Subtle Differences in Bond Risk Numbers

[kayy]

Hi Guys –

I came across a Risk Management system that produces 3 different risk numbers for bonds.

The specs in the system says that -

IRDV01 - obtained by 1basis parallel shift in the underlying IR curve

IR_SHIFT – Obtained by 1basis shift in the zero curve

CR01 - Obtained by 1basis shift in the Z-spread

All the above risk numbers look different from one another (not huge difference tough). But I don’t understand the differences arising here - whether we shift Z - Spread / Underlying IR curve / Zero curve by 1 bps shouldn’t the change in Present values be the same in all the cases? Since the new discount factors obtained should be the same after the shift in all the cases.


Many Thanks in advance.

 

[Ken Abbott]

When credit spreads are high, small shifts in the credit spread can imply movement closer to the default "frontier," with its LGD assumption. You shouldn't see any difference in PV01 and CSPV01 for hi-grade bonds, but as quality gets lower, you should expect them to diverge.

IRDV01 here is the coupon curve, as opposed to the zero curve, with all the reinvestment factored in.

Anyway, that's my intuition and I'm sticking to it.