The use and misuse of statistics

In today's WSJ, William McGurn has an article on John James Cowperthwaite ("Go for Bust, Mr. Romney"), which provides some background into Hong Kong's phenomenal success (he cites, for example, the rise in per capita income relative to Great Britain: it was 28% of GB's in 1960, and 137% in 1996). He credits Sir John with much of their success.

He also quotes Cowperthwaite, who "forbade officials from keeping numbers even for such things as gross domestic product":

"If I let them keep statistics...

they can only misuse them."

Perhaps Sir John was also a fan of Mark Twain, who famously said that there were three types of lies:

• lies,
• damned lies, and
• statistics.

Sadly, we all know about the misuse of statistics in the performance and risk measurement profession. Numbers are too often thrown around in improper fashions, sometimes accidentally, but too often intentionally. We've seen everything from completely made up numbers (e.g., Bernie Madoff, who is only one of many), to numbers that were concocted in a way that meets no industry standard or practice that could be considered "best." At times numbers are used to impress, even when they have no relevance to what is being sold.

And so, should we all adopt Sir John's rule of not showing statistics? There have been suggestions that this would be appropriate. For example, not many years back, as I recall, a UK regulator questioned the appropriateness of reporting returns, if "past performance is no indication of future results."

The U.S. witnessed the appearance of misleading performance presentations in the mid 1980s, when for the first time we saw institutional asset managers advertise their past performance. Many of us believe this served as a catalyst for the FAF (Financial Analysts Federation) Performance Presentation Standards, which led to the AIMR-PPS(R), and then to GIPS(R) (Global Investment Performance Standards).

And this, of course, is the answer to those who question the use of statistics: rules. GIPS applies to much of the investment market, and new UAPS (Universal Advisor Performance Standards) will apply to other segments. Those who receive statistics should ensure that the provider adheres to the appropriate rules. Of course, their claim of compliance (and sadly, even sometimes when they've been verified) is no guarantee, as a few have still managed to get away with fraud. But, this is at least a start.

And so, with all due respect to Sir John and Mark Twain, statistics have value and should be reported. The recipient just needs to ensure they understand what's being shown and ideally that it conforms to industry standards of practice.