Opinion

All the Speculation: Mitsubishi $350MM Loss

Try not to roll your eyes. Another rogue trader. It’s pretty early days, but we can probably make some reasonable assumptions about what happened based on Mitsubishi’s Petro Diamond Singapore division’s brief explanation:

Mitsubishi Petro Diamond Singapore stated the trader, “Was discovered to have been repeatedly engaging in unauthorized derivatives transactions and disguising them to look like hedge transactions since January of this year.” Bloomberg reports that he was trading Brent.

How does this even happen? Let’s run a couple of scenarios:

Let’s assume that the rogue trader is using exchange trades. He does not have the authority to speculate aggressively, so the rogue trader figures out a way to book the trades in the ETRM system as a hedge. How could he do this?

  • Rogue trader might have had his hands on a hedge clearing account. Speculative trades are from one account and hedge trades are from another. When this trader booked trades they just came into the ETRM as a hedge. This is not altogether uncommon.  Usually this happens by accident and can be corrected by end of day.  
  • Rogue trader was allowed to book his own trades. Simply flipped the drop down to hedge. This is terribly self-serving since we sell direct connectivity. People touching trades causes problems. Usually it’s a pain in the neck, I’m still at the office at 8 PM kind of problems. $350MM loss?  Yea, that too.
  • Rogue trader might have hacked the ETRM system. Unlikely but possible.

What was missed in this scenario: 

  • No Delta Report: Delta report is a daily report that evaluates physical underlying positions relative to hedges. If you have Barrels coming in on a barge(s), how much of that have we hedged?  Fully hedged has a delta of one. Here? Probably zero. 
  • No Effective Notional Report: Every trading company operating in a G20 location has had to go through some form of notional analysis (de minimis, NFC, etc.).  Risk and compliance teams have been forced to go back and forth between positions and their hedges. This kind of analysis usually includes hedges which in many instances can reduce reportable notional.  
  • Unengaged Credit: Exchange trades require margin. Little back of the envelope math would indicate this rogue trader ran up a margin tab of over $100MM.  This trader was only on board since November 2018. That means in 7 or 8 months Mitsubishi should have seen increased margin of over ~$150MM or somewhere around $20MM per month. For a unit that only has 30 or 40 employees, this should have caught someone’s attention. How about a margin by clearing account report?  
  • Position Limits: Rogue traders tend to focus on making outsized positions in a single product.  When this happens there is a pretty good chance the firm will bump up against an exchange limit.  A spot limit breach will certainly get you a call from an exchange, but an accountability limits breach might not (It’s not always a hard limit). So unless you are closely watching limits there is nothing to alert to send someone to look and see: what are all these trades?  A lot of firms likewise allocate these limits to trade desks.  

Insult to injury? Mitsubishi was alerted to this only upon the price of Brent falling into the 50s.  They liquidated the position and took the loss. It’s back into the 60s. Does this remind anyone of Metallgesellschaft?  

Anyway, if you need assistance or have questions about how to book trades into any ETRM without touching them, gathering data for reports or limits please give us a ring.