Build a comps valuation for Tanvika Pharma from five peers trading between 18x and 42x earnings. Which peers do you keep, median or mean, and is a premium justified for Tanvika's higher US exposure?

Keep A restated, B, C and D; drop E; use the median of about 26.5x, and give Tanvika a premium of about 10%, roughly 29x. Peer A's one-off gain hides a normal 27x multiple, and Peer E is a contract manufacturer. Tanvika's faster growth argues for up to 37x, but heavy reliance on US plants argues for caution. At 29.2x, EPS of Rs 40 implies about Rs 1,166 a share.

When you price a flat by looking at nearby sales, you leave out the penthouse and the one sold in a family distress sale, not because they are high or low but because they are different. A peer stays if its business is comparable and its earnings are normal; a peer with a comparable business but distorted earnings is restated, not dropped. Peer A's 18x is low only because a one-time exclusivity gain inflates next year's earnings; remove the third that will not recur and its multiple is about 27x. Peer E's 42x reflects a contract manufacturing business with different growth and risk, so it goes.

Use the median, and say why. With five or six peers, one outlier moves the mean a lot and the median hardly at all. The raw mean of all five is 27.6x because Peer E drags it up; the raw median is 26x. On the clean set the two nearly agree, 26.5x and 26.25x, which is itself a sign the set is now sensible. When mean and median disagree by more than a turn or two, the peer set, not the average, needs work.

Exposure alone does not; what the US business earns does. Plain US generics face steady price erosion, which would argue for a discount. Tanvika's US revenue is hard-to-make injectables with few competitors, which is why it grows at 18% against the peers' 13%. Giving full credit for growth, a PEGP/E divided by the expected growth rate. Used to compare multiples of companies growing at different speeds.-style scaling of 26.5x by 18 over 13 gives 36.7x; that is the ceiling, not the answer. The offset is concentration: if one US plant supplies most of that revenue, a single failed regulatory inspection can stop it. A 10% premium, about 29.2x, credits part of the growth and keeps a haircut for the plant risk. At EPS of Rs 40 that is about Rs 1,166 a share, against Rs 1,060 at the plain median.

The limitation to say out loud: comps tell you what the market pays for similar businesses today, not whether the market is right. If the whole sector is expensive, comps will say Tanvika is fairly priced at an expensive level. Pair it with a discounted cash flow before taking a view.