Math: the new magic bullet

BAKER (2006) BUSINESS WEEK/ 1/12/ 54-66

Math will rock your world if you are in business. Nearly 12 pages is devoted to the uses of math.

However, as good as it sounds, one hopes that business still relies on qualitative analysis to compliment the math models.

It is called triangulation. You can use up to 3 different strategies, one of which is qualitative, to assess and optimize a solution(s.)

The article does deal with some of the fuzziness of human behavior. Although there is a model to quantify words. However, here are some of the problems that have troubled sociology, political science and psychology. Economics is now entertaining BEHAVORIAL ECONOMICS that is more qualatative than quantitative. So here are some problems.

HARD NUMBERS/ To master a human institution like business, you need to use hard numbers in multiple variable modeling and number crunching. However, many human activities such as panic stock selling, and economic bubbles are more amenable to soft numbers. Soft numbers are not as easily put in multiple variable analysis. General human reasoning is more likely to compliment soft numbers.

RATIO DATA/ Ratio data means that there is an absolute base line, that each number is mutually elusive from the others, are equidistant from each other, and that number multiples are exactly always greater or lesser than each other. How do you define profitability when measuring customer satisfaction and repeat business when the seller has recently merged?

INACCURATE DATA/If you have multiple variables and one is flawed it throws the entire model in disarray. You believe that you can control multi-collinearity when the best linear unbiased estimator is flawed by the spurious variable. Further, some profitability is cooked like expensing.

CHAOTIC BUTTERFLIES/ Your model does not include a low probability but high impact variable that acts a wild card and is not quantified in the math model. It is treated as noise and is over looked to the chagrin of the statistician.

Math should be considered a welcome addition as long as the qualitative is not overlooked.

Further, in the past all methods have failed. Qualitative ones are easily remembered, but quantify ones like the collapse of hedge funds and econometric models are easy to forget by many because they don't understand them in the first place.

So, hold  your wallet, until the modeling is perfected and the most qualified number crunchers admits that a non-math assessment may be of value.