Written Testimony of Peter Weiland, JPMorgan's former Head of Market Risk at the CIO for the Senate Permanent Subcommittee on Investigations March 15, 2013
'Chairman Levin, Ranking Member McCain, and Members of the Subcommittee:
My name is Peter Weiland. I was employed at J.P. Morgan as its Head of Market Risk in J.P. Morgan’s Chief Investment Office (“CIO”) from October 2008 until October 2012. I am here in response to the invitation I received from the Subcommittee, and I will attempt to provide information and answer questions to the best of my recollection and knowledge.
I am proud of the CIO and its accomplishments. It was a hard working team that took its job seriously. The events of 2012 were a substantial disappointment but I am here today to help explain the facts surrounding those events.
In your invitation to me, you asked for information on several different topics. I will try to respond to those questions as directly and as succinctly as I can.
You asked me about the risk management structure in the CIO and my role in managing risks, including within the London office. The risk management structure changed over the four years I was there. I was hired in October, 2008, by Ina Drew to be the Head of Market Risk. There was no Chief Risk Officer for the CIO at that time. When I was hired, I had a direct reporting line to Ms. Drew with an indirect line to J.P. Morgan’s Chief Risk Officer (“CRO”) Barry Zubrow. In mid-2009, my official reporting line changed to a direct line to Mr. Zubrow with an indirect line to Ms. Drew.
In September, 2011, a search was announced to bring in a Chief Risk Officer for the CIO to whom I would report. In January, 2012, Irv Goldman was selected to be the CRO for the CIO. Upon his hiring, I reported directly to Mr. Goldman.
The second question posed to me by the Subcommittee was to describe the risk management practices at the CIO, including the risk metrics and the limits that were used, who set the limits, and how the breaches were handled. During most of the period in question, the CIO’s activity was divided into SAA (strategic, longer term investment activity mostly in AFS accounting) and TAA (tactical, shorter term activity, mostly mark-to-market accounting). The SAA risk framework was generally not based on delegated trading limits; the SAA Committee specifically approved portfolio activity.
The TAA was treated with a traditional delegated limit framework, including limits on Value-at-Risk (“VaR”), stress, and non-statisticals. Synthetic credit, although intended to offset strategic credit exposures, was included in the TAA because accounting rules require derivatives to be mark-to-market. Therefore, CIO market risk limits largely applied only to TAA activity with some exceptions, including individual issuer limits in SAA (“single name limits”). In late 2011 and early 2012, it was agreed to include the SAA in firm-wide stress limit in the firm-wide stress limit framework. This was implemented during the first quarter of 2012.
Limit approvals were to be made at various levels; limits were agreed by the nature of the business and the Risk. Top (Level 1) CIO limits were approved by the CEO, the CRO, and the CIO. Within the CIO, Level 2, limits were approved by Ms. Drew, regional CIOs, and myself. Some regional limits were approved by regional CIOs and myself. Breaches of the limits were handled according to firm-wide Market Risk policy; the designated approvers were notified when the breaches occurred.
The third question that the Subcommittee posed was the CIO’s efforts to develop alterative risk and capital models during 2011 and 2012, in particular to recalculate the Value-at-Risk, the Comprehensive Risk Measure (“CRM”), and the Risk Weighted Assets (“RWA”) results, and my role in those efforts.
In early 2011, I discussed expected Basel 2.5 risk requirements with central Risk QR staff. Risk QR was developing a CRM model to encompass all J.P. Morgan synthetic credit activity. In addition, it became clear that the VaR for CIO synthetic credit would need to be upgraded. The target date for this upgrade was to be the end of 2011. In this regard, several historical factors are relevant:
• In 2009, the CIO synthetic credit VaR was initially modeled after the Investment Bank (“IB”) for consistency;
• the IB upgraded its VaR with a heavy dependence on individual credit curves in 2010. Because CIO activity is purely index based, it was clear that the IB’s individual name approach would be inappropriate for the CIO;
• in 2011, the Market Risk Technology group developed a VaR tool that was expected to take data feeds from both the IB and the CIO. In August, 2011, the initial feeds from CIO failed and it was determined that the VaR tool did not work without IB transformation tools;
• in September, 2011, the CIO office began work on a new VaR model that was intended to meet all the requirements of Basel 2.5. The development of this new VaR model was led by Pat Hagan of the CIO, with close monitoring by the Risk QR to ensure fulfillments of all regulatory requirements;
• concurrently, the Risk QR was developing a CRM model, another contributor to RWA for synthetic credit (RWA = VaR + stress VaR + CRM). The CIO business quantitative team had questions about the details about the central CRM model development. The European CIO team requested approval to develop a CRM model specifically for CIO, but their request was denied;
• the CIO business quant team developed a parallel model CRM to try to better understand the central CRM model;
• ultimately, the CIO synthetic credit VaR was approved by the Risk QR and then implemented at the end of January, 2012. Subsequently, it was determined that coding errors resulted in an incorrect VaR calculation and the VaR was rolled back to its previous model;
• my role throughout the process was to maintain focus to ensure that the new VaR was implemented so that the firm would meet regulatory deadlines as well as maintain communication with the Risk QR regarding CRM model development and testing.
The fourth question that you submitted asked for information about my role in analyzing the risks associated with the Synthetic Credit Portfolio (SCP) and the SCP’s record of risk limit breaches and actions taken in response to those breaches in 2011 and 2012. As always, the answers to SCP questions are complicated. But I was responsible for all Market Risk within the CIO, including synthetic credit, during my tenure as the Head of Market Risk. This included a wide range of activities, including:
• the AFS securities portfolio consisting of a wide variety of asset classes including MBS, CMBS, CLOs, student loan ABS, credit card ABS, uni-bonds and so forth. This was a portfolio of approximately $400 billion;
• Mortgage Servicing Rights hedging activity, including very large interest rates, volatility, and prepayment risks;
• structural Foreign Exchange (FX) hedging activity;
• tactical positioning in New York, London, and Hong Kong, primarily in rates and FX and also including synthetic credit;
• CIO private equity activity.
Supporting me in London was one senior risk manager for whom the synthetic credit portfolio was part of his responsibilities. There were also two junior resources dedicated to the synthetic credit portfolio.
The main limits on CIO synthetic credit in 2012 were VaR, stress, csbpv, and csw10%. The initial breaches were csbpv breaches. The csbpv values for all synthetic credit positions were added to arrive at the usage. During early 2012, as investment grade (“IG”) positions began to be used in size to offset high yield (HY) positions, the total csbpv grew because risk equivalent IG and HY positions had a very different csbpv value. Given this change in the business mix, it was determined that the csbpv was no longer a useful limit.
VaR limit breaches also occurred during January, 2012, sometimes resulting in a firmwide VaR limit breach. Given that the CIO VaR was known to consistently overstate P/L volatility and because it was believed that an improved model was near implementation, the new VaR model output was an important consideration in the approval of one-off increases and gave management comfort that the VaR breaches would not have occurred if a new model were already in place. However, because of further breaches of the csw10% limits and stress limits that were occurring at the end of March and April, 2012, there was increased attention to the portfolio from senior management due to the RWA, the losses, and the press.
The fifth question you asked was an inquiry as to what actions were taken to inform senior bank managers about the SCP risk issues. Let me first note that the size of the synthetic credit portfolio was well understood among the senior management at J.P. Morgan. The portfolio dominated CIO VaR for most of the period from 2008 to 2012, and it had a significant impact on stress test results as well. Because the track record for managing such a significant risk had been good, namely there were significant gains in 2009 and relatively stable results during the rest of the period, matters proceeded in the normal course. There were active discussions on regulatory capital usage.
The final question you asked me related to what actions were taken to inform regulators about SCP issues. During the period in question, I had occasional meetings with regulators with regard with to the CCAR regulatory stress exercises. In addition, the CIO management team met with regulators in April 2012 to discuss the SCP events. I also recall that there were some additional meetings about synthetic credit following up on the meeting with regulators in April, 2012.
I hope the above answers are helpful to your inquiry. Because this subject area is so complicated, it is hard to give succinct answers. I am happy to answer any questions about the above topics.
image: © brydeb