A client is offered an 11% NCD from an A-rated housing finance company against 8% on AAA paper. With a 3% yearly default chance and 50% loss if it defaults, what is the expected return, and is the extra yield enough?

After expected losses the NCD yields about 9.5%, a pick-up of 1.5 points over AAA, half the headline gap. Expected loss is 3% times 50%, or 1.5% a year. The extra yield pays for the average loss with some left over, but over three years there is a 8.7% chance of a default costing Rs 25.0 lakh on this ticket. Hold it only as a small, diversified slice, well below Rs 50 lakh.

An insurer who charges a higher premium on riskier drivers is not richer for it if those drivers crash more often. Credit spread is partly a premium and partly a payment for losses that will, on average, happen. The piece that pays for losses is the expected lossThe average yearly credit loss: the chance of default multiplied by the share of the money lost if default happens.: 3% times 50%, or 1.5% a year. Only what is left over is extra return.

It is the standard shortcut, yield less expected loss. A stricter one-year version, where a default also costs that year's coupon, gives about 9.2%, so the pick-up is between 1.2 and 1.5 points depending on the convention. Say which one you are using. Either way, the NCD beats AAA on average, and the interview turns on whether the average is the right thing to look at.

On the average, yes, just. On the tail, it depends on size. Over three years there is about a 8.7% chance of a default, and on a Rs 50 lakh ticket that is a Rs 25.0 lakh loss, 12.5% of the client's Rs 2 crore. The expected gain from the pick-up is about Rs 75,000 a year. A client who would be shaken by a Rs 25 lakh hole should not buy a Rs 50 lakh ticket for Rs 75,000 a year of extra expected income.

Add the two costs the yield does not show. NCDs from smaller issuers can be hard to sell before maturity, so this is money that must be able to wait three years. And a housing finance company's troubles tend to arrive with a funding squeeze that hits several similar issuers at once, so two A-rated NCDs from the same sector diversify less than they seem to. The judgement: acceptable as a small slice spread across issuers, not as a Rs 50 lakh single bet.