Winners and Losers
Sometimes quants who get involved with building new economic capital models have the opinion that their work will reveal the truth about the risks of the group and that the best approach is to just let the truth be told and let the chips fall where they may.
Then they are completely surprised that their project has enemies within management. And that those enemies are actively at work undermining the credibility of the model. Eventually, the modelers are faced with a choice of adjusting the model assumptions to suit those enemies or having the entire project discarded because it has failed to get the confidence of management.
But that situation is actually totally predictable.
That is because it is almost a sure thing that the first comprehensive and consistent look at the group’s risks will reveal winners and losers. And if this really is a new way of approaching things, one or more of the losers will come as a complete surprise to many.
The easiest path for the managers of the new loser business is to undermine the model. And it is completely natural to find that they will usually be completely skeptical of this new model that makes their business look bad. It is quite likely that they do not think that their business takes too much risk or has too little profits in comparison to their risk.
In the most primitive basis, I saw this first in the late 1970’s when the life insurer where I worked shifted from a risk approach that allocated all capital in proportion to reserves to one that recognized the insurance risk as well as the investment risk as two separate factors. The term insurance products suddenly were found to be drastically underpriced. Of course, the product manager of that product was an instant enemy of the new approach and was able to find many reasons why capital shouldn’t be allocated to insurance risk.
The same sorts of issues had been experienced by firms when they first adopted nat cat models and shifted from a volatility risk focus to a ruin risk focus.
What needs to be done to diffuse these sorts of issues, is that steps must be taken to separate the message from the messenger. There are 2 main ways to accomplish this:
- The message about the new level of risks needs to be delivered long before the model is completed. This cannot wait until the model is available and the exact values are completely known. Management should be exposed to broad approximations of the findings of the model at the earliest possible date. And the rationale for the levels of the risk needs to be revealed and discussed and agreed long before the model is completed.
- Once the broad levels of the risk are accepted and the problem areas are known, a realistic period of time should be identified for resolving these newly identified problems. And appropriate resources allocated to developing the solution. Too often the reaction is to keep doing business and avoid attempting a solution.
That way, the model can take its rightful place as a bringer of light to the risk situation, rather than the enemy of one or more businesses.