Confessions of a risk manager

[nnja]

It boils down to a position-sizing algorithm.



It may seem like that, but it doesn't. As rrp alludes to, there are a lot of intangible benefits that get factored in here. For example, at a certain late great IB, there was a lot of social pressure to hold employer stock, even after vesting. It's not like if you sold they were going to ding you at next bonus time (that is, even if anybody knew), but you drink the kool-aid in a lot of ways or else you are not seen as a team player, which definitely would hurt in the long run. Such a social construct actually is in the shareholder's benefit because of the lower tolerance for doubling down - which in turn becomes the employee's benefit (as they become shareholders). But there is no silver bullet...

[nodoodahs]

Doing something stupid because of social pressure doesn't make the thing itself any less stupid.



I understand the psychology of it; it is what makes gang membership flourish, soccer hooliganism possible, and it inspires soldiers to leap off amphibious landing vehicles and climb out of trenches.  Social pressure is powerful and often outweighs rational, goal-oriented decisionmaking.



I agree there is some marginal utility in spending money to conform to social norms, but once it gets past the point of discretionary income and minor choices like clothing, grooming, small donations to charities, etc., and into the realm of large portions of one's total wealth ... it's stupid.



Just out of curiousity:  how much does "team player" count in IB-land, all other things being equal (like talent, ability, etc.)?  And what is the typical length of a "long run" over which not being a "team player" would hurt?  My impression from outside the industry is that tenure is typically short with a median point counted in years on one hand.

[rowdyroddypiper]

Just out of curiousity:  how much does "team player" count in IB-land, all other things being equal (like talent, ability, etc.)?  And what is the typical length of a "long run" over which not being a "team player" would hurt?  My impression from outside the industry is that tenure is typically short with a median point counted in years on one hand.



It  matters a lot.  While IBs might put out a lot of marketing materials about being "beyond the leading edge", "outside of the box", "innovators in innovation", etc.  they are really lousy environments for free thinking individuals.  There is a huge value for people that go along to get along. 



Notice that at IBs there were not traditionally a lot of outsiders running it; the head honcho was often times a lifer.  That leads to management being filled more often than not with people that have similar career paths.  I'm not saying that you have to be a team player to survive, but if you aren't one it has a high probability of limiting your prospects for advancement.  There are plenty of loners, iconoclasts, whatever toiling away in IBs but they're probably doing similar stuff to what they were 5 years ago, including complaining about the talentless hacks running the place.

[dgn2]

I would argue that being a 'team player' on most trading floors - particularly for junior people - is extremely important. To even get on to the floor you need to fit into the culture...which is probably one of the biggest reasons I didn't end up and the floor and moved to the buyside. I couldn't go along with a lot of the shit that goes on. But if I could have I probably would have.

[Johnny]

Whilst interesting, this is perhaps getting away from the point. Whether or not they `should', many bank employees do in practice hold on to their shares for a long time. This means that we can get rid of some mis-conceptions. Here are a few:



Long-term holding of shares by employees has the effect of tying pay (as experienced through subsequent share price movements) to realised profits.



People claiming that bankers have a short-term horizon (not past bonus day) should understand that bonus day is the start of closer financial ties between employees and the company; of more closely aligned incentives between shareholders and staff.



People arguing that deferred equity compensation is seen as `LEAPS (at best)' should understand that these shares often represent a substantial proportion of the wealth of a bank employee.



The point is that many of the radical changes to the banking industry being suggested already occur in the current set-up.

[sfca]

"Getting away from the point" is an understatement.  This conversation swerved off the highway and ran into a tree.  Tinkering with compensation packages is not a solution to the problem of risk management failing to avoid disasters.

[Maggette]

"Tinkering with compensation packages is not a solution to the problem of risk management failing to avoid disasters."..

Sorry, I guess the threadjack started because of my comment, so I apologize. But maybe it is not a stretch to state that risk management was faced by a lot of resistance when trying to enforce some actions that would have avoided disaster or at least damped some of its effects. And some (if not all) of this "in house resistance" seems to be motivated by compensation/bonus issues. Please correct me if I'm wrong

thanx

[sfca]

Please do not apologize; the conversation just does not address the issue.  And I would disagree with your last point.  You seem to be approaching this from a microeconomic view of "given perfect information, no transactions costs, ..."  and all, then appropriate compensation can repair any economic inefficiency.  I would reject that framework.  I could go on at length about this, but I don't have a beard and don't much like poker, so no one would listen.  So my brief response would be to suggest behavioral economics as another choice.  Investors and managers especially become overconfident in their own abilities.  When you are a pedestrian in the financial district and you see the best and the brightest unable to walk or drive at the same time they are talking on a cell phone, it should give you pause for thought about people's limitations.   There are a lot of ineffiencies and limitations and failures in structure that will not be addressed by changing compensation packages.

[Maggette]

"I could go on at length about this, but I don't have a beard and don't much like poker, so no one would listen."

Please go on length about it, if you don't mind. I think you have an interesting point. So obviously, I would listen...

thx,

maggette

[sfca]

Perhaps not at length, but I'll toss out some observations.  First, there is the behavioral econ thing.  Bob Schiller, et. al.  Investors engage in herd behavior, and if they end up investing stupidly in their own portfolio during bubbles, why would one think they would be better judges in their profession.  Cycles happen, over and over despite all the bestseller books.  That is just what people do, they often move in herds and will fall off the cliff together.  Associated with that is people are way overconfident in their own abilities.  A manager can easily dismiss warnings of impending recession because he has not seen one for a while, he is making lots of money, and he can cherry pick Business Week articles indicating things are going great and ignore the warnings.  People generally don't listen to contradictory information well and hear what they want to hear.  Another problem is the mismatch between line unit and risk.  The senior risk people I've seen were not chosen for their risk skills, but rather for organizational abilities and such.  The managers in charge of risk simply may not have the skills to tackle the problems.  And they are not simple problems, they cross all products and units and computer systems and time and different personalities and so on.  How can any one person be able to do all that?  They can't.  And since line units make the money, and the risk unit would not be there without them, risk groups usually have much less political power.  Senior management is also not a good judge of risk unit information.  I learned never to use the phrase standard deviation.  I literally once had to rewrite a memo to say the current VAR model's errors over the past month should only be seen once every ICE AGE, about every 30,000 years or so.