Earlier this week, Golden Retriever (who is a bit of a dog) asks
Would also be great to get any thoughts you have on why Network Rail bonds yield more [20bp] than gilts even though they are guaranteed by the UK government
This is an objection to my earlier observation that illiquid spreads amount to a matter of basis points.
Now my observation was anecdotal. In 2008 I was responsible for the risk management of a large bond portfolio dominated by a single position of nearly £1bn in an illiquid French bond. I can’t remember exactly the issuer, but it was a form of sub-sovereign issued by a French region. It was around March 2008 when the storm was raging with unimaginable intensity, many ships on the rocks etc, and we were desperate to unload the position, which we finally did, within 5bp of risk free (might even have been 2bp).
That’s anecdotal, I agree. I checked out Network Rail Bonds and true, their yield is at a premium to gilts. This document tells us that “although Network Rail has held a UK government backed guarantee since 2004, its debt has nevertheless traditionally traded at a premium to gilts due to its bonds being less liquid, not featuring in government bond indices and due to a residual perception of enhanced credit risk.”
As always, the de-composition of liquidity and credit risk is awfully hard to quantify. Another Eumaeus correspondent suggests that, putting ongoing liquidity preferences aside, the transaction costs of simply acquiring the Network Rail bond may be greater than those incurred in buying a gilt. Also, he suggests that if we find ourselves in a world where the UK government is thinking about defaulting on stuff, it might default on its Network Rail guarantee before its gilt payments. So the credit risk must be greater, though quantifying that difference is not trivial.
But leaving all that aside, it is undeniable that insurers want illiquidity premia to be triple figures in basis points, so they aren’t very interested in inferring them from Network Rail bond spreads. Putting a ceiling of 20bps on a reasonable illiquidity premium assumption, and putting a limit like that on the MA yield pick-up would be pretty impactful for UK life, no?
Let’s leave it at 20bp max. Woof.