The Empirical Manager

RMA Journal, The, Feb, 2001 by Mark Southern

The risk manager of the future not only will safeguard the assets of the institution but will help the institution realize economic value. Intelligent use of information can help an institution become more adaptive, more quickly, to the competitive environment. Statistical analysis is important--but not everything.

In 1938, Dr. Frank Benford, a physicist at the General Electric Company, noticed that pages of logarithms corresponding to numbers beginning with the numeral 1 were more worn than other pages. He supposed that it was unlikely users had a particular preference for the numeral 1, so he set out to explain this phenomenon. To conduct his research, he analyzed more than 20,000 sets of numbers from entirely unrelated sources, including, for example, the areas of rivers, baseball statistics, stock quotations, the molecular weights of compounds, and even electricity bills from the Solomon Islands.

It may seem that any number from 1 to 9 would show up about 11% (1/9) of the time. Benford found, however, that for a given set of numbers, the numeral 1 is likely to be the first digit about 30% of the time. His discovery, confirmed by other mathematicians since then, was that this phenomenon could be described by the formula: [log.sub.10](1 1/d), where d is the digit in question.

Benford's Law increased our understanding of the environment in which we live and work. Once knowledge like this is revealed, it is inevitable that practical applications will follow. And so they did. For example, Dr. Mark J. Nigrini gained notoriety a few years ago by applying a system based on Benford's Law to detect tax fraud in the city of Brooklyn. [1] Perhaps Benford's Law could be used to detect other fraudulent behaviors as well such as accounting irregularities or expense report mischief.

Quantitative methods are poised to become essential tools for managers--especially senior managers. According to Peter Drucker, managers should focus the majority of their attention on deviations from the norm.

Large companies with volumes of underutilized data, run by people who are increasingly asked to do more with less, are fertile ground for the Empirical Manager. The Empirical Manager is someone who can improve performance by applying quantitative methods to discover profitable opportunities. Even simple statistics such as confidence intervals and standard deviation have enormous potential--especially in areas where they have never been applied.

Understanding the environment in which we operate is fundamental to systematically creating value. Measurement is a key ingredient of that understanding.

Evolving Measurement of Economic Value

For centuries, accounting data has been the mother's milk of measuring economic value. Accounting as we know it was invented in the 15th century when Luca Pacioli, a Venetian monk and mathematician, introduced the principle of double-entry bookkeeping. At that time, there was no standard for valuing a business. Double-entry bookkeeping was the elegant solution and it led to the creation of the balance sheet (note that intangibles weren't much of a factor 500 years ago). Income statements were the next major development. They gave insight into how and why the net worth section changed over the course of a given period. Gash flow statements followed. Cost accounting, the next major step, wasn't invented until the 1920s. It was invented to catch accounting up to 1800s economic theory that emphasized cost control. [2]

Each step increased understanding of economic value. Accounting data, valuable though it may be, is just the ante. It is required to play but, because it is widely understood and available, it provides little competitive edge.

Quantitative tools are being developed to catch us up to mid-1900s economic theory, which emphasizes uncertainty. In the world of statistics, Luca Pacioli's counterparts are the French mathematicians Blaise Pascal and Pierre de Ferment. Their analysis (circa 1654), relating to games of chance, led to a formal mathematical basis for probability theory. The first recorded business application came 42 years later when English mathematician Edmund Halley demonstrated how life tables could be used to price insurance. The impact on the insurance industry in the last 300 years cannot be overstated.

Two seminal papers and applications that followed demonstrate how the economic world has been forever changed by statistical analysis. The two papers come from:

1. Harry Markowitz in 1952 on portfolio diversification.

2. Fischer Black and Myron Scholes in 1970 on option pricing and the applications that followed.

Better Information Enables Better Decisions

Better information enables better decisions. Better understanding leads to fewer surprises. Fewer surprises lead to customer and shareholder loyalty.

"Our Customers Feel the Variance, Not the Mean"

--Jack Welch, CEO

General Electric

Formal Deming/Baldrige-type quality initiatives have been attempted by many organizations but there are few successful implementations. Following our theme of focusing on the abnormal we will consider two success stories.