Here is a wonderful sentence:
A key insight from the enhanced BIS credit derivatives data is that non-rated multi-name credit risk sourced from multiple sectors has been transferred from derivatives dealers to IFGCs, SPVs and OFCs.
Yeah! Wh … what?
It’s from the quarterly review of the Bank for International Settlements, which is a delightful hodgepodge of hard-to-read charts, hard-to-read sentences, and general oblique glances at the guts of the global financial system. It is both glancing and gutsy. There are reams of tables. Give it a read.
The quote above is about this:
Everything clear now?
I love charts and all but I mostly think in stories, and I’m trying to parse together the story for these facts because they seem somehow important. It seems to go like this.
These charts show the movement of credit exposure via credit derivatives. Various big banks and dealers report their derivatives exposure to their central banks, which report it to the BIS. There’s some $32 trillion of gross CDS notional outstanding, but obviously dealers take mostly-offsetting positions; these bars represent the gross long exposure in various categories minus the gross short exposure in those categories. A positive number means that the dealers are long protection / short credit; a negative number means they’ve sold protection / are long credit. The three charts are three ways of slicing it; the left-hand one, “by sector,” seems the most interesting to me.
So what are those bars?
BSDs – banks and securities dealers. Despite the acronym, these are probably mostly smaller and less important banks, that is, the non-reporting dealers, meaning in the U.S. banks and broker-dealers that are not Fed primary dealers. The reporting dealers are not listed here because their net exposure to each other should be zero.
IFGCs – insurance and financial guarantee companies, and (for some reason) pension funds
SPVs – special purpose vehicles – presumably, securitizations
OFCs – other financial companies; per the report, “Other managed funds, such as money market mutual funds, are well represented in the OFC category”
NFIs – non financial institutions, meaning I guess corporates
HFs – hedge funds
The most interesting slices of the bars are the forest-greeny-bluey ones for “multiple sectors.” In a footnote, the BIS explains that these are probably mostly synthetic CDOs and CDS index and basket products:
Information is not available on positions in CDS that are both non-rated and reference multiple sectors, but supplementary BIS data do show that the majority of non-rated risk transfers to IFGCs, SPVs and OFCs occurred through multi-name CDS. Multi-name CDS that are likely to reference multiple sectors and be classified as non-rated include basket CDS, synthetic collateralised debt obligations (CDOs) and CDS index tranches. Where multi-name CDS did not have a rating, reporting dealers were asked to allocate these instruments to a rating bucket on the basis of the credit quality of the underlying debt, unless this was “not possible or very burdensome”. The products listed above would probably fit this description.
In short, the reporting dealers seem to own (eyeballing it here) $1 trillion of credit protection from a fairly even mix of smaller banks, insurers, SPVs, and other financial companies, plus a small slug of non-financials. And they’ve sold about one-third of that amount to hedge funds.*
This seems to abstract this into two big flows of money:
1. The bigger flow is: Bank lends money to a person or business to buy a home or build a machine. Bank then offloads credit risk to a series of acronyms – IFGCs, SPVs, OFCs, etc. – that are loosely speaking in the shadow banking system. Bank does the underwriting in the individual names but sells the risk mostly in vast undifferentiated pools of stuff so that the shadow banks don’t have to do too much credit work.
On the one hand, great. Banks get more balance sheet to lend, and systemically important banks are less likely to go bankrupt due to bad credits – they reduce their credit risk except in a turtle-y, what-if-the-shadow-banks-go-bankrupt way, which the BIS notes. (“The new data also show that BSDs and SPVs had sold on a net basis credit protection on financial debt. The risk of simultaneous default of protection sellers and reference entities is often higher when these institutions come from a common sector, rather than different sectors.”) Banks do credit work by lending to single names; shadow banks are essentially credit indexers, willing to take on broad classes of credit risk rather than individual names – in other words, banks provide the lending expertise, shadow banks, with apparently longer time horizons, supply the risk appetite. This seems like a decent division of labor.
On the other hand, hmm, transferring risk from people supposed to underwrite individual credit risks to index players who buy big pools of stuff without examining the underlying credits – haven’t we heard about that before? The BIS quote above starts some general nervousness on the part of the otherwise fairly unruffled BIS:
A key insight from the enhanced BIS credit derivatives data is that non-rated multi-name credit risk sourced from multiple sectors has been transferred from derivatives dealers to IFGCs, SPVs and OFCs. Such risk transfers are likely to have been generated by basket CDS, synthetic CDOs or CDS index tranches. These types of CDS can be difficult to value and have experienced significant price jumps in the past. To the extent that such risks remain, they appear to have been passed on from the banking sector to parts of the non-bank financial sector often known as shadow banks.
2. Same as above, except instead of sourcing credit risk by lending money to people to do risky things, the bank sources credit risk by writing CDS to hedge funds to give them short credit exposure to the things they want short credit exposure to:
Accordingly, in the past year or so hedge funds have been using CDSs to express increasingly bearish views on debt markets, which the report indicates primarily relate to sovereign and (non-financial) corporate debt. And because these graphs show that dealers are net long in their CDS positions, hedge funds’ CDS shorts are probably ultimately being absorbed by the other categories of counterparties.
This one is more opaque and the BIS statistics don’t tell a complete story. Maybe all those hedge funds are buying CDS to hedge their own real-economy contributions: maybe they’re supporting Greece by buying Greek bonds but hedging using CDS; maybe they’re lending to companies and proxy hedging with correlated CDS. Or maybe they’re just making zero-sum directional bets against credits that they don’t like. Maybe they’re doing the credit work to find the worst sovereigns and companies, and using dealers to facilitate their short bets against those entities – with the other side taken, again, by indexing schlubs who aren’t interested in doing credit work. We’ve heard that one before, too.
* That’s about one-third of notional (again, from eyeballing). It’s probably less than that in terms of actual risk exposure, measured by DV01 or market value or whatever, because for instance the hedge fund short exposures tend to have shorter tenors than the non-HF long exposures. This seems to be true based on the BIS’s chart of market values: