Luck and Risk Management

Each year, our holding company holds its strategic management offsite in the conference room of a casino somewhere. At the end of each hard day of strategic thought, most managers end up hitting the tables and trying their luck on various games. Some ultimately do better than others. The lesson in this exercise is that while luck is important in gaming, it leaves much to be desired in the field of risk management. Those who don't play the odds are relying on luck; and luck will sooner or later run out. While probabilities and statistics are often overlooked in daily operations, they should be the lifeblood of any financial institution. What is the probability of any given loan defaulting over its life? If your portfolio approaches the average, it can be as high as 3.8%. Like Vegas, if you don't know the odds you might get lucky for awhile, but you will not remain lucky for ever. Consider the banker walking up to the craps table for the first time. They take the two dice and attempt to roll a 7. If they roll it on the first try, they got lucky. However, by rolling the dice 10k times, it is a foregone conclusion that they will roll a 7 approximately 16.7% of the time. For those bankers armed with proper data and proper modeling, luck becomes an item for rabbits. This same concept is behind value-at-risk methodology becoming popular with many banks. By looking at the historical or expected prices of loans and securities, potential future valuations can be modeled. Running the model 10k times delivers a range of possible outcomes and increases the level of certainty. The higher quality the data, the better the model results will be. Operating without proper data or models and running on hunches leaves an institution more reliant on luck than they probably would like to be. What senior manager wouldn't like analysis completed that provides a definite conclusion to a problem, particularly with some degree of certainty. The ultimate goal of risk management is to facilitate a consistent implementation of both risk and business policy. In so doing, managers often place more value on rewards, since earnings are tangible and known, than on risk (which represents something intangible and uncertain). By collecting and quantifying data on processes such as credit impairments, liability performance, security values and operational failures, the first steps can be taken to make the intangible, tangible. As casinos know, people who feel lucky can win large amounts of money. However, the longer they stay in the casino and the more they play, the more likely they are to eventually give it all back. Those who analyze the risk before walking into the casino (and read the odds on the payout signs) know that the best games to play are usually craps, blackjack and video poker. Those who don't do this analysis usually rely on luck to carry them through and may end up playing keno (usually the worst odds in the casino). Luck is important, but risk analysis and quantification is better.