Sum-of-the-parts and holding-company discounts

Some companies are one business. Many of India's largest listed groups are not — they are a consumer arm, an infrastructure arm, a chemicals arm and a finance arm, bolted onto one balance sheet and reported under one share price. For these, the tidy tools break. A single P/E or a single EV/EBITDA on the whole is a category error, because it forces one number to describe businesses that could not be more different — a steady, high-return consumer brand and a lumpy, capital-hungry construction contractor do not deserve the same multiple, and blending them buries the truth in an average.

The honest method is older and more patient. Value each part separately, on the multiple its own economics deserve, then add them up. This is sum-of-the-parts — SOTP for short. It treats a conglomerate as a portfolio of businesses rather than a single organism, prices each on its own terms, and only then reassembles the total, subtracting the group's debt and central costs to reach a value for the shares.

And then there is a special, stranger case: the holding-company discount. A company whose main asset is stakes in other companies almost always trades below the market value of those stakes. Understanding why — and when that gap is a trap versus an opportunity — is the second half of this module, and one of the most misunderstood ideas in Indian markets.

A multiple is shorthand for a business's growth, its returns, its risk and its capital hunger, all compressed into one figure. A stable consumer brand that grows steadily, earns high returns on little capital, and rarely stumbles genuinely deserves a rich multiple. A construction contractor whose profits swing with the order cycle and who must sink cash into every project deserves a modest one. These are not opinions; they are the market's consistent verdict on the two kinds of economics.

Staple them together and value the pair on a single blended multiple, and you commit two errors at once: you overvalue the weak business by lending it the strong one's rating, and you undervalue the strong business by dragging it down to the weak one's. The blend feels balanced. It is in fact wrong in both directions simultaneously, and the "balance" is precisely what hides the double mistake. Each segment carries its own story — its own growth, returns and risk — and a story that specific cannot be told by one number stretched across all of them.

SOTP exists to respect that. It insists you say, out loud and separately, what each part is worth and why — this multiple for the consumer arm because of its returns, that multiple for the cyclical arm because of its swings, this method for the finance arm because banks are valued differently again. The discipline is not the arithmetic (adding is easy); it is being forced to justify each part on its own merits rather than hiding a weak business inside a strong group's average.

SOTP is a four-step build, and the arithmetic is deliberately plain. illustrative

Take Bharat Diversified Ltd, a composite group with three arms. Value each on the multiple its own economics deserve, sum to a group enterprise value, then subtract net debt to reach the value of the shares.

Read the steps. One: each operating arm is valued on the multiple that fits its economics — 15× for the steady consumer business, 7× for the lumpy contractor. Two: the listed finance stake is taken at its own market value rather than a multiple, because the market already prices it (and a finance arm needs its own method — the next module). Three: the three add to a group enterprise value of ₹5,450 crore. Four: subtract net debt of ₹950 crore to reach an equity value — the net asset value, the worth of everything owned less what is owed — of ₹4,500 crore. That is your SOTP fair value for the shares.

Now compare it to what the market pays. If Bharat Diversified's market cap is ₹3,600 crore, the group trades at a 20% discount to its own parts — the conglomerate discount, the SOTP cousin of the holdco discount. The market is saying: owning these four businesses bundled together, run by one central management, is worth less than the four would fetch apart.

The purest form of this puzzle is the holding company — a firm whose main asset is shares in other firms. Watch one read. illustrative

Anchor Holdings Ltd owns a 60% stake in a listed operating company. That subsidiary has a market cap of ₹10,000 crore, so Anchor's stake is worth 60% × 10,000 = ₹6,000 crore. Anchor has little else — a small cash pile, some central costs. Yet Anchor's own market cap is only ₹4,200 crore. It trades at a 30% discount to the very stake it holds. You can see the ₹6,000 crore on the screen; the market values Anchor's slice of it at ₹4,200 crore.

Why would anyone accept less for the same shares held one layer up? Because owning through a holdco carries real frictions. You do not control the subsidiary's cash — dividends arrive at the parent's discretion, not yours. If Anchor ever sold the stake, capital-gains tax would take a bite. The parent has its own overhead. And crucially, you are trusting Anchor's management to allocate the dividends it receives well — a promoter who hoards cash or invests it poorly can trap value at the holdco level indefinitely. The discount is the market pricing every one of those frictions.

A wide, unjustified holdco discount can be a genuine margin of safety — but only when something is likely to close it; a discount with no catalyst is not a margin, it is a permanent cost of ownership. That distinction — discount-with-catalyst versus discount-forever — is the entire art of reading a holding company.

Sum-of-the-parts looks rigorous — a neat table, each line justified — and that neatness hides several things it cannot capture.

It cannot see the interactions between the parts. SOTP assumes the businesses are separable, but a group may share a brand, a distribution network, a treasury, or a management team. Sometimes those links create value the standalone sum misses; more often, a weak arm quietly drains cash and attention from a strong one in ways no per-segment multiple records.

It cannot value the central management's capital allocation. The single biggest driver of a conglomerate's long-run value is what the centre does with the cash the parts throw off — and that is a matter of judgement about people, not a multiple. A brilliant allocator makes the whole worth more than the parts; a poor one makes it worth far less, and no SOTP table shows this until years have passed.

It cannot tell you the discount will ever close. This is the hardest limit. A conglomerate or holdco discount can be entirely rational and entirely permanent. The market price already reflects the frictions of the structure; assuming those frictions will melt away just because you can compute a higher "parts" number is wishful arithmetic.

And it inherits every weakness of the multiples it uses. If you value the consumer arm at 15× because peers trade there, and those peers are in a bubble, your "parts" are as inflated as the market you borrowed from.

Holding companies and conglomerates are a rich field for self-deception. The recurring traps:

1. "The discount is free money." The most seductive error. A 30% discount is not a coupon you clip — you cannot reach the underlying value without control, a sale, tax and time. The discount exists because those frictions are real.

2. "Deeper discount, bigger bargain." Depth alone guarantees nothing. Some of the widest, most durable discounts in India belong to holdcos that never return value to shareholders. Without a catalyst, a deeper discount is often just the market's firmer verdict that the value will never arrive.

3. Ignoring capital allocation. Two holdcos at the same discount are worlds apart if one pays through its dividends and buys back stock while the other hoards and empire-builds. The discount is half the story; what the controller does is the other half.

4. Using peak-cycle multiples on the parts. Valuing a cyclical arm at its boom-year EBITDA times a boom-year multiple double-counts the good times. The parts must be valued on normalised, mid-cycle terms, or the sum is a mirage.

5. Forgetting the taxes and costs between you and the value. SOTP that ignores the capital-gains tax on unlocking a stake, the central overhead, and inter-company debt overstates what a shareholder could ever actually receive.