May

11

It's worth bearing in mind that unlike most of us here, sell-side bank desk traders don't really get much opportunity to pick and choose their positions as they want and when they want. If customer wants to do big size in x, they might get saddled with the other side of that trade, bad gamma and all. The trader might not even want the position, but the head of sales "authorizes" the deal because the trade is for a new, big, shiny mutual fund customer they want to facil– just before the market does something the sales dudes thought was "unthinkable"…like spiking down 8% in the space of 10 mins.

It's a bit like flight tests when releasing a new aircraft. Yes, you might be able to handle a situation that you can almost completely control…normal flight situations and some 2 sigma variations…but what happens when there's a catastrophic failure that you're not ready for? That's why airbus et al do lots of tests to see what happens when fan blades come off at maximum engine rpm; testing whether the vibrations at the edge of the flight envelope will tear the plane apart. That kind of thing. Should be the same with trading– especially given the excess-kurtosis environment we deal with as a matter of course…at least aero-physics is reasonably predictable and stable.

There's much we can learn from this. Maybe we can set up simulators to give us a random position in xyz, then run some randomly-selected real-world crazy historical scenario against us. The emotional hit wouldn't be the same, but at least it might help develop a protocoli agree that if you are the captain of your own ship, you have no-one to blame but yourself. But many participants in the market can get nailed with positions that aren't their direct choice. No one's asking you to cry for the bank traders….but, as Taleb says, it's never a good idea to ask a man if he's from Sparta.


Comments

Name

Email

Website

Speak your mind

2 Comments so far

  1. Anonymous on May 12, 2010 1:09 am

    Too many banks emphasize quant skills over trading nous or street smarts. The politics of trading can count more for your remuneration than your p/l at such places.

  2. David Riffer on May 14, 2010 10:08 pm

    The problem with scenario analyses is that it is tough to imagine the right scenarios. The things you are really trying to protect against are susceptible to failures of imagination, and it isn’t just the sales guys. One way to solve this is to use actual historical scenarios - Fall 2001, Summer 1998, August 2007, etc. That in turn focuses too much attention on recent historical experiences which are already over-weighted anyhow. It is a tough problem to solve.

Archives

Resources & Links

Search