A client needs Rs 60 lakh for an MBA abroad. Should she take an education loan at 10.5%, with the interest deduction treated as a framework to confirm, or redeem from a portfolio expected to earn 11%?

Sell the Rs 20 lakh of debt funds and borrow the other Rs 40 lakh. Debt funds leave about 4.9% after tax, less than the loan costs even with the deduction, so they should go first. Equity is expected to leave about 10.5% after tax against a loan cost of 7.35% once she earns, a thin edge that fails about 33% of the time over five years. Borrowing Rs 40 lakh costs about Rs 4.2 lakh a year before the deduction.

The after-tax cost of the loan against the after-tax return on each holding she would otherwise sell. Keeping a fixed deposit at 7% while paying 12% on a credit card balance loses money every month; the question is the same, done asset by asset. The loan costs 10.5% until she has income and, if the deduction applies at a 30% slab, about 7.35% after that. A portfolio's average of 11% is the wrong yardstick, because she can choose which part to sell.

Because they earn less than the loan costs. 7% taxed at a 30% slab leaves about 4.9%, below even the loan's after-deduction cost of 7.35%; borrowing to keep them would pay more in interest than they earn. Selling Rs 20 lakh of debt funds covers a third of the need with no leverage at all. The only reason to keep them would be an emergency reserve, and she should keep a separate reserve of a few months' costs for that.

Borrow, knowing the edge is thin. Equity's expected after-tax return of about 10.5% beats the loan's 7.35% once she works, but with yearly swings of about 16 points, the five-year average falls below the loan cost about 33% of the time. During the study years, with no deduction, the comparison is 10.5% against 10.5%, a loss about 50% of the time. The gain, about 3 points on Rs 40 lakh a year on average, is real but modest; in Benjamin Graham's terms the margin of safetyThe gap between the expected result and the level at which a decision stops paying, kept as a cushion against being wrong. is thin, which argues against borrowing more than the Rs 40 lakh.

Close with what could change the answer. If her post-MBA income is uncertain, the deduction may be worth less and the loan's cost closer to 10.5%, which removes most of equity's edge. A loan also adds a fixed payment to her first working years. The decision rule is simple and portable: sell what earns less than the loan costs, borrow instead of selling what earns clearly more, and keep the borrowing small when the edge is thin.