JPMorgan Chase loss seems to be an idiosyncratic trading loss

The JPMorgan Chase loss seems to be a plain-vanilla trading loss – unfortunate and bad news, but not a sign of something more serious in the financial system. Bad as they are, we have lived with these kind of trading losses for decades.

Details about the loss announced by JPMorgan last week are sketchy, but what I can discern tells me that it is more akin to a standard trading loss (an idiosyncratic loss, having to do with specific issues at JPMorgan) than a systemic loss (related to macroeconomic or financial system issues).

There is a long history of trading losses at financial firms, and studying these losses both helps us to understand the sources and responses to such events, and also helps put this specific event into perspective.

The announced $2bn loss is large, but by no means the largest – LTCM in 1998 was $4.60bn, Amaranth Advisors $6.50bn in 2006, and Société Générale $7.22bn in 2008 – and were we to adjust for inflation and growth in the economy those would be even larger today. And LTCM and Amaranth were small compared with today’s Morgan Chase. According to the Financial Times the portfolio within Morgan Chase that suffered the loss was $360bn, so although $2bn is large in absolute terms it is only 0.6% of the portfolio – this does not minimize the loss but again helps put it in perspective.

When we examine trading losses over the years, two facts become apparent:

  • They occur more regularly than we would like to think
  • The sources and reasons are varied, but most are not “rogue traders”, and transparency and mark-to-market matter

In chapter 4 of my recent book (Quantitative Risk Management published by Wiley, also available in the CFA monograph
A Practical Guide to Risk Management, available for free download)
I examine 35 cases of trading losses over the years, ranging from the large and well-known (LTCM’s collapse in 1998) to the smaller and more obscure (traders at National Australia bank losing $310mn in 2004 on FX trading). Fully 40% involved no fraud, and only 26% involved fraudulent trading that fits our image of the rogue trader. Many cases originated in standard business which either went badly wrong (for example Metallgesellschaft in 1994 where a poorly-designed hedge against oil prices went disastrously wrong, or Askin Capital Management, also in 1994, where investments in principal-only securitized mortgage securities lost heavily when interest rates rose) or got out of control (Aracruz Celulose and Sadia, where FX hedging apparently morphed into outsize speculation).

One lesson we can take from examining these cases is something which should be blindingly obvious – financial markets are risky and sometimes bad luck happens. Investors and managers often make mistakes and foolish decisions, but sometimes there is simple bad luck. Personally, I think LTCM made foolishly large one-way bets on swap spreads and long-term equity volatility, but they undoubtedly suffered the bad luck of having Russia default on its debt at a time when the financial markets were somewhat unsettled. (In fact, a friend of mine escaped the Russian debacle only because she had the good luck to forget to roll an expiring position in Russian bills when they came due on a Friday and thus owned no bills at the time of default.)

Financial markets are risky and bad things do happen, and so transparency and mark-to-market are vitally important. When losses do not come to light quickly they tend to grow. It is human nature to avoid bad news. Case after case shows the danger of allowing losses to sit and the all-too-human tendency to try and fix the problem, try to trade out of the loss. JPMorgan has at least owned up to the loss with some alacrity.

One final lesson we should take from examining idiosyncratic trading losses is that they are bad, but pint-sized relative to losses from systemic events. Consider Hypo Group Alpe Adria (discussed on p. 120 in the CFA monograph, p. 132 in the Wiley book). They had an idiosyncratic loss of eur 300mn in 2004 related to a currency swap. But then they got caught in the global financial crisis, and in December 2009 had to be rescued by the Republic of Austria – after-tax loss of eur 1,600mn for 2009. Another example – Dexia Bank – idiosyncratic loss of eur 300mn in 2001, but losses for 2008 of eur 3,300mn (and more recent nationalization). And don’t forget Fannie Mac and Freddie Mac – as of May 2010 the US government had provided $145bn of support, and it has only grown since then. Idiosyncratic losses are measured in the hundreds or million dollars, maybe a couple billion. Systemic losses start in the billions and run to the hundreds of billions. Idiosyncratic are one or two orders of magnitude smaller than systemic. Idiosyncratic losses can bring down a single firm. Systemic events can bring down a nation.

We should remember that the loss JPMorgan announced last week is firmly in the category of idiosyncratic loss, rather than systemic. The discussion of reasons and responses should be related to the nature of that idiosyncratic loss. Unfortunately, much of the discussion has been, and will continue to be, around regulation and responses for systemic risk.

About Thomas Coleman

Thomas S. Coleman is Senior Advisor at the Becker Friedman Institute for Research in Economics and Adjunct Professor of Finance at the Booth School of Business at the University of Chicago. Prior to returning to academia, Mr. Coleman worked in the finance industry for more than twenty years with considerable experience in trading, risk management, and quantitative modeling. Mr. Coleman earned a PhD in economics from the University of Chicago and a BA in physics from Harvard College.
This entry was posted in Musings, Risk Management and tagged . Bookmark the permalink.

Leave a Reply