Fitch Ratings’ report on repo and shadow banking from February 3 has just been posted on the Fitch Ratings website — available to all who sign up for a login.
While its key themes were covered extensively by our FT colleagues back in February — namely that the use of lower-rated debt as collateral had returned to pre-crisis levels — one table buried deep in the report did catch our eye:
It seems, the increased use of lower-rated debt between the second half of 2010 and August 2011 came at the amazing cost of Treasury collateral.
Which begs the question, where did all that Treasury collateral get re-directed to? If anywhere?
Update 14:52 – Some of our commenters have wondered why there is a discrepancy between Fitch’s data and the repo data collected by the New York Fed.
According to the authors of the Fitch report their analysis is focused on a sample of the public disclosures of the 10 largest US prime money funds, rather than the market as a whole.
The dataset is supposed to complement the FRBNY data, rather than go against it.
We agree with the authors that the data subset actually provides an interesting view of developments in the repo market at that time. Readers might recall that the summer of 2011 was plagued by an extremely volatile US Treasury repo market, in part fueled by the political standoff over the debt ceiling and in part due to the negative/low repo rate regime which came before it.
It also coincided with rocketing US M2 money supply — almost as if money makret funds didn’t quite know what to do with the money.
Here in any case is a further explanation from Fitch’s Robert Grossman:
The unique advantages of focusing on prime money fund disclosures are that they:
(1) provide the most detailed publicly available historical information on repo haircuts, pricing, collateral, and counterparties;
(2) enable construction of a time series of repo attributes , allowing us to capture trends before, during, and after the US credit crisis;
(3) with the recent SEC Form N-MFP, provide unprecedented granularity into security-level details, which enabled us to determine that at end-August 2011, the median value of the structured finance collateral was roughly 43 cents on the dollar and that about 50% of the structured finance collateral pool consisted of legacy subprime and Alt-A RMBS and CDOs; and
(4) isolate and focus on trends in prime money funds, which play a prominent role in tri-party markets, particularly as lenders against “non-traditional” collateral (e.g., corporate debt, equities, and structured finance).
On this last point, our estimates of the share of government securities collateral will necessarily be below the FRBNY’s, since our study excludes money funds that transact exclusively in government securities.
As to what led to that sudden decrease in UST use, here are Grossman’s thoughts:
The question posed in your piece is an interesting one. A possible explanation is repo pricing, particularly given the low yield environment facing the money fund sector. While repo pricing is beyond the scope of the FRBNY data, our study indicates that, as of end-August 2011, repos backed by structured finance collateral yielded more than 50 bp on average, while repos backed by Treasurys yielded roughly 5 bp.
Interestingly, repos backed by Agency securities yielded roughly 15 bp on average, which might also explain why there appeared to be a reallocation from Treasury to Agency securities relative to our prior observation period. It will be interesting to monitor this trend going forward, as money funds face a possible tension between ongoing risk aversion versus the potentially lower yield of repos backed by Treasurys.
For example, in a study we recently published on money fund exposure trends, roughly 90% of repos with French banks were backed by Treasury and agency collateral (as of end-January 2012).
That last point is very interesting indeed we *think*.