The Vidyadhara Scholarship Trust has a Rs 10 crore corpus and must spend 5% a year on scholarships, with costs rising 6% a year. What return must it earn to keep the corpus intact in real terms (about 11.3%), how much equity does that need, and what spending rule protects scholarships in a bad year?

The corpus needs about 11.3% a year: 5 points to pay out, 6 points to grow with costs, and a little more because scholarships go out during the year rather than at its end. With debt assumed at 7% and equity at 12.5%, that takes about 78% in equity, which exposes the corpus to falls of 23% or so. A rule that spends 5% of the average corpus over the last three years, backed by a two-year reserve, cuts scholarships about 6% after such a fall instead of 28%.

A family that lives on the interest from a fixed deposit finds, ten years on, that the same interest buys far less; the capital was never touched, and it still shrank in what it could do. A perpetual spender has to earn its spending plus inflation, because the corpus must grow as fast as costs for the same 5% to keep buying the same scholarships. If scholarships were paid in one cheque at the end of the year, the sum is exact: 6% growth plus 5% spending is 11.0%. Paid in one cheque at the start, the corpus has only 95% working all year and the requirement is 11.6%. Paid through the year, as tuition actually falls due, it is about 11.3%, and that is the number to plan on.

On the trustees' own assumptions, a debt portfolio returns 7%, 4.3 points short. Equity adds 5.5 points over debt for every 100% of the portfolio, so closing the gap needs 78% in equity. A trust that must spend 5% real for ever is forced into a mostly equity portfolio, whatever its trustees' temperament, because nothing else is assumed to earn enough. The price of that is volatility: a 30% equity fall takes 23% off the corpus in one year. Compromise has a price too: at 60% equity the expected return is 10.3%, 1.0 points short, and a shortfall that small, compounded, leaves the corpus about 16% smaller in real terms after twenty years. The assumptions are illustrative; the trustees should rerun them with their own figures and with fund costs deducted.

Run the three obvious rules through a crash. Years one and two go to plan; in year three the corpus falls 23%; year four recovers 18%; year five is on plan. A strictly inflation-linked plan pays Rs 60 lakh in year four, which is 7.3% of a shrunken corpus and eats capital the deed says must last. A rule of 5% of the current corpus pays Rs 41 lakh, a cut of 28% that lands on students mid-degree. A rule of 5% of the average corpus over the last three years pays Rs 50 lakh, a cut of 6%, because the two good years before the fall still count. That is a smoothing ruleA spending rule that sets the payout from an average of past corpus values rather than the latest one, so that a single bad year changes spending only a little., and it is how endowments spend.

Two practical pieces complete the rule. Keep about two years of scholarships, Rs 103 lakh, in a liquid or short debt fund so that no scholarship is ever paid by selling equity in a fall; that reserve sits comfortably inside the 22% debt allocation. And write the rule into the trust's investment policy so the trustees do not have to decide under pressure. The averaging rule spends 6.1% of the corpus in the bad year, above the deed's 5%, and it lags on the way back up too: year five pays Rs 48 lakh against a plan of Rs 63 lakh, because the crash year stays in the average. That lag is the price of steadiness, and the trustees should confirm the deed allows the average to be the base, and if not, amend it. A spending rule chosen in a calm year is the only kind that survives a bad one.