falling through the high windows of a factory building.

Key takeaways

  • A carve-out is two transactions: the purchase, and the construction of a standalone business. Only the first has an agreement, a closing date and a fee attached to it.
  • Carve-out accounts are an allocation of the parent's costs, not a record of the business's own. What a standalone company actually needs usually exceeds what was ever allocated to it.
  • The transitional services agreement does more to determine the outcome than almost anything in the purchase agreement, and its scope has to be settled service by service.
  • People who make the business work are often outside the legal transfer population. Identify them by name during diligence, because after signing they belong to somebody else.

Corporate sellers are shedding non-core divisions at a pace that has outrun the market’s ability to staff the separations. For buyers, this is the most interesting supply of assets available: businesses that were never managed for their own performance, priced by sellers who want them gone, and improvable in ways that a well-run standalone company simply is not.

It is also the deal type where the gap between the model and the outcome is widest. Not because buyers overpay, carve-out entry multiples are usually defensible, but because the second transaction, the construction of a standalone business, is systematically under-scoped.

The carve-out financials are not financials

The seller will provide carve-out accounts. They are an allocation exercise, and they are optimistic in a specific, predictable way.

Allocated corporate cost is not the cost you will incur. The parent allocates overhead by revenue or headcount. Your standalone business needs a CFO, a controller, an HR function, a general counsel or an outside firm, an IT organisation, insurance in its own name, an audit, a treasury function and a board. In almost every case, the sum of what you actually need exceeds what was allocated. The delta is not a rounding error; on smaller carve-outs it can consume a material share of EBITDA.

Shared contracts do not travel. Software licences, logistics agreements, insurance, key supplier terms and customer master agreements are frequently held at parent level, at parent-level pricing. When you re-paper them at your own scale, prices move against you. Ask, early, for a schedule of every contract the business uses, who signs it, and what volume tier it sits in.

People in scope are not the same as people needed. The legal transfer population is defined by employment law and organisational charts. The people who actually make the business work include several who sit in the parent’s shared services and are not in scope at all. Identify them by name during diligence, because after signing they belong to somebody else.

Stranded cost is the seller’s problem: until it becomes a negotiating point. The parent will be left with overhead that no longer has a business under it. Sophisticated sellers try to move some of that into the TSA pricing. Know what you are being asked to absorb.

It is worth keeping the two cost categories apart, because they behave differently and sit with different parties. Stranded costs are the parent’s overhead left without a business under it. Dis-synergies are the costs that persist in the carved-out business for the foreseeable future, because it now buys its own insurance, audit, IT and procurement at its own scale rather than at the group’s.1 On the IT side in particular, stranded personnel, vendor contracts and licences can continue to inflate the retained cost base well past completion.2 Both belong in the model, with an owner against each, before the price is agreed.

The question to ask about every carve-out model is simple and rarely asked early enough: what does this business cost to run when nobody else is paying for anything?

The transitional service agreement is the deal

Most buyers treat the TSA as a schedule attached to the SPA. It deserves the same attention as the price.

Scope, not headings. “IT services” is not a scope. What you need is a service-by-service inventory with volumes, service levels, named dependencies and a defined output. If a service is not in the schedule, you will pay for it as a change request at a price set by a counterparty with no incentive to be generous.

Duration with genuine optionality. Twelve months feels comfortable in the data room and is frequently short. Buy a base term with unilateral extension rights at a pre-agreed price, and accept a step-up rather than an open negotiation. Sellers want you off their systems; extensions negotiated under time pressure are expensive.

Pricing and the cost-plus trap. Cost-plus sounds fair and is unauditable. Where possible, fix prices per unit of service. Where cost-plus is unavoidable, negotiate audit rights and a cap.

Exit ramps, per service. You will not leave the TSA in one event. You will leave payroll in month four, the ERP in month fourteen and one warehouse system in month twenty. Each service needs its own termination right without triggering a penalty on the rest.

Reverse TSAs. Frequently forgotten. The parent may need services from you: data, manufacturing capacity, a shared facility. Price it, scope it, and put a term on it.

Day one, and what actually has to be true

The separation plan that survives is the one built backwards from a small number of things that must work on the first morning.

People get paid. Payroll cutover is the single highest-risk workstream in any carve-out and the one with the least tolerance for slippage. It requires legal entities, bank accounts, tax registrations and a payroll provider: several of which have lead times measured in months, not weeks.

Customers can order and be invoiced. New entity details, new remit-to instructions, re-papered master agreements, EDI connections rebuilt. Every customer touchpoint that changes is an opportunity for a payment to go missing, and working capital is the first thing to deteriorate in a botched separation.

Suppliers keep shipping. Credit is granted to the parent, not to you. A new entity with no credit history and no trading record will be asked for prepayment or guarantees by suppliers who extended sixty days to the seller.

Somebody can close the books. A standalone business needs a first month-end. That means a chart of accounts, a general ledger, a consolidation, and a finance team who have done it in this configuration before, which, by definition, they have not.

The entity legally exists everywhere it trades. Entity formation, licences, permits, VAT and payroll registrations across jurisdictions is a long-lead workstream that is invisible on a synergy slide and fatal if late.

Where the returns actually come from

Carve-outs reward buyers for work, not for financial engineering. In the separations that work, value tends to come from three places with reasonable consistency.

The first is pricing discipline that the parent never applied. Divisional businesses inside larger groups are often priced to protect a group relationship or a bundled offer. Standing alone, they can price for their own margin, and frequently discover their product is worth more than the group believed.

The second is capital allocation attention. A division competing for capital against the parent’s core business loses that competition by design. The same capital expenditure request, approved in six weeks instead of six quarters, changes the growth rate.

The third is management who own something. For the first time, the people running the business have equity in the thing they run, and a board that talks about their business for the entire meeting rather than for eleven minutes.

None of those show up in the first year while the separation is consuming the organisation. That is the trade a carve-out buyer is making, and it is worth making: provided the model has been honest about the year in which nothing improves.


General commentary on market practice. Not legal, financial or tax advice.

Sources

  1. KPMG, Separation in practice and Building the value case, on distinguishing dis-synergies from stranded costs and assigning accountability for each. ↩

  2. Deloitte, The Art of Speed: accelerating the exit from IT Transition Service Agreements. ↩