In Tilottama Mutual Fund's Rs 800 crore corporate bond scheme, the top 10 investors hold 58%, and one treasury holding Rs 150 crore gives notice to exit. The scheme has 11% in liquid assets. What happens to the other investors, and what should the AMC have been monitoring?

The exit uses up the whole liquid buffer and still needs Rs 62 crore of bond sales, so the investors who stay end up in a scheme with no cash cushion. The direct cost is small, about Rs 31 lakh. The damage is what is left: the next two large holders are 17% of a smaller scheme. The AMC should have sized liquidity to its largest investors and watched concentration every day.

Think of a shared taxi where one passenger is paying most of the fare. When that passenger gets out early, the others either pay more or walk. A mutual fund pays every leaving investor at NAV, so any cost of raising the cash lands on the investors who remain. The treasury leaves with Rs 150 crore at a fair price. Whatever discount the fund takes to sell bonds afterwards comes out of the assets the other investors still own.

Here the redemption is 18.75% of the scheme against a buffer of 11%. The Rs 88 crore of liquid assets covers only 59% of it, so Rs 62 crore of bonds must be sold within the settlement window.

There are two ways to raise the money. The easy way spends all Rs 88 crore of liquid assets and then sells Rs 62 crore of the most tradeable bonds. At an assumed discount of 0.5% to fair value, that costs about Rs 31 lakh, or 4.8 basis points on the Rs 650 crore left. The explicit cost is small; the real cost is the portfolio left behind, with no liquid assets and the hardest to sell bonds now a larger share.

That matters because the next two investors hold Rs 110 crore, now 16.9% of the scheme. If they follow, there is no buffer, and the fund must sell thinly traded bonds at perhaps 1.5% below fair value: Rs 1.65 crore, 31 basis points on the Rs 540 crore then left. The second wave costs about ten times the first. The better way is a vertical sliceSelling a proportional share of every holding, liquid and illiquid alike, so the remaining portfolio keeps the same mix as before the redemption.: sell about 19% of every holding, including some less liquid bonds, so the investors who stay keep the same mix and the same 11% cushion. It costs a little more today and protects them far better tomorrow.

Three numbers, every day. First, concentration: the share of the largest single investor and of the top ten. Second, coverage: liquid assets divided by the holding of the largest investor, here 0.59x, when anything under 1.0x means one phone call empties the buffer. Third, a redemption stress test that asks what happens if the top few leave in the same week. SEBI requires debt schemes to run liquidity stress tests and disclose risk measures; confirm the current requirements, but the logic does not depend on the rule. A scheme's liquidity must be sized to its largest investors, not its average one.

Covering the largest holder is not free. Moving the buffer from 11% to 18.75% at a 1% yield give-up costs about 7.8 basis points a year for every investor. That is the price of insurance against a single phone call, and it is cheap next to a forced sale in a bad week. The other levers are to cap how much any single investor may hold, to talk to large holders about staggering big exits, and to keep the relationship team telling the fund manager about any exit it hears of early.