Insurance-Canada.ca - Where Insurance & Technology Meet

Lies, Damned Lies, and Spreadsheets

When the only tool you have is a hammer, lots of things can begin to look like nails. And when the only tool you have is a spreadsheet, even mistakes can look good in pie charts. Some recent examples in financial services and government suggest better discipline is needed.

While we like the convenience of tools such as Excel, there are serious risks in indiscriminate use. We’d like to hear from you on how to make use of tools while avoiding pitfalls.

Excel:  Faster than  a Speeding Lotus, More Powerful than Ginseng

There is no doubt about two things:

  1. Excel is a very powerful analytic tool
  2. Excel is the predominant user tool is numeric-heavy industries, such as insurance.

And the ubiquity of Excel is spreading well beyond the natural habitat of accounting and actuarial.  Consider a recent announcement of a course for lawyers involved in insurance claims:

Excel is an incredibly useful tool for insurers and policyholders, but it is underutilized. Lawyers mistakenly shy away from Excel and view it as a tool for accountants and engineers. That’s a mistake. This program is designed to help lawyers use Excel in resolving insurance coverage disputes.

In 1.5 hours, the on-line  course  promises to show how Excel can assess different allocation approaches, weigh different outcomes, and facilitate settlement.

What could go wrong?

A recent post in the on-line blog Quartz, titled ‘Damn You Excel’,  shows some examples which suggest a grain of skepticism might be in order when looking at spreadsheets.

  • JP Morgan underestimated risk in  its synthetic credit portfolio, resulting in a write down of $6 billion.  The cause?  According to James Kwak, a combination of a faulty formula and cut-and-paste errors.
  • Researchers Carmen M. Reinhart and Kenneth S. Rogoff published a widely cited book This Time It’s Different which contended that empirical data showed that exceeding public debt levels of 90% placed economies at peril. The book was used by several European countries to justify austerity.  However, a subsequent review by the University of Massachusetts found that the fundamental research was flawed due, in part, to Excel errors.

There is nothing inherently wrong with applications like Excel.  However, indiscriminate use of these tools, with blind acceptance of results is a recipe for ridicule at best and expensive mistakes.

What do you think?

Are we too trusting of technology like Excel?  Do you have some horrror stories of your own?  Better, do you have safety measures you have put in?

 

Tags:

2 Comments

Ian Totman

It’s not so much the technology as it is the process. If your process has manual steps or potentially dangerous steps (e.g. copy/paste; an unprotected spreadsheet) then you need to assess risk and address it.
One reason we get these situations is that Excel is so quick and easy for prototyping solutions. Then it becomes engrained into the process and doesn’t have the appropriate controls!

Insurance-Canada.ca BlogEditor

Excellent point, Ian … In the age of doing more, in less time, with fewer resources, we are seeing lots of prototypes that get promoted to production with little testing and virtually no thought to security, consistency, maintainability, and support.
-Editor

Comments are closed.