Relative Valuation India

Relative Valuation (India)

Multiples are not a valuation method. They are a way of asking "what is the market assuming about this company relative to that one?" Used carelessly they smuggle in every assumption you thought you were avoiding.

Choosing the multiple

Rule: use at least two multiples with different denominators. If P/E and EV/EBITDA disagree, the gap is leverage, other income, or tax — find out which.

Building the comp set

An Indian comp set is not "companies in the same sector on the exchange". Screen on:

1. Business model, not sector label. An IT products company is not comparable to an IT services company.
2. Size within roughly an order of magnitude
3. Growth within a comparable band
4. Return on capital — the single biggest driver of multiple dispersion in Indian equities
5. Capital intensity and working capital cycle
6. Governance and float quality — a promoter-pledged, thinly traded name is not comparable to an institutionally-held large cap even in the same business

Include, and label separately, the global comparables. Indian names in the same business routinely trade at 1.5–2.5× the multiple of their global peers. Do not "correct" this — explain it.

Why Indian multiples are structurally higher

When asked why an Indian company trades at 40× while its global peer trades at 18×, the honest answer has these components. Name them; do not hand-wave "India premium".

• Higher nominal growth. Indian nominal GDP growth in the high single digits versus 3–4% in developed markets mechanically supports a higher terminal multiple.
• Higher cost of equity — which works the other way. India's Ke is higher (Rf ~6.5–7% plus ~7% ERP), which should compress multiples. That it does not means growth expectations are doing heavy lifting.
• Domestic flow structure. Sustained monthly SIP inflows into domestic equity mutual funds create a price-insensitive bid, particularly in mid and small caps.
• Limited free float. High promoter holding shrinks investable float; index and fund demand chases a small supply.
• Scarcity of listed exposure to certain themes.
• Quality skew. Indian large caps that survive selection have genuinely high ROCE.

The test of whether the premium is justified: does the implied growth-and-return combination survive a reverse DCF (see dcf-india)? Usually the honest conclusion is "partly justified, partly flow-driven".

India-specific adjustments — do these every time

1. Ind AS 116 (leases)

Since Ind AS 116, operating leases sit on the balance sheet as a right-of-use asset and a lease liability. Rent leaves EBITDA and reappears as depreciation and interest.

• EBITDA is inflated for retail, QSR, aviation, hospitals, hotels and warehousing.
• EV must include the lease liability as debt, or EV/EBITDA is systematically understated.
• When comparing an Indian company to a peer on a different standard or a pre-116 history, restate one side. Say which.

2. Holding company discount

Indian conglomerates and promoter holdcos trade at large discounts to the sum of their stakes.

NAV = Σ (stake % × market cap of each listed holding)
      + fair value of unlisted businesses
      + net cash − net debt at the holdco

Holdco discount = 1 − (holdco market cap / NAV)

Observed discounts in India commonly run 40–70%, driven by: no operating control transfer, dividend leakage and double taxation, poor capital allocation history, and the absence of a catalyst. Do not model the discount closing without naming the catalyst that closes it (demerger, buyback, listing of an unlisted arm, a stated capital-return policy).

3. Cross-holdings and treasury

Strip investments in group companies out of the operating entity before computing an operating multiple, and value them separately. Otherwise a company looks expensive on EV/EBITDA because a third of its EV is a stake in something else.

4. Standalone vs consolidated

Use consolidated for both the subject and every comp. Mixing them is a silent error.

5. Other income

Indian companies frequently carry large treasury books. Compute a core P/E:

Core earnings = PAT − post-tax other income
Core P/E = (Market cap − surplus cash − investments) / Core earnings

6. Promoter and float effects

Adjust for: promoter pledge (a discount), a very low free float (a liquidity premium in bull markets that reverses violently), and pending or likely index inclusion or exclusion.

Where to anchor a "fair" multiple

Three anchors, in order of usefulness:

1. The company's own history. 5-year and 10-year median forward P/E and EV/EBITDA, plus the standard deviation band. Where does it sit today, and has the business changed enough to justify a re-rating?
2. The peer set median, regressed against ROCE and growth. A simple cross-sectional regression of EV/EBITDA on ROCE and revenue CAGR across the comps will usually explain most of the dispersion and tells you whether the subject is cheap for a reason.
3. The implied multiple from a DCF. If the DCF and the peer median disagree by more than ~25%, one of them contains an assumption you have not examined.

Also useful for context: Nifty 50 and Nifty Midcap 150 index-level trailing and forward P/E versus their own long-run medians, published by NSE.

Output format

# <Company> — Relative Valuation, <date>

## Comp set
| Company | Mcap ₹cr | Rev CAGR 3y | EBITDA mgn | ROCE | Net D/E | P/E (fwd) | EV/EBITDA | P/B |
|---|---|---|---|---|---|---|---|---|
| <subject> | | | | | | | | |
| peer 1 ... | | | | | | | | |
| **Median** | | | | | | | | |
| *Global comps (labelled separately)* | | | | | | | | |

## Adjustments applied
- Ind AS 116 lease liability added to EV: ₹X cr
- Cross-holdings valued separately: ₹Y cr
- Core P/E (ex-other income): ...

## Own-history band
| | 10y median | 5y median | Today | z |
| Fwd P/E | | | | |
| EV/EBITDA | | | | |

## Assessment
<Where it trades, versus peers and versus itself, and the two or three reasons for the gap>

## Implied value
| Method | Multiple | Value/share |
| Peer median EV/EBITDA | | |
| Own 5y median fwd P/E | | |
| DCF cross-check | | |

Hard rules

• Date the multiples and name the earnings basis — trailing twelve months, FY26E consensus, or your own estimate. Comparing your forward estimate to a peer's trailing multiple is not a comparison.
• Never average multiples across companies with very different leverage without moving to an EV basis.
• Exclude outliers explicitly, and say why. A loss-making peer has no P/E; dropping it silently biases the median.
• A cheap multiple is a question, not a conclusion. Run forensic-redflags-india before calling anything a bargain.
• Analysis, not advice.