Tuesday, May 15, 2012

What happened and why at JP Morgan

The news is filled with reports and articles on JP Morgan’s loss of $2 billion dollars on a poorly executed trade. The question people are asking is how this could happen at a bank with a very good reputation for controlling risk. In this blog entry I want to explain what happened, and provide a psychological explanation for why it happened. At its core, the story is about a game of poker between JP Morgan and hedge funds, as the former doubled down on a bad bet, hoping that the hedge funds would "fold."  JP Morgan's loss is still on paper, but as it unwinds its positions,  paper losses could turn into real money.

The story begins when the bank’s office in London, part of its Chief investment office (CIO), was buying insurance on corporate bonds that the bank owned. This insurance, called “credit default swaps,” would compensate the bank for losses, should the companies that sold these bonds go bankrupt and not be able to pay its creditors. It was a perfectly sensible hedge. It is the same as protecting the value of your house by buying fire insurance. You pay the premiums, but your house is insured.

At some point however, this investment strategy changed, and the CIO began selling rather than buying insurance. It seems that the traders in the CIO now believed that the finances of companies backing the original bonds would be improving. JP Morgan could make money by selling insurance at rates that would be very favorable to them, once it became clear that the original corporate bonds were safer than once thought. In other words, buyers would wind up paying too much in premiums. It is as if a company selling fire insurance knew that a new fire retardant was about to be introduced, so that the risk of a house burning down was going to fall substantially. If homeowners did not know this, they would wind up paying more than they had to for their fire insurance policies. Should it choose to do so, the insurance company that owned these policies could then sell them to other insurance companies, at favorable prices. While the risk of fires had gone down, the earlier, higher, insurance rates were locked in.

But when JP Morgan switched from buying to selling insurance, its previous hedge changed to what is called a “directional bet.” JP Morgan was now betting that the finances of the companies that stood behind the bonds would improve. This was a pure bet as opposed to a hedge, since it presumed an outcome that was uncertain.

How did this bet fare? Not well. JP Morgan was selling insurance on a composite of bonds, called an index. Hedge fund traders soon realized that the cost of buying the insurance with the index was cheaper than the cost of buying insurance on each bond in the index. This is called a "mispricing." It is as if the price of a bundle of groceries is substantially less than the sum of the prices of each item in the bundle. So to make money all the hedge funds had to do was to buy the insurance, lock in the lower price, and wait for prices to rise. As one trader at Merrill Lynch said, “Fast money has smelled blood.”

But prices did not rise, because the CIO traders continued to sell the insurance, effectively keeping the premiums -- the cost of insurance -- down. They were willing to throw good money after bad, effectively "doubling down" on their bet,  because they had what Jamie Dimon, JP Morgan’s CEO, once called, the bank's "fortress balance sheet" behind them. In effect, the CIO traders were like the poker player with lots of chips, who ups the ante on a bet, so that his opponents, scared by the size of his bet, will fold -- in this case, sell rather than hold onto the index.

The hedge fund owners were pissed. From their point of view JP Morgan was trading like someone who “corners a market,” and so controls the price and quantity. But since this is an unregulated, over-the-counter market, they could only complain to journalists rather than federal officials. This was how Jamie Dimon and his leadership team first got wind of the trade and shut it down. In other words, Dimon first learned of this trade and its mispricing from the newspapers.

It is tempting to think of the professionals in the CIO as rogue traders. But the office's reputation was solid. It had put on successful hedges many times before, and had even made money for the bank from some of its directional bets.  Diamond trusted the executive Ina Drew, who oversaw the unit, which is one reason that he and his top team did not monitor the trade closely. Upon realizing the extent of the mispricing, Drew did the honorable thing and offered to resign, several times in fact, until Dimon accepted her resignation. The trader directly involved in the transaction, Bruno Iksill, was skilled and talented and had a good reputation in the close knit world of Credit Default Swap traders.

Some reporters have speculated that the CIO traders were tempted by the prospect of making money for the bank, and perhaps for themselves, assuming they were paid for their performance. But the incentive system at JP Morgan was much more nuanced than this, and traders were assessed on how they managed risk, not simply on how much money they made for the bank.

I want to suggest instead that CIO traders were induced into the “game” of trading, which despite all the quantitative controls and models that evaluate risk, can feel quite personal, just as if you are playing poker face-to-face with known opponents. Derivatives trading, as opposed to trading on the stock market is a “zero sum game.” For every winner there is a loser. By contrast, on the stock market, all investors can experience an increase in the value of a share. The zero sum nature of derivatives trading makes trading feel personal since you are causing someone else considerable pain to extract personal gain. This is why the Merrill Lynch trader used the metaphor “smelling blood” to describe the contest between JP Morgan and the hedge funds. When trading switches from a discipline informed by quantitative models, to a personal contest, people are vulnerable to taking outsized risks. The markets become a game, albeit one with big stakes and personal reputations to defend. This is especially true when the number of players is small.

This hypothesis is strengthened when we consider the character of Achilles Macris, who was in charge of the London desk of the CIO, and oversaw Bruno Iksil’s trading. While this is speculative, he appears to be the kind of person who would relish a game. He once told the Financial News that he kept his personal art collection on his office walls because, "The idea is to be part of an organization that is forward looking, intellectually curious and keen to reflect passion for new ideas and creativity. We want the work environment to reflect who we are." This is not the voice of a bureaucrat or a risk modeler, but of someone who takes pleasure in putting his personal stamp on a setting. 

It is probably true that the “quants,” the quantitative modelers, are taking over Wall Street. But as long as trading is experienced as a game, the human element--with its dares, contests, reputations, winners and losers -- will persist.  


  1. Very inspiring, thank you Larry.
    There is another question: Why did the leaders of JP Morgan allow that history repeats itsself? The group of Archilles Macris, Evan Kalintjis and Javier Martin-Artajo caused in 2008 a $ 5.7 Billion loss for Dresdner Kleinwort Wasserstein with exactly the same 'gambling'modell. Their former DKW boss Andrew Pinsker did not resign as a consequence but their current female boss does.

  2. Very interesting, I did not know this. As my later blog shows Boaz Weinstein who won big this time betting against JP Morgan, also lost big in 2008 for Deutsche bank.

    Any hypotheses about the resignations or lack thereof? My sense from the press is that Ina Drew was quite honorable in the way she behaved. I don't know what stance Pinsker took.

  3. nice posting.. thanks for sharing.