Search My Blog
 About Donald

Don founded Kamakura Corporation in April 1990 and currently serves as its chairman and chief executive officer where he focuses on enterprise wide risk management and modern credit risk technology. His primary financial consulting and research interests involve the practical application of leading edge financial theory to solve critical financial risk management problems. Don was elected to the 50 member RISK Magazine Hall of Fame in 2002 for his work at Kamakura. Read More

 Connect
 Now Available

An Introduction to Derivative Securities, Financial Markets, and Risk ManagementAdvanced Financial Risk Management, 2nd ed.

 Contact Us
Kamakura Corporation
2222 Kalakaua Avenue

Suite 1400
Honolulu HI 96815

Phone: 808.791.9888
Fax: 808.791.9898
info@kamakuraco.com

Americas, Canada
James McKeon
Director of USA Business Solutions
Phone: 215.932.0312

Andrew Zippan
Director, North America (Canada)
Phone: 647.405.0895
 
Asia, Pacific
Clement Ooi
President, Asia Pacific Operations
Phone: +65.6818.6336

Australia, New Zealand
Andrew Cowton
Managing Director
Phone: +61.3.9563.6082

Europe, Middle East, Africa
Jim Moloney
Managing Director, EMEA
Phone: +49.17.33.430.184

Tokyo, Japan
3-6-7 Kita-Aoyama, Level 11
Minato-ku, Tokyo, 107-0061 Japan
Toshio Murate
Phone: +03.5778.7807

Visit Us
Linked In Twitter Seeking Alpha

Careers at Kamakura
Technical Business Consultant – ASPAC
Asia Pacific Region
Business Consultant – ASPAC
Asia Pacific Region

Consultant
Europe

Kamakura Risk Manager Data Expert
Europe, North America, Asia & Australia 

 

 Archive
  

Kamakura Blog

  
Dec 13

Written by: Donald van Deventer
12/13/2011 4:23 AM 

On August 25, 2011, Charles Schwab Bank and Charles Schwab Corporation filed suit alleging 12 banks conspired to manipulate the U.S. Dollar London Interbank Offered Rate (“Libor”) over an extended period of time beginning in 2007.  The firms named in the lawsuit were affiliates of Bank of America, Credit Suisse, JPMorgan Chase, Citigroup, HBOS, HSBC, Barclays, Lloyds, WestLB, UBS, Royal Bank of Scotland, and Deutsche Bank. The complaint contains anecdotal and statistical evidence that Libor was manipulated downward.  This blog uses publically available information and reaches the conclusion that the allegations are highly likely to be correct.

The lawsuit was filed by Lieff Cabraser Heimann and Bernstein as counsel to Charles Schwab.  For a copy of the complaint, use this link:

http://www.lieffcabraser.com/media/pnc/4/media.904.pdf

During the period of the complaint, the U.S. Dollar Libor panel consisted of 16 members.  Libor was determined by eliminating the four highest quotations and the four lowest quotations and averaging the remaining quotations.  The process is managed by the British Bankers Association. Background on the U.S. Dollar and other panels is available at www.bbalibor.com.

Five banks which served on the U.S. Dollar Libor panel between April 26, 2007 and November 9, 2009 were not named in the Charles Schwab complaint:

Rabobank of the Netherlands
Royal Bank of Canada
Norinchukin of Japan
Bank of Tokyo-Mitsubishi UFJ
Societe Generale of France

Societe Generale replaced HBOS on the U.S. Dollar Libor panel on February 9, 2009.

Types of Evidence of Manipulation

There are a number of different classes of information that could shed light on possible manipulation of U.S. dollar Libor:

  1. Internal bank data which confirms that actual bank funding costs were much higher than Libor rates indicated by the issuing bank on that day, or that the bank was unable to issue Eurodollar deposits in U.S. dollars at that maturity on that day.  For large international banks like those named in the Charles Schwab complaint, this information is certain to have been collected and will clearly be an object of the discovery process in the trial.
  2. “Smoking gun” conversations or e-mails by management at one or more banks confirming the existence of manipulation.  Some anecdotal evidence of this type is included in the Charles Schwab complaint
  3. Statistical evidence that indicates that posted Libor indications deviated from other third party measures of Eurodollar funding costs in a significant and persistent way.
  4. Ranking of Libor rate indications that is consistent with collusion among two or more Libor panelists in their Libor postings for that day.

This blog uses Libor indications from each of the 17 banks named above over the period from April 26, 2007 to November 9, 2009 or a shorter period (in the case of HBOS and Societe Generale) if the firm was not a U.S. dollar panel member for the full period.  The focus was on the key 3-month Libor maturity, which is very commonly the basis for interest rate swap contracts.  The data was reported by Bloomberg under the auspices of the British Bankers Association.  The author wishes to thank a follower of dvandeventer on twitter.com for generously providing the data.

Manipulation Strategies

There are two alternative manipulation strategies that could be followed separately or jointly to peg Libor at a lower level than actual funding costs.  The first strategy leaves the ordinal ranking of N manipulators’ Libor indications relatively unchanged, but each of the participants lowers the true Libor indication by m basis points.  This type of manipulation could be engineered with no communication among manipulators.  The second manipulation strategy requires communication.  In this strategy, Libor panel members reveal their intended quotations so that the 13th bidder’s (with the highest bidder being number 1, second highest number 2, and so on) indication is revealed.  The 13th-16th bids, the four lowest, will be thrown out.  With knowledge of this cut off, up to N of the manipulators could revise their indications to the same level as the 13th bid. 

We test for each of these strategies in what follows.

Previous Research on Manipulation

In previous Kamakura blogs (June 7, 2010 and October 25, 2010) we used a comparison of default probabilities to study the difference between reported British Bankers Association Libor rates and the matched maturity cost of Eurodollars reported daily by the Federal Reserve on its H15 statistical release.  We concluded in the June 7 blog;

“This brief analysis shows the same conclusions that we reached in the CDS market: that there is the potential for manipulation of a key financial statistic, and market participants should beware.”

In the October 25 blog, we explained that the risk of manipulation of Libor made a move to the overnight index swap rate structure a safe and essential move for market participants.  The June 7 blog focused on the one month maturity.  In what follows, we focus on the 3 month maturity.

CDS Quotations as an Indicator of Default Risk

In the Charles Schwab complaint, evidence from the credit default swap market is presented in order to measure the possibility that Libor has been manipulated.  There are four reasons why this evidence is less useful than one might initially think.  First, there is a severe maturity mismatch, as we are focused on 3 month Libor and the most traded credit default swap maturity is five years.  Second, there is next to no liquidity among non-dealers in the credit default swap market, as we documented in blogs on May 13, 2010 and June 2, 2010.  Third and most importantly, the credit swap market is controlled by five firms, four of which are named in the Charles Schwab complaint.  According to a report from the Office of the Comptroller of the Currency (“OCC’s Quarterly Report on Bank Trading and Derivatives Activities, Second Quarter 2011“), these firms have the most derivatives outstanding as of June 30, 2011:



Given that JPMorgan, Citibank, Bank of America and HSBC are named in the Charles Schwab complaint, a sophisticated conspiracy would either (a) insure that the relative CDS quotes are consistent with their Libor indications or (b) that any inconsistency could be explained away using information only available to one of the five firms that dominate the credit derivatives market in the United States. A fourth reason for avoiding CDS quotes is also important.  The CDS quote is highly affected by the probability of a government rescue and as such, it is an impure indicator of the default risk of banks that are “too big to fail.” Kamakura’s blog of September 29, 2009 shows how the probability of failure for FNMA and Citigroup diverged from CDS spreads once it became obvious that both firms would be bailed out by the U.S. government.  Almost all of the U.S. and European firms on the Libor panel during the period of interest were beneficiaries of government rescue funds as documented in the Kamakura “Case Studies in Liquidity Risk” blogs listed below.

For those reasons, we use other data in what follows.

Measuring Deviations from Third Party Measures of 3 Month Eurodollar Costs

In this section, we compare the 3 month U.S. Dollar Libor indications of the 17 firms that served on the 16 member U.S. Dollar Libor panel from April. 26, 2007 to November 9, 2009 with a third party measure of 3 month Eurodollar costs.  The Federal Reserve reports the cost of 3 month Eurodollars daily on its H15 statistical release.  That release identifies the source as the broker ICAP via Bloomberg.  No other description of the process of data collection is made available by the Federal Reserve or by ICAP on its website.  We assume, but cannot confirm, that the rates reported represent an average of actual transactions via ICAP as a Eurodollar deposit broker.  This contrasts with the Libor postings, which are indications, not traded rates. 

The graph below shows the difference between 3 month Libor rates reported by Citibank between April 26, 2007 and November 9, 2009 and the H15 3 month Eurodollar cost reported by the Federal Reserve. A positive difference means that the Citibank Libor indication was higher than the Federal Reserve Rate and a negative difference means that the Citibank Libor indication was lower than the Federal Reserve rate.



The Citibank Libor quotation first turned negative on a regular basis beginning August 8, 2007. Beginning with the collapse of Lehman Brothers on Sunday, September 14, 2008, the Citibank Libor indications were very negative compared to the rate reported by the Federal Reserve, with the difference exceeding 200 basis points within a month after Lehman’s collapse.  In the chart below, we compare the difference versus the Fed H15 quotation for all 17 banks and rank them for the period from September 15, 2008 to November 9, 2009.  We also note which banks were named in the Charles Schwab complaint, which banks were major CDS dealers, and which banks were among the top borrowers from the Federal Reserve in a long series of Kamakura blogs entitled “Case Studies in Liquidity Risk” listed below.



HBOS shows the largest “underpricing” of Libor from the period from September 15, 2008 to November 9, 2009, but only because it was Libor panel only until February 9, 2009, the period when negative spreads were highest.  Note that 8 of the 9 firms with the biggest underpricing to the Federal Reserve’s H15 measure of 3 month Eurodollar costs were named in the Charles Schwab complaint.  Appendices 1 through 17 detail the daily spreads over 3 month U.S. Treasury yields (also from the H15 release) for each bank’s Libor indication and for the Fed’s H15 3 month Eurodollar funding cost measure.  Other than the firms which were not publicly listed (Rabobank, WestLB, Norinchukin), each graph has an overlay of the probability of failure as measured by the Kamakura Risk Information Services 3 month Jarrow-Chava version 5 default probability.  For more on the KRIS default probability service, see www.kamakuraco.com.

We conclude that all 17 of the firms showed significant and persistent underpricing of Libor versus the H15 Eurodollar quotations reported by the Federal Reserve. We now measure the likelihood of Libor manipulation by changing the ordinal ranking of Libor quotations by “bunching” of quotations at the 13th highest level.

Manipulating the Ordinal Ranking of Libor Quotations

As noted above, the calculation of Libor during this period employed a 16 member panel where the four highest and four lowest quotations were eliminated.  Libor in effect is determined by the remaining eight quotations.  Those remaining banks, if they were aware of the four highest and lowest quotes, could manipulate Libor as follows.  If they wanted to lower Libor (as the Schwab complaint asserts), they would move their indication to match the 13th highest level.  If they wanted to raise Libor, they would move their quote to the 4th highest level.  We examine the number of times that there was more than one indication at the 4th highest Libor rate and the 13th highest Libor rate, and we measure the probability that each of the individual banks fell into one of these two groups.

We note that this is a clumsy strategy because it’s so easily detected.  A more subtle strategy is the one measured above, where N manipulators all lower their quotes without bunching.  Another problem with detecting a bunching at the 4th highest rate or 13th highest rate is the case where the 16 panel members indicated only 2, 3, or 4 different rates, rather than 16 distinctly different rates.  If the 16 members indicated that 3 month Libor were either 5.35% or 5.36%, random chance alone could create bunching at the 4th highest or 13th highest rates.  More than 94% of the Libor indications from April 26, 2007 to November 9, 2009 were rounded to the nearest basis point.  About two-thirds of the remaining quotes were rounded to the nearest half basis point.  The graph below shows the differences between the highest and lowest quotations on a daily basis for the period in question, an indirect measure of the maximum number (up to 16) of the number of quotes that are possible:



We focus on suspicion of downside manipulation, focusing on the period from September 2008 onward.  Any quotation for which the bank’s quotation matches the 13th highest quotation on the panel is labeled “suspicious” because it indicates bunching.  The results are as follows:



HSBC had 86 suspicious quotations out of 421 total observations, a ratio of 20.4%.  By contrast, Norinchukin had no observations out of 421 in which its indications matched the 13th highest quote.  As noted above, this could come about by random chance.  We measure the standard deviation of the measured “suspicion” ratio and find that 8 of the 17 banks had a statistically significant (more than 2 standard deviations) difference from the null hypothesis that there was only a 1/15 chance of matching the quote of one of the other 15 panel members.  The null hypothesis here is not easy to articulate, because on a day where there were only two rates indicated by the 16 panel members, the null hypothesis probability would be substantially different.

Our conclusion from this data is that the “bunching” phenomenon is more difficult to establish with a high degree of confidence.  Because of its crudeness and easy detection, an institution would be foolish to follow such a manipulation strategy, so it is not surprising to find the results above.

Conclusions

Appendices 1 through 17 show convincingly that most of the Libor panel members persistently indicated Libor levels well below the H15 3 month Eurodollar rate.  This was particularly true from September 2008 onward, when (ironically) the failure probabilities of the firms were rising sharply and governments in the U.S. and Europe moved to bail out these banks, presumably because they were unable to fund themselves completely.

We conclude that it is highly likely that the first two categories of data (bank funding costs and “smoking gun” internal communications) will confirm that the banks tacitly manipulated Libor downward by collective lowering of quotations, rather than consistently moving to cluster at the 13th highest rate on that day.

References

van Deventer, Donald R.  “HBOS: A Case Study in Risk Management Failure,” Kamakura blog, www.kamakuraco.com.  March 10, 2009.

van Deventer, Donald R. “CDS Spreads and Default Probabilities: What is the Linkage?” Kamakura blog, www.kamakuraco.com, March 18, 2009.

Robert A. Jarrow, Li Li, Mark Mesler, and Donald R. van Deventer, “The Determinants of Corporate Credit Spreads: An Update,” Kamakura blog, www.kamakuraco.com, September 23, 2009.  Redistributed on www.riskcenter.com on September 24, 2009.

van Deventer, Donald R. “Comparing Default Probabilities and Credit Default Swap Quotes: Insights from the Examples of FNMA and Citigroup,” Kamakura blog, www.kamakuraco.com, September 29, 2009.  Redistributed on www.riskcenter.com on September 30, 2009.

van Deventer, Donald R. “Corporate Credit Default Swaps and Non-Dealer Trading Volume,” Kamakura blog, www.kamakuraco.com, May 13, 2010.  Redistributed on www.riskcenter.com on May 16, 2010.

van Deventer, Donald R. “Kamakura Blog: Daily Trading Volumes in Credit Default Swaps,” Kamakura blog, www.kamakuraco.com, June 2, 2010.  Redistributed on www.riskcenter.com on June 3, 2010.

van Deventer, Donald R. “Kamakura Blog: Default Probabilities and Libor,” Kamakura blog, www.kamakuraco.com, June 7, 2010. Redistributed on www.riskcenter.com on June 8, 2010.

van Deventer, Donald R. “The Links between CDS Spreads and Default Probabilities,” Kamakura blog, www.kamakuraco.com, July 14, 2010.  Redistributed on www.riskcenter.com on July 19, 2010. Reprinted in Japanese in Market Solutions Review, August 2010.

van Deventer, Donald R. “Overnight Index Swaps versus Libor Swaps,” Kamakura blog, www.kamakuraco.com, October 5, 2010.  Redistributed on www.riskcenter.com on October 6, 2010.

van Deventer, Donald R., “Institutions that Used Primary, Secondary or Other Extensions of Credit From the Federal Reserve, February 8, 2008 to March 16, 2009,” Kamakura blog, www.kamakuraco.com,  May 5, 2011.

van Deventer, Donald R.,” Top 100 Institutions Ranked by Number of Transactions in Primary, Secondary or Other Extensions of Credit from the Federal Reserve,” Kamakura blog, www.kamakuraco.com, May 9, 2011

van Deventer, Donald R. “Top 100 Transactions in Primary, Secondary or Other Extensions of Credit from the Federal Reserve, February 8, 2008 to March 16, 2009,” Kamakura blog, www.kamakuraco.com, May 10, 2011

van Deventer, Donald R. “A Credit Crisis Chronology, Part 1 Through February 2008: This Time Isn’t Different,” Kamakura blog, www.kamakuraco.com, May 13, 2011. Redistributed on www.riskcenter.com, May 16, 2011.

van Deventer, Donald R., ”Case Studies in Liquidity Risk: AIG,” Kamakura blog, www.kamakuraco.com, May 16, 2011

van Deventer, Donald R., “Case Studies in Liquidity Risk: Bank of America,” Kamakura blog, www.kamakuraco.com, May 17, 2011

van Deventer, Donald R., ”Case Studies in Liquidity Risk: Countrywide Financial,” Kamakura blog, www.kamakuraco.com, May 18, 2011

van Deventer, Donald R., ”Case Studies in Liquidity Risk: Merrill Lynch,” Kamakura blog, www.kamakuraco.com, May 20, 2011

van Deventer, Donald R., “Case Studies in Liquidity Risk: Consolidated Bank of America, Countrywide and Merrill Lynch,” Kamakura blog, www.kamakuraco.com, May 25, 2011. Reprinted in Bank Asset and Liability Management Newsletter, November 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Lehman Brothers (Dick Fuld was Right, Updated June 3, 2001),” Kamakura blog, www.kamakuraco.com, May 31, 2011

van Deventer, Donald R., “Case Studies in Liquidity Risk: Morgan Stanley (Updated June 3, 2011),”  Kamakura blog, www.kamakuraco.com, June 1, 2011

van Deventer, Donald R. “Case Studies in Liquidity Risk: Citigroup,” Kamakura blog, www.kamakuraco.com, June 6, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Dexia Credit Local New York Branch,” Kamakura blog, www.kamakuraco.com, June 7, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Depfa Bank PLC New York Branch,” Kamakura blog, www.kamakuraco.com, June 9, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Barclays,“ Kamakura blog, www.kamakuraco.com, June 14, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Goldman Sachs,“ Kamakura blog, www.kamakuraco.com, July 7, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: JPMorgan Chase, Bear Stearns and Washington Mutual,“ Kamakura blog, www.kamakuraco.com, July 8, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Wachovia,“ Kamakura blog, www.kamakuraco.com, July 12, 2011.

van Deventer, Donald R. and Stephanie Yasunaga, “A Credit Crisis Chronology, Part 2 from March 2008 Through March 2009: This Time Isn’t Different,” Kamakura blog, www.kamakuraco.com, posted July 13, 2011 and dated May 14, 2008 to match part 1 on May 13, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: State Street,” Kamakura blog, www.kamakuraco.com, July 20, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Bank of New York Mellon,” Kamakura blog, www.kamakuraco.com, July 21, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: HSH Nordbank AG New York Branch,” Kamakura blog, www.kamakuraco.com, July 25, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Societe Generale SA New York Branch,” Kamakura blog, www.kamakuraco.com, July 27, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Bank of Scotland PLC New York Branch,” Kamakura blog, www.kamakuraco.com, August 1, 2011.

van Deventer, Donald R. “Case Studies in Liquidity Risk: Royal Bank of Scotland PLC New York Branch,” Kamakura blog, www.kamakuraco.com, August 4, 2011.

van Deventer, Donald R. “How Joe and Mary Six Pack Saved Wall Street, London, Frankfurt and Big Corporates in the USA and Europe,” Kamakura blog, www.kamakuraco.com, August 23, 2011.

van Deventer, Donald R. “In Which the Fed Uses $348.5 Billion of Funds from Joe and Mary Six Pack and Lends More Than Half of It to European Firms,” Kamakura blog, www.kamakuraco.com, August 25, 2011.

van Deventer, Donald R. “Ranking of the 82 Borrowers under the Federal Reserve's Commercial Paper Funding Facility,” Kamakura blog, www.kamakuraco.com, August 29, 2011.

Appendices are given below.



Donald R. van Deventer
Kamakura Corporation
Honolulu, December 12, 2011


Appendix 1



Appendix 2



Appendix 3



Appendix 4



Appendix 5



Appendix 6



Appendix 7



Appendix 8



Appendix 9



Appendix 10



Appendix 11



Appendix 12



Appendix 13



Appendix 14



Appendix 15



Appendix 16



Appendix 17



Tags: