Our market research is showing some movement in the last year, but also the last few months, in ideas about collateralization levels in securities lending. This article looks at some of our initial findings.
Here is what we’ve got so far:
1) In Europe, the growth of equities as non-cash collateral has pushed a reevaluation of the risk that taking equities entails. Beneficial owners, working with agent lenders and market data vendors, have found that the correlation risk of taking equities as collateral in a typical stock borrow loan transaction are fairly reasonable. Not perfect, but reasonable for the risk that lenders are taking on. The flip side of this conversation is that beneficial owners are asking for, and getting in multiple observed cases, 110% collateral when taking equities as collateral. Beneficial owners report better fees for loans against equities than for government bonds or cash.
2) In the US, equities remain a rare collateral class for securities lending – we know of just one big beneficial owner that will happily take equities. On the other hand, we see plenty of treasuries, agencies and 15c3-3 eligible collateral (this includes credit card debt, student loans and the like). We see a very limited amount of corporate bonds. The collateralization levels have not moved however – equities get lent at 102% and treasuries go lower in a customary trade. We haven’t seen US beneficial owners get excited about raising collateralization rates based on the types of collateral they receive. The real movement here is in fees, and there are some parallels between the rates paid by borrowers for alternative collateral against securities lending trades and the rates that repo dealers pay to cash borrowers for these collateral types.
3) In an indemnified trade, why should beneficial owners be concerned about margin levels? In the case of counterparty default, an agent lender will liquidate collateral first and then be obligated to make up the difference out of their own pocket. As we are still in the world of indemnification, this means that concerns about collateralization moves over to the agent lender and doesn’t reside at the beneficial owners’ office.
4) For beneficial owners not using an agent lender, we’re seeing the same patterns. We think that the reason for this is competitive pressure; given that self-lenders are much smaller than the agent lender community, any action to increase collateralization levels taken by self-lenders on their own will fail unless the biggest agent lenders move in this direction also. With the exception of rare cases like equities as collateral, we don’t see this happening.
5) We’ve done some thinking about agent lender risk models and their comfort with counterparties. In truth, if an agent lender is willing to lend to a given borrower, a difference of 1% or 2% in collateralization probably isn’t going to make much of a difference in their decision matrix. 25% might make a big difference, but we have no evidence of that happening in the agent lender world right now. So really, if agents are willing to deal with a selected counterparty, they are largely comfortable that the counterparty will return their borrowed securities.
6) Meanwhile, outside of the bilateral world, some CCPs think that the right collateralization levels for securities lending are 120% or 140%. As one example, the Taiwan Stock Exchange is reported to have this range. This suggests that for a non-agent lender transaction, made between two counterparties that meet certain minimum risk criteria but that are otherwise unknown to each other, a neutral intermediary thinks that a much higher level than 102% is required. This is an interesting conclusion and makes us wonder about the outcome of the Financial Stability Board’s phase 2 of its securities lending and repo inquiries (a link to the press release for the Quantitative Impact Study is here.)
What haircuts/margin levels are right in securities lending is a question for which many people have different answers. Sometimes the same person has more than one answer. But we think that this conversation is not laid to rest, especially due to both the upcoming FSB report and the generally increased attention that beneficial owners are paying to collateral across asset classes and financing products. We expect additional experimentation, testing and of course regulatory recommendations and rules in the next few years.