Sum of the parts
Also known as: SOTP, Break-up value
Each division valued separately and added up — used to argue a group is worth more apart than the market says it is worth together.
1 · SnapshotThe one idea to remember
2 · BeginnerWhat actually happens?
Some companies are really several companies wearing one name. A group might own a stable utility, a fast-growing software business and a property portfolio. The market gives the whole thing one share price and one multiple.
A sum-of-the-parts analysis takes it apart. Each business is valued against companies that do only that — the utility against utilities, the software against software — and the answers are added up.
Very often that total is higher than what the market is paying for the group. The gap is called the conglomerate discount, and this analysis is how it gets measured. It is the standard argument for selling a division, spinning one off, or breaking the whole thing up.
Two things have to be true for it to mean anything. The parts must be genuinely separable — two divisions sharing a factory and a sales force are one business however they are reported. And the head office costs have to go somewhere. Leaving them out is the easiest way to make any group's parts add up to more than its share price.
- 1
Defining segments2–4 days
Which businesses are genuinely separable, which is harder than the reporting segments suggest.
- 2
Allocating costs3–6 days
Central costs have to be pushed down, and the answer moves a great deal depending on how.
- 3
Valuing each3–7 days
Each segment gets its own peers and its own multiple, which is the whole point of doing this.
- 4
Bridging to equity1–2 days
Debt, pensions, minorities and unallocated costs are deducted from the total.
Are they separable — Operational reality decides. Two divisions sharing a factory, a sales force and a brand are one business however they are reported.
Where do the central costs go — The analyst decides. Leaving head office out of the segments is the single commonest way this analysis flatters the answer.
Who is on the deal
| Who | Side | What they are actually for |
|---|---|---|
| The analyst | Neither | Decides where the central costs go, which moves the answer more than any multiple does. |
| The company | Sell side | Reports segments in a form chosen for accounting rather than for this analysis. |
| Activist investors | Buy side | Frequently commission exactly this work to argue a group should be broken up. |
| The market | Neither | Applies a discount to conglomerates that this analysis is an attempt to measure. |
- Desk
- Valuation & Deal Analysis
- Values
- Each business on its own peers and its own multiple
- Used to argue
- That a group should be broken up
- Moves the answer most
- Where the central costs are allocated
- First test
- Are the parts genuinely separable
What decides whether it completes
Not how hard this is, and not a rating — there is deliberately no total. It says which of five blockers decides whether this transaction happens at all, in the same order on all 70 transaction types so they can be compared. This publication's own reading; see the notice below.
- Pricedecides it
- Financingbarely applies
- Approvalbarely applies
- Diligencedecides it
- Executionmatters
What decides it here. Whether the parts are genuinely separable and where the central costs are allocated decide the answer, and both are judgements made before any multiple is applied. Leaving head office out of the segments is the commonest single way this analysis flatters its own conclusion.
3 · IntermediateHow it runs in practice
The steps
- Define the segments — which businesses could actually stand alone.
- Get each one's earnings on a standalone basis, which means allocating the costs the parent currently absorbs.
- Value each with its own peers and its own multiple.
- Add them up to get a total enterprise value.
- Bridge to equity — deduct net debt, pensions and minorities, and deduct the capitalised value of any central costs that would remain.
The central cost problem, precisely
A head office costs money every year. If the group is broken up, some of that disappears and some of it is duplicated across the new companies. Three treatments are used:
- Allocate to the segments — most conservative, and it lowers each division's earnings.
- Capitalise the total and deduct it as a negative stub.
- Ignore it — which is common, wrong, and produces the flattering answer.
Why the discount exists at all
- Investors cannot buy the piece they want and pay a lower price for a package containing something they did not choose.
- Capital allocation inside a group is decided by a head office rather than by each division's own cost of capital.
- Disclosure — a division inside consolidated accounts is a segment note; as a listed company it publishes everything.
- Analyst coverage follows sectors, and a group that spans three does not fit any of them.
Who uses it
Activist investors, most visibly, to argue for a break-up. And boards themselves, before deciding on a carve-out or a spin-off.
4 · AdvancedThe numbers & the documents
Separability, tested properly
The question is not whether the divisions are reported separately. It is whether they could operate separately, and the checks are concrete: shared customers, shared systems, shared brand, shared manufacturing, shared distribution, shared people. A division that depends on the group for any of those is not a standalone business, and valuing it as one is a category error rather than an aggressive assumption.
This is the same analysis a carve-out has to do for real, and it is why carve-outs take a year of separation work before anything can be sold.
Dis-synergies, which are usually omitted
Separating businesses destroys some value as well as releasing it: lost purchasing scale, duplicated functions, contracts negotiated as a group, and the one-off cost of separation itself. A sum-of-the-parts that counts the release and not the loss is only half an analysis, and the missing half typically consumes a meaningful part of the gap.
Tax, which frequently decides it
A group cannot always be separated tax-neutrally. A break-up that triggers a large charge may destroy more than the discount is worth, and the tax analysis is therefore a gating question rather than a detail. It is the reason many demonstrably discounted conglomerates stay as they are.
When the discount is correct
Sometimes there is no discount to close, because the market is right. If the group genuinely shares customers, technology and management, the parts are not separable and the single multiple reflects that. The analysis then measures something that cannot be captured, and saying so is a better answer than a number.
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.
5 · Desk notesHow people on the deal think about it
Now say it back
Close the page and give Sum of the parts in four sentences. It takes a minute and it is the only way to find out whether reading it was enough.
- Who wants what — name both sides and what each one is actually trying to get.
- What has to happen, in order — the three or four stages, not the whole timetable.
- Where the money comes from — cash, new shares, or borrowed; somebody has to fund it.
- What kills it — the ordinary way, not the dramatic one.
Where this transaction shows up elsewhere
- EasyActivist campaignDealA small stake and a public argument
- MediumCarve-outDealSelling part of a group that was never a company
- MediumValuationDeskThe arithmetic underneath every transaction: discounted cash flow, comparables, precedent transactions, the buyout…
- HardCost of capitalDealThe rate everything is discounted at, assembled from inputs that are mostly estimates of things nobody can observe