Shostack + Friends Blog Archive


How Much Is Risk Management Worth?

David Akin blogs that Fitch Ratings has purchased Toronto’s Algorithmics for $175M (the press release is datelined New York, so I’m guessing that’s a US dollar figure).

Algorithmics makes risk management software, focusing on market risks for banks, things like hedging strategies and BASEL II compliance (based on a quick read of their site.) So one answer is that better risk management is worth $175m.

But presume that you know a lot about other banks’ risk management strategy, because you make the software that drives it. Can you anticipate their actions and use that against them? (Fitch seems to be only in the ratings business, and so may not be positioned to do this.)

2 comments on "How Much Is Risk Management Worth?"

  • James Nesfield says:

    Most risk programs require time to build the data for it to become viable. There are some that claim they can do this in less than one year and others that have achieved success in two years. Basically it is what is generally acceptable and Fitch would know because they are the most accepted. Fitch is buying market share plain and simple the user base Algorithmics has and Fitchs ability to copy what they have done it is easier to buy them. EBITDA times 8 or 6 or whatever mulitple they paid over to obtain the client base. Risk matrix also fill a big hole in the securities fair value model meaning that without a rating a method must be found to place a value on the stuff. So risk provides that without a liquid marketplace. There was a firm RiskMetrics that was for sale and the salaries of the those that created it made it an unprofitable venture although it was eventually purchased by one the banks. After they removed the high priced executives they had a going concern that they serviced out to their clients for the purpose of price assignment of illiquid securities and assets.

  • adam says:

    Interesting! How much of the building data is collecting the raw data, versus sifting it with your new theories and testing which ones have predictive power?

Comments are closed.