Cost of debt

The discount rate has two components. This is the easier one, and getting it right is mostly a matter of refusing three tempting shortcuts.

It is the marginal rate, not the historical one

The cost of debt is what the company would pay to borrow today.

A firm with loans taken three years ago at 7%, whose bonds now yield 9.5%, has a cost of debt of 9.5%. The 7% is a historical accident of when the money was raised.

The reason is the question being asked. Capital budgeting evaluates the next project, funded with the next rupee — and the next rupee costs today's rate. Chapter 2's incrementality rule, applied to the right-hand side of the balance sheet.

So the answer to the problem is 9.5%, before tax. The 7% is a sunk financing decision.

It is after tax

Interest is deductible, so the government funds part of it:

$k_d^{\text{after-tax}} = k_d \times (1 - t)$

At 9.5% and a 25% tax rate:

$9.5\% \times 0.75 = 7.125\%$

That is the number that goes into WACC. The deduction is why debt appears cheaper than equity by more than the risk difference alone justifies, and chapter 10 examines how far that argument can be pushed.

Two conditions on the shield, both real:

It requires taxable profit. A loss-making company gets no immediate benefit; the deduction becomes a carried-forward loss, which is the deferred tax asset of chapter 11 of the Accounting subject, worth something only if profits arrive.

It depends on the applicable rate, which may differ from the headline statutory rate once exemptions and set-offs are counted.

Where to get the number

In descending order of reliability:

Traded bonds. The yield to maturity on the company's own bonds is the market's direct answer. Not the coupon — the yield, which is chapter 5 of the Fixed income subject's distinction.

A recent loan. What it actually paid to borrow in the last few months.

Rating-based estimate. Take the government bond yield for the matching tenor and add the credit spread for the company's rating. This is the usual method for an unlisted or unrated borrower, and the spread is the part to be careful with.

Interest expense ÷ average debt. The last resort, and the one most often used. It yields the historical average rate, which is the number this chapter opened by rejecting. In a period of changing rates it can be badly wrong in either direction.

The Indian specifics

Two features of the Indian market matter for a sensible estimate.

Most corporate borrowing is bank debt, not bonds. The corporate bond market is thin outside the top ratings, so for most companies there is no traded yield to read — the estimate comes from recent loan pricing or a rating-based build-up.

Bank lending rates are benchmark-linked. The RBI's directions on interest rates on advances govern how banks price loans, through the external benchmark and marginal cost of funds regimes. The practical consequence for a forecast: a floating-rate borrower's cost of debt moves with the benchmark, so a cost of capital computed in a low-rate year understates the cost of a project financed through a rising one.

What the cost of debt is not

It is not the interest actually paid this year. That is history.

It is not the risk-free rate. A company is not the government, and the spread between them is the market pricing default risk.

It is not zero for a company with no debt. An unlevered firm still has a cost of debt — the rate it would pay — which is needed for any analysis of what a different capital structure would do.

The point

The cost of debt is the yield the company would pay to borrow today, not the coupon on loans raised years ago, because capital budgeting funds the next project with the next rupee. It enters WACC after tax, since interest is deductible — a shield that requires taxable profit to be worth anything now. In India most borrowing is benchmark-linked bank debt, so a floating-rate borrower's cost moves with the policy cycle.