JPM and JPM Bank

In 2012, JPM had assets with a book value worth $2.4 trillion, making it the largest financial holding company in the United States, as well as the world’s largest derivatives dealer and credit derivatives trader. JPM’s main bank subsidiary, JPM Bank, was also the largest bank in the United States, having branches in twenty-three states. With historic roots in the nineteenth century, JPM started as a commercial bank and then evolved into the large and highly diversified company it is today. During the twentieth century, JPM’s growth was spirited by passage of the 1999 Gramm-Leach-Bliley Act, repealing the 1933 Glass-Stegall Act and permitting financial institutions to engage, once again, in both commercial and investment banking activities. JPM responded quickly and assertively by (1) merging with Chase Manhattan Bank (2000) and Bank One (2004); (2) acquiring Bear Stearns (2008), an investment bank; and (3) purchasing most of the banking operations of Washington Mutual (2008), a major thrift institution.

JPM (parent) had six lines of businesses, (1) Retail Financial Services; (2) Credit Card Services and Automobile Loans; (3) Commercial Banking; (4) Investment Banking; (5) Treasury and Securities Services; and (6) Asset Management, but none of them caused its $6.2 billion credit derivatives losses. Rather, the SCP caused them (see “The SCP” section, below).

The CIO

Prior to 2005, JPM Bank’s internal treasury department managed the risks of both its long-term and short-term assets. This changed in 2005 when the Treasury and CIO Office (CIO) was spun off as a separate group, with the intended purpose of focusing on JPM Bank’s long-term asset management. Short-term asset management remained the responsibility of the bank’s internal treasury department. Led by banking veteran Ina Drew, the CIO had a staff of about 425 people, including 140 traders located mainly in New York and London. Drew had direct reporting access to top management echelons of both JPM Bank and JPM Parent. Because the CIO was part of JPM Bank, its main regulator was the OCC, but oversight was loose because the CIO had no client-facing activities.

The CIO had two major responsibilities, optimizing the bank’s “excess deposits” and managing JPM Bank’s risks.

From the beginning, the CIO’s mandate was nebulous and became even more so as it morphed over time, largely due to the Great Recession (December 2007–June 2009), which reorganized many of JPM Bank’s priorities. Threatened by the potential failure of both the U.S. and global financial systems, customer funds flowed massively into safe bank deposits. This flight to safety, combined with JPM’s acquisition of Bear Stearns and Washington Mutual Bank, greatly expanded JPM Bank’s customer deposit base. In 2008, alone, deposits rose by approximately $100 billion, and by 2012, the CIO found itself managing an investment portfolio worth approximately $350 billion, which would have ranked this office (alone) as the seventh largest bank in the United States.

The SCP

Because the CIO’s responsibilities included credit risk management, in 2006, the CIO’s International Chief Investment Officer, Achilles Macris, sought and won internal approval for a new business initiative, focusing on “credit trading.” Soon thereafter, the CIO had credit derivatives positions on individual securities, security indices, and security tranches.

JPM Bank’s credit derivative activities had a checkered history. Prior to 2006, they were in various parts of JPM Bank with (portfolio) names, such as the “EMEA Credit Tranche,” “CIO International Synthetic Tranche Book,” “Synthetic Credit Tranche Book,” “Tactical Asset Allocation Portfolio,” “Discretionary Trading Portfolio,” and “Mark-to-Market Overlay Portfolio.” To the average customer or investor, most (if not all) of these names were nebulous and confusing. It was not until 2008 that the “Synthetic Credit Portfolio” was created as part of the CIO.

More importantly for the London Whale saga is it was not until January 27, 2012 (roughly four years after its creation) that JPM Bank informed its main regulator, the OCC, about the SCP functioning as a separate portfolio. During 2011, the SCP grew by 1,175% (from $4 billion to $51 billion) and, by January 26, 2012, had already lost $100 million, with at least $300 million more expected to come. Despite its size and rapid growth, the CIO was still making major presentations in March 2012 without listing the SCP as one of its primary investment portfolios.

As late as April 2012, some senior managers within JPM knew little or nothing about the SCP’s activities and size. Whatever camouflage may have existed, the Wall Street Journal and Bloomberg News bluntly removed it on April 6, 2012, when they revealed the SCP’s enormous trading volumes and nicknamed its lead trader, Bruno Iksil, the “London Whale.” JPM Bank’s Operating Committee waited until April 13 to inform the Securities and Exchange Commission (SEC), and it was not until May, just a few days before JPM Bank disclosed the SCP’s initial $2 billion loss, that JPM Bank provided the OCC with detailed information on the portfolio’s positions, losses, and collateral disputes. These collateral disputes amounted to $690 million, involved ten different counterparties, and centered on disagreements over credit index valuations.

The SCP earned positive returns from 2006 to 2011. Therefore, its mandate did not come under intense internal scrutiny. A consensus view was JPM Bank created the SCP to hedge the bank’s tail-end risks, which are low-probability events that can have oversized negative financial consequences. There were considerable differences of opinion over exactly what the SCP’s responsibilities were supposed to be.

Mistake #1: Ignoring the SCP’s Strategic Purpose

From 2006 to 2011, the SCP’s activities seemed to be broadly and consistently in line with the strategic purpose for which it was created, namely to mitigate JPM Bank’s credit risks. JPM Bank’s loan portfolio exposed the company to credit risk; so, the SCP bought credit protection to hedge these positions. During the 2008–11 period, prices changed in ways that allowed SCP traders to earn revenues of $1.8 billion. Life was good because the SCP’s strategic purpose (i.e., hedging) dovetailed so nicely with evolving market conditions. When substantial losses were incurred in 2012, JPM took a second look at the SCP’s trades and positions, revealing many discomforting issues. U.S. Senate investigations that followed the SCP’s losses exposed even more.

As early as 2007, an internal audit labeled the SCP’s trading activities as “proprietary position strategies” (i.e., speculations, not hedges) and scolded the portfolio’s managers for their multiple calculation errors. In May 2012, five years later, an OCC bank examiner called the SCP a “make believe voodoo magic composite hedge.” Post-crisis testimony before U.S. Senate subcommittees uncovered wide differences of opinion among senior JPM managers about whether the SCP’s job was to hedge or speculate, whether its time horizon was short-term or long-term, and whether the portfolio was a conservative user of known trading strategies or a test-ground for new ones. This testimony also exposed considerable differences of opinion about whether the SCP was supposed to focus on controlling JPM Bank’s tail-end risks, the CIO’s risks, exposures of dedicated assets in JPM Bank’s balance sheet, or international positions.

Critics rightly asked: If the SCP was really hedging, then: Why weren’t its trades identified as such, beforehand, as they are supposed to be? Why weren’t they tracked and tested for effectiveness? Why did JPM Bank’s firm-wide risk managers know little or nothing about the SCP’s activities and have no role in creating, approving, coordinating, or executing its trades? Defiantly, one senior JPM Bank manager went so far as to criticize the OCC for expecting there to be a “reasonable correlation” between the SCP’s macro hedges and JPM’s assets. In his 2012 testimony before a U.S. Senate subcommittee, CEO Jamie Dimon asserted that the SCP’s purpose was to make “a little money” when market conditions were benign but be positioned to earn more significant returns if there was a credit crisis. By contrast, Ina Drew testified that the CIO’s (and, by inference, the SCP’s) mandate was to “optimize and protect the firm’s balance sheet from potential losses and create and preserve economic value over the long term.”

Table 11.1 shows the SCP’s record of success from 2008 to 2011. As satisfying as these returns must have been, scratching just slightly below the surface revealed a disturbing fact. SCP was not hedging; rather, it was speculating. In post-crisis investigations, no documents could be found detailing the SCP’s hedging objectives or strategies, and there were no records identifying the specific assets, portfolios, or tail-end events that this portfolio was supposed to protect. Nowhere was there information on the size, nature, or effectiveness of the SCP’s hedges. When asked for correlation estimates between changes in the value of the SCP’s holdings and any other JPM Bank activity, senior managers had none. Similarly, there was no justification for why the SCP’s hedges were treated differently from the CIO’s other hedges, such as mortgage servicing rights and interest rate hedges. Only weak justification could be given for why the SCP’s senior managers were compensated on the profits they earned, rather than the effectiveness of their hedges. Some CIO employees were among the highest paid workers at JPM. Summing just two years of compensation (2010 and 2011), the salaries of Achilles Macris, Ina Drew, Javier Martin Artajo, and Bruno Iksil amounted to $32 million, $29 million, $24 million, and $14 million, respectively. Most people in the world could live comfortably for the rest of their lives on what each of them earned in just one year.

Two additional pieces of information raised special concerns about the SCP’s true strategic purpose. First, between January and March 2012, the SCP’s balance sheet footings increased from $51 billion to $157 billion (i.e., more than 300%), totally out of line with changes in both the size and composition of JPM Bank’s loan portfolio. Second, instead of reducing its RWA by $25 billion (from $51 billion to $26 billion) during 2012, as directed, the SCP significantly increased them, costing JPM Bank millions of dollars in additional capital requirements.

Mistake #2: A Failed Trading Strategy

In 2012, JPM Bank ordered the SCP to reduce its credit protection positions. The most direct way to do so would have been to sell them at market value, but this option was dismissed quickly because the SCP had become such a large part of the market that finding counterparties to its trades would have adversely affected credit spreads, causing estimated losses between $400 million and $600 million. In retrospect, this would have been a bargain. To get an idea of how large the SCP was, consider this: In April 2012, the notional value of the SCP’s short credit protection positions was $836 billion and its long credit protection positions was $678 billion. Traders estimated that hedging these positions would take eight to fifteen days, assuming they had 100 percent of the daily trading volume. In some contracts, the SCP held more than 50 percent of open interest. Neutralizing so much in such a short period would have significantly narrowed credit spreads and increased the SCP’s losses.

Despite the order to reduce the SCP’s RWA, Ina Drew, who headed the CIO, complicated her traders’ strategic task by simultaneously requesting that they position the portfolio to earn profits from credit events caused by defaults of high-risk borrowers. It is likely that Drew had become accustomed to profits the SCP had earned since 2008 and was especially smitten by the $453 million gain in 2011, of which $400 million came from AA’s declared bankruptcy. Drew’s “competing and inconsistent priorities” led the SCP to implement a dangerous trading strategy, which was made even more so by its size.

By late January 2012, after brainstorming the conflicting goals facing them, the SCP’s senior managers and traders found (what they thought was) an answer that might satisfy both JPM Bank’s and Ina Drew’s orders. The strategy had two legs—front and back. Figure 11.2 shows this two-legged trading strategy and its net results. Notice that the top portion of the figure shows the strategy and the bottom portion its net results. The SCP’s strategy can be explained using two credit derivative contracts, a one-year (short-term) contract on non-investment grade securities maturing in 2012, and a six-year (long-term) contract on investment-grade securities maturing in 2017.417

Short-Term Front Leg

To satisfy Ina Drew’s wish that the SCP position itself to profit from adverse credit events, traders bought credit protection on non-investment grade credit indices and tranches maturing at the end of 2012 (i.e., with one-year maturities). Their hope was that these relatively high-yield, high-risk bets would, by chance, pay off, as they did with AA’s bankruptcy in 2011. The problem was finding a way to finance the new positions, which was a major reason the back leg became so important.

Long-Term Back Leg

The SCP had a financing problem larger than just funding its new one-year front leg. The portfolio also needed to finance: (1) the credit protection positions with which it entered 2012, (2) mounting losses, and (3) new positions. To solve these funding problems, the SCP sold credit protection on investment-grade credit indices and tranches maturing in 2017 (i.e., six-year maturities). As a result, during the first year (2012), the SCP paid credit spreads on the investment-grade, one-year credit protection positions bought prior to 2012 (i.e., long positions) and received credit spreads on the new six-year credit protection positions it sold (i.e., short positions). Because the short positions were larger than the long ones, the SCP earned net revenues for the first year. From the end of 2012 to 2017, the SCP’s short credit protection positions (i.e. the ones taken out at the beginning of 2012 and expired in 2017) were fully exposed because there were no existing investment-grade credit protection positions to offset them (see Figure 11.2). The result was clear. The SCP had turned itself from a net buyer of investment-grade credit protection into a net seller, meaning it was no longer hedging JPM’s credit risks, but increasing them.

Gearing problems

For the new strategy to work, the SCP had to scale the sales of its protection contracts to satisfy all its funding needs. This proved to be problematic because short-term and long-term credit indices do not always move in tandem. Not only do the different maturities cause problems, but the composition (i.e., constituent companies) in short-term and long-term credit derivatives indices differ, adding to the misalignment. In addition, the SCP had become its own worst enemy because its massive sales of credit derivatives adversely affected credit spreads, causing its revenues per contract to fall and complicating efforts to estimate the correct number of credit protection contracts to sell.

The SCP created arbitrage opportunities

At times, the SCP sales of credit protection were so large that credit spreads narrowed to the point where the cost to insure the entire credit index was less than the cost to insure its component parts (i.e., the single-name credit derivatives in the index). As a result, the SCP created arbitrage opportunities for its trading counterparties, such as Blue Mountain Capital, BlueCrest Capital, CQS, Hutchin Hill, Lucidus Capital Partners, and Saba Capital Management, and they responded in force, using three different strategies.

The first was to simultaneously buy inexpensive credit protection on investment-grade credit indices, at spreads made low by the SCP’s massive sales, and then sell protection on each of the individual securities in the index. The second was to purchase the underlying securities in the index and hedge their credit risks by buying credit protection at profitably narrow spreads, due to the SCP’s sales. If held to maturity, these trades should have ensured net profits. The final strategy was to speculate against the “London Whale.” Believing that credit protection spreads were artificially low and unsustainable, they bought underpriced credit protection and waited until credit spreads rose—and they did!

Results of the SCP’s Two-Legged Strategy

The SCP’s aggressive new strategy suffered losses almost immediately. On January 19, 2012, Eastman Kodak defaulted, resulting in portfolio losses of $50 million. With the SCP’s large holdings of long credit protection positions, one might have expected Eastman Kodak’s default to boost revenues, but the SCP was under orders to reduce its RWA. As a result, some of its risk positions (including Eastman Kodak) expired on December 20 and were not renewed. Similarly, on February 13, 2012, Ally Financial, Inc. announced bankruptcy preparations for Residential Capital LLC (ResCap), its mortgage subsidiary. Almost unbelievably, this announcement resulted in the SCP losing value on both its long and short positions.

The Eastman Kodak and ResCap defaults were not symptomatic of general market conditions. Overall credit health improved, causing credit spreads to fall, reducing the value of the SCP’s long credit protection positions on non-investment-grade indices, and decreasing the income it earned from sales of long-term investment-grade credit protection contracts. Finally, the SCP’s explosive trading resulted in a collection of approximately 100 different credit derivatives instruments, most on indices and tranches but each in need of monitoring and control. The level of complexity they added to the job of managing this already-complex portfolio increased the chances that the new strategy would run aground—and it did!

Table 11.2 shows the SCP’s monthly and quarterly losses from January to September 2012. Because virtually all its positions were transferred to JPM’s Investment Bank on July 2, 2012 and closed out as soon as possible, the reported losses end in September.

Table 11.2: The SCP’s Reported Mark-to-Market Losses, January–September 2012

Source: JPMorgan Chase & Co., 2012 SEC filings; OCC spreadsheet, OCC-SPI-00000298. The PSI Report, pg. 94.

During January 2012, the SCP had losses on seventeen of twenty-one trading days, followed by losses on fifteen of twenty-one days in February, and sixteen of twenty-two trading days in March. The CIO’s and the SCP’s senior managers were perplexed but convinced that, with time, their new strategy would prove its worth. Partial relief came when the SCP was given permission by JPM Bank to unwind its existing positions “opportunistically.” But instead of reducing positions, the SCP doubled down and then tripled down by dramatically increasing long-term credit protection sales, causing the portfolio’s notional value to rise from $51 billion in January 2012 to more than $157 billion by the end of March—with $40 billion added in just a couple weeks during March.

Mistake #3: Disregarding JPM Bank’s Internal and External Risk Measures

Setting the SCP’s risk limits was supposed to be the combined responsibility of CIO head, Ina Drew, JPM Bank’s Chief Risk Officer, John Hogan, and the CIO’s Chief Risk Officer, but it was not until January 2012 that the SCP hired its Chief Risk Officer. Consequently, the SCP’s traders and risk managers both reported directly to Drew. These reporting lines created potential conflicts of interest. At JPM Bank, risk management and derivative trading decisions were supposed to be separated. Combining them removed important checks and balances from bank operations.

Under stressful conditions, the effects were easy to predict. From 2011 to April 2012, the SCP routinely breached its five most important risk limits (see Risk Notepad 11.3) and did so without internal or external repercussions. Instead of investigating, disciplining, and immediately reducing risks, JPM’s senior managers raised them for both the CIO and SCP and then trivialized these risk measures. There were periods in 2012 when the SCP breached all five of its major risk measures at the same time. Others were breached for long periods, and some violations were so blatantly large that they almost begged for intervention and supervision. For example, in January 2012, the SCP exceeded its CS01 limit (see Risk Notepad 11.3) by 100 percent, and then exceeded it by 270 percent in February and by more than 1,000 percent in April. On April 17, 2012, the CIO had exceeded its CS01 limit for seventy-one straight trading days. All totaled, during the first four months of 2012, the SCP breached its risk limits and advisories more than 330 times. An internal list of the many risk-limit violations was nine pages long.

Mistake #4: Manipulating JPM Bank’s Risk Metrics

As the SCP’s losses mounted, senior managers reacted by changing the CIO’s risk models to something “more sensible.” In its crosshairs were the Comprehensive Risk Measurement (CRM), Stressed Value at Risk (SVaR), and Value at Risk (VaR). Prior to 2012, the SCP had never exceeded its VaR limits, but that changed. Repeated violations in 2012 led senior managers to conclude that the SCP’s VaR measure was “too conservative.” Consequently, focus was put on developing a new VaR model, and by the end of January 2012, one had been created, internally approved, and implemented. Normally, new risk measures are back-tested and compared for several months to current models to determine their accuracy and to ferret out any mistakes. The CIO chose not to follow this procedure and implemented its new VaR measure immediately, without back testing and without correcting problems identified, beforehand, by JPM’s Model Risk and Development Office.

Where were internal controllers?

The new VaR had the effect of reducing, overnight, the SCP’s VaR by 50 percent, from $132 billion to $66 billion. It also reduced JPM’s RWA and JPM Bank’s required capital. Criticism of the SCP’s new risk model should have come from a May 2012 investigation by JPM’s Controller’s office, but astonishingly, even though the SCP had understated its losses by approximately $500 million, JPM’s internal investigators determined that its methods were “consistent with industry practices.” Criticism should have also come from the CIO’s Valuation Control Group, which was supposed to crosscheck and validate, monthly, the accuracy of the SCP’s derivatives valuations, but all was quiet on this front as well.

Where were Federal Regulators?

JPM Bank notified the OCC that it had developed a new VaR model but did not explain why the new estimation methodology was more accurate or appropriate than the old one and provided no explanation for why the SCP’s risks changed so dramatically from one day to the next. Problems with the new model quickly surfaced. First, it required manual data entry, reducing scalability and making the model prone to human error. Furthermore, it contained formula and calculation mistakes, making it inaccurate and undependable. In May 2012 when the SCP abandoned its new VaR measure, JPM Bank notified the OCC but did so without explanation.

Data Manipulation

Changing its VaR model was only a part of the effort SCP put into window-dressing its deteriorating positions. The staff also began manipulating financial data entered into the unit’s new VaR model. An unbiased way to calculate the value of a credit derivatives index or CDS is to average the bid and ask values from several independent dealers or to hire an outside valuation service to provide fair market values. The SCP decided that market-based, midpoint values were inferior to those reflecting its own traders’ perceptions of prevailing market conditions. So, the SCP began using more favorable valuations in late January 2012 and by July 2012, the estimated difference between the SCP’s free-wheeling (dealer perception) method and the usual consensus, midpoint-valuation method was a whopping $660 million! Not only did the SCP’s new VaR metric reduce VaR by 50 percent but also kept this risk measure 30 to 50 percent below the midpoint valuation method until it was abandoned and replaced by the old model in May 2012.

From its start in 2008, the SCP was supposed to coordinate with and use the same credit spreads and valuation methodologies as JPM Investment Bank, but this rule somehow got lost in translation. There were times when the SCP and JPM Investment Bank were on opposite sides of the same credit derivatives deal, each valuing it differently. Responsibility for marking the SCP positions to market fell on the shoulders of Julien Grout, a junior trader who was supervised by Bruno Iksil. Grout must have had his hands full. For example on March 23, 2012, recorded banter among the SCP’s traders estimated that daily losses would be between $300 million and $600 million, but Grout’s internal report showed only a $12.5 million loss for the day. The SCP ended its data manipulation in May 2012, when the unit was ordered to use JPM Investment Bank’s methodology.

Mistake #5: Publicly Misrepresenting the SCP’s Financial Condition

By March 2012, JPM’s senior management was well aware that the SCP was in trouble. Nevertheless, in mid-April, CFO Douglas Braunstein hosted a meeting in connection with an SEC filing (8-K Report) and earnings call. Braunstein tried to sooth investors’ angst by stating that the SCP’s: