Two items today. A friend of Eumaeus from Ireland writes to say they are disappointed that we did not cover the Jeffery and Smith paper (Equity Release Mortgages: Irish & UK Experience) as extensively as we might have done. Another friend wrote to express a puzzle about the Matching Adjustment principle. The principle suggests that we can construct a synthetic non-sovereign bond rate that is risk free, but which has a higher return than a (risk free) gilt, which we can use to discount the future liability. But if the gilt and the synthetic bond are certain to pay the same amount at maturity, how can their returns differ?
First, I am sorry we did not give the Jeffery and Smith paper more attention, but as we explained at the time, Eumaeus was taking his dogs on a Spring break. But I will discuss one point now, as it is connected with our other friend’s question about illiquidity. They argue thus:
How could we have two bonds with the same cash flows, but different prices? Does this not create an arbitrage for traders issuing the liquid bonds and collecting the illiquidity premium on the illiquid bonds?
To see why this arbitrage does not work, consider an illiquid but default-free bond. Suppose a trader borrows at the liquid risk-free rate to invest in this bond. The liquid risk-free borrowing must be liquid from the perspective of the lender, who can call in their loan at any time. If the lender calls in the loan early, the borrower must sell their asset into an illiquid market. The achievable price in a forced sale may be less than amount borrowed, so the lender now faces default losses. Lenders, anticipating the possibility of such losses, would not lend at the risk-free rate, and so the whole arbitrage unravels.
As it stands, that’s a really terrible argument. First, why should any market participant have to borrow in order to invest in the illiquid bond? The Matching Adjustment given to firms is of the order of 140bp upwards, which even without leverage is a considerable return. Forget borrowing, why not just buy the bond outright and earn the premium every year?
Second, even if there is an illiquidity spread, this is highly unlikely to be greater than 5%. Having traded illiquid bonds in the past, I can personally guarantee that the illiquidity spread was a matter of perhaps five basis points, not five hundred. So why not borrow 95% of the money at close to risk free, supply the other 5% myself as a haircut of ‘equity’, and earn close to 140bp on an initial investment of 500bp. That’s a guaranteed 28% return. Not even London Capital and Finance could beat that, and this would be absolutely risk free, from the premiss of the Jeffery and Smith argument (i.e. that the premium is due to illiquidity alone, and is not a risk premium).
There might be other reasons why the illiquidity premium exists, but this does not look like one of them. I will come back to other points in their paper next week.