When people buy a used car in the United States, they often consult prices paid by thousands of others for the same model car that they are considering. Kelly Blue Book is a common source for this kind of information. Another alternative that is much less informative is the prices that have been paid on one used car dealer’s lot. When it comes to the recent near-hysteria in the credit default swap market about Greece and other sovereigns, we need to analyze the volume and sources of information we’re getting on credit default swap quotations. Is the sovereign CDS market the equivalent of the Kelly Blue Book for used cars, with thousands of quotations? Or is it just a few quotes from the equivalent of one used car lot? That is the subject of today’s blog.
As an avid twitter user (www.twitter.com/dvandeventer), I’ve been stunned at how much digging my fellow financial twitterati have done into the nitty gritty data behind the headlines. Watching credit default swaps for Greece rise and fall by hundreds of basis points in an hour or less has been a real education. The blogger who goes by the pseudonym “Tyler Durden” on www.zerohedge.com was the first to call my attention to the data behind the numbers. In today’s post we analyze data on credit default swaps reported by the Depository Trust and Clearing Corporation on www.dtcc.com for the week of April 26-30, 2010. Our purpose is to get some feel for the depth of trading in sovereign credit default swaps.
Our first table, taken from www.dtcc.com, shows a huge $14 trillion volume (notional principal, US dollar basis) in all credit default swap contracts during the week.
More than 2.1 million credit default swap contracts were traded. Of greater interest, however, is the fact that almost 1.8 million of these contracts were traded between dealers, 83.7% of the total number of trades. Only 16.3% of the trades, 351 thousand, were bought or sold by non-dealers.
Who are the dealers in credit default swaps? There are many sources of this information. For purposes of this blog I used the quarterly report of derivatives outstanding of the Office of the Comptroller of the United States for the fourth quarter of 2008, which is available at this link: http://www.occ.treas.gov/ftp/release/2009-34a.pdf
This is of course a U.S. focused report, but DTCC is a U.S. focused organization as well. The OCC report lists this volume of credit derivatives outstanding by notional principal:
Positions at JP Morgan Chase represented 52.8% of the total, and the top 6 organizations (6 after recognizing the merger of Wells Fargo and Wachovia) accounted for 99.6% of the total outstandings in credit default swaps. We need to keep this in mind when we examine the DTCC data on trading in sovereign default swaps.
The raw data from DTCC reports on approximately 1,000 reference names in which there were at least 50 new trades or assignments of existing credit default swap contracts on all entities traded during the week of April 26-30, 2010. Only twenty of these entities were sovereigns, and DTCC reported this volume of new credit default swap contracts for the week of April 26-30, 2010 (we ignore assignments for the rest of this blog post):
Note that there were only 49 CDS contracts traded per day for Greece (“Hellenic Republic”) and a miniscule 7 contracts traded per day for the Kingdom of Belgium. Even those numbers are inflated volume estimates in the same way that used cars traded between auto dealers are not representative of the true demand by real end users, retail buyers of automobiles. How much was the true demand by real end users of credit default swaps on sovereigns? We can use the estimate from the full week of trading of CDS contracts reported by DTCC, in which 83.7% of all trading was dealer to dealer and 16.3% was by non-dealers, our true “end user” of the credit default swaps for sovereigns. We also apply the average notional amount on all CDS trades by non-dealers of $6.9 million per trade in a modified version of the table above:
The implications of this table are stunning. If the proportions of non-dealer trades for all CDS apply to sovereign CDS as well, we can see that the trading in the CDS of Greece by true non-dealer end users amounts to only 7.9 trades per day. For the Argentine Republic, there is the equivalent of only 1.4 trades per day. For Greece, the approximate notional principal involved is just under $55 million. For the Argentine Republic, an average day’s volume by non-dealers is less than $10 million. For the top 20 sovereigns traded during the week, there were only an estimated 4.4 trades per day by non-dealers.
This miniscule volume by non-dealers is particularly stunning in light of the huge publicity on movements in CDS quotes on sovereigns who are having their 15 minutes in the spotlight.
But isn’t the 83.7% of trading by dealers “real”? If there are only 6 firms in the United States controlling the credit derivatives market and slightly more than that internationally, how can that be “real”? Since CDS positions are a zero sum game, all dealers combined can only be net long what non-dealers are net short, and vice versa. Just as in the used car analogy, we care only about the trading by retail car buyers, not the movement of volume between dealers, who still need to sell whatever cars they buy en masse. Moreover, we want as many price points from retail car buyers as we can get, so we want the www.kbb.com Kelly Blue Book prices, not the prices from one dealer’s lot.
So what are we to make of the constant moves in the CDS quotes of Greece? We know that 83.7% of the real trades are just the “big six” playing among themselves. Even more important, we can imagine that much of the “movement” in Greece CDS is the equivalent of one car dealer moving the price on one used car during the day, even though no one bought the car. If there are two traders of Greece CDS in each dealer firm, that means there are just 12 people moving the price around in the dealer community, the same as a small group of car dealers moving the price on one car model with no sales for the day.
The sad conclusion is that the sovereign CDS market, based on DTCC data, is no more important in terms of daily trades by real end users than the number of used car sales by a mid-sized car dealer. In fact, at 4.4 end-user trades per day in the top 20 sovereign names, there is LESS volume in sovereign CDS per day than one average used car dealer would generate. The accuracy of “sales prices” and “offered prices” on one car lot pales in comparison to the thousands of data points represented by the Kelly Blue Book summary. Sadly, we have no Kelly Blue Book for sovereign CDS because there’s only one car dealer, the one owned by the six-plus market makers.
That’s a very good reason why observers of sovereign CDS spreads should keep both hands on their wallets.
Donald R. van Deventer
May 11, 2010