A small cap fund grew from Rs 5,000 crore to Rs 30,000 crore in four years while its yearly alpha fell from 4.1% to 0.8%, and its median holding trades Rs 22 crore a day. Is it still worth holding, and at what size does the strategy stop working?

The alpha decline tracks size closely, about 1.3 points lost for every doubling of AUM, and on liquidity the strategy's comfortable capacity is roughly Rs 9,400 to 12,600 crore. At Rs 30,000 crore the median holding takes about 48 trading days to sell, so the manager can no longer act on his ideas quickly. With 0.8% alpha against 5% tracking error, tell the platform to stop recommending fresh money and to look for a smaller fund.

A good street food stall can serve a hundred customers with the same quality; ask it to serve ten thousand and it has to change its recipe. A small cap manager's edge is finding mispriced stocks that trade thinly, and the same thinness that creates the mispricing limits how much money can use it. As the fund grows, each idea must be a larger position, which takes longer to buy and sell and moves the price against the fund on the way in and out. The manager either holds more stocks, diluting his best ideas, or moves up to bigger, better-researched companies where the edge is smaller.

The numbers fit that story closely. Fitting alpha against the logarithm of AUM gives a line that loses about 1.28 points of alpha for each doubling of size, and it reaches zero around Rs 45,000 crore. Five points do not prove cause, but a steady decline that follows size this closely, with the same manager and process, is the classic capacity signature.

Estimate capacity from liquidity, not from returns, because returns are noisy and liquidity is measurable. A common desk rule is to trade no more than about 20% of a stock's daily volume, here Rs 4.4 crore a day for the median holding. If the manager wants to be able to exit a median position within 15 trading days, that position can be at most Rs 66 crore, and at 0.7% a position the fund can be at most about Rs 9,400 crore. Allowing 20 days gives about Rs 12,600 crore.

At Rs 30,000 crore the median holding takes about 48 trading days to sell, so a change of view takes two months to act on. In a falling market, when every small cap fund is selling at once, it would take longer still. That is the hidden risk of a large small cap fund: its stated holdings look liquid only until everyone needs to sell.

Alpha of 0.8% against 5% tracking error is an information ratio of about 0.16, statistically indistinguishable from zero. Before blaming size alone, check one alternative: if small caps rallied broadly, many active funds' alpha may have shrunk at the same time, so compare Tarkshya with smaller peers over the same years. If smaller peers kept their alpha, size is the cause. The recommendation to the platform is to stop directing new money to the fund, keep existing investors informed rather than forcing a switch that triggers exit loads and capital gains tax, and favour funds still inside their capacity, or a low-cost small cap index fund. An AMC that soft-closes a fund by limiting lump-sum inflows is showing the discipline you want to see.