DTCC-Euroclear GlobalCollateral Ltd and PricewaterhouseCoopers recently published a white paper, “Implications of Collateral Settlement Fails.” The paper claims that “the total value of collateral settlements that are expected to fail is likely to exceed $27bn.” The implication then is that better infrastructure like GlobalCollateral Ltd will reduce fails substantially. What to make of this big number? Let’s unpack what went into this calculation and see if it makes sense.
The authors collected a data set from mostly North American banks on collateral movement volume, collateral settlement fail rates, collateral settlement team size, and time spent remedying collateral settlement fails. We know the group who did the survey and think it is safe to assume that they spoke with the right people and that the underlying data are correct.
We can also validate that following introduction of non-cleared derivatives margin rules later this year (or next if they get postponed again), that the volume of collateral movements will increase substantially. In a January 2014 paper, “Trends, Risks and Opportunities in Collateral Management,” DTCC guessed that the increase might be 500% to 1000% times today’s volumes. This was a pie-in-the-sky figure that has not been proven anywhere in particular, but we can all agree that everyone posting Initial Margin to each other plus Variation Margin, none of it nettable, means a lot of messages and movements going back and forth. A proportionate number of fails will follow without some sort of intervention. That fails will increase proportionately is at the heart of the DTCC and PwC argument for US$27 billion.
Part of their argument also revolves around operational complexity. They say: “Participants with portfolios containing both cleared and bilateral products will be required to separately post collateral to FCMs/SCMs and then to CCPs and counterparties. Additionally, the segregation of counterparty IM will result in a need to engage independent segregation services in third/tri-party segregation models.” They further segment the causes of operational fails by “Miscommunication, Constrained Technology, Insufficient Collateral, and Counterparty Insolvency.”
Here’s where we’ll take a step back though, because while all of these new connectivity points are real, it is also the case that vendors and collateral management outsourcers are well on top of resolving constrained technology. They are also helpful but not solely responsible in resolving miscommunication, which the paper notes is the main cause of fails. In fact, in our October 2015 survey of collateral management technology companies, we found that market connectivity was the #1 priority by far of technology vendors. Our November 2015 survey of collateral management outsourcers found a similar importance on connectivity, although not by as wide a margin compared to other priorities. So while market participants do have many new operational processes, their vendors are already working on centralization of the process. That said, there is no solving for insufficent collateral and a counterparty default with internal technology or market infrastructure; you have to leave that one up to the other side acting responsibly.
The paper goes through the key assumptions made in getting to the US$27 billion figure, which we won’t repeat here, but they make general sense. The big difference we see is that current collateral settlement fails rates should decrease, all things being equal. This will in turn reduce the US$27 billion figure by a proportion relative to this decrease. For example, the paper takes a 3% settlement fails rate * an ISDA figure of collateral received and delivered of US$889.5 billion. This gets you to US$27 billion. Redo that same calculation with a 1.5% collateral fail rates and you are looking at US$13.4 billion in cost. This is a crap shoot – either you buy into the idea of the same rate of collateral fails or you think they are going lower (or higher, if you are pessimistically inclined), and that gets to your total industry cost figure. We are inclined to see general improvement on the horizon, which would decrease but not eliminate the urgency of the paper’s subliminal argument.
We are more readily willing to accept some other figures in the paper, notably increases in expected FTEs across the buy-side and sell-side due to the increase in collateral movements. This headcount translates into direct costs, which the paper cites at $3.6 million per buy-side firm in 2020 and $2.4 million per sell-side firm in 2020. This makes sense: “Although the sell-side is also likely to witness an operational cost increase for remedying collateral settlement fails due to the uncleared/bilateral margin rules, this increase is expected to be more pronounced for buy-side industry participants.”
This is a good paper all in all. The US$27 billion figure feels like a media-friendly punt, but the close look at FTEs and costs for each buy-side and sell-side firm make better sense. Whatever your opinion on fail rates, a breakdown of assumptions makes it possible to redo the calculations at the top level. We recommend giving this one a read.
The full paper is available at DTCC-Euroclear GlobalCollateral Ltd’s website with free registration.
(Note: our thanks to the authors of the paper for jokingly suggesting not just the title of this review but also a chunk of the content!)