The Excel Mistake Heard Around the World
Research done by two Harvard economists, Ken Rogoff and Carmen Reinhart, found that when a country owes more than 90 percent of their GDP, it slides into recession. This research has been used around the world to justify austerity and the growing concerns of the size of government debt. The pair met with 40 senators in 2011 and "they told them, you need to act now and that we can't afford to spend more money to stimulate the economy." Also reading research by the two Harvard economists were budget chairs from both parties, then Treasury Secretary Timothy Geithner, the Simpson-Bowles Commission and financial leaders in countries overseas. "When we were talking about the budget deficit and the debt in 2010, 2011 and 2012, everybody had this 90 percent threshold on their minds," says Tim Fernholz, a business reporter with Quartz
However, it Reinhart and Rogoff made a glaring mistake in the Microsoft Excel spreadsheet they used to calculate their averages. "They left off five countries. And that changed things pretty significantly," says Fernholz. Instead of a mild recession, carrying that much debt means a country is probably going to have mild growth -- slow, but growth all the same.
In their defense, Reinhart and Rogoff point to other studies that show high debt leads to slow growth. But Fernholz says "it's not clear if countries that are growing slowly have high debt or if high debt causes countries to grow slowly." The Reinhart-Rogoff research suggested causation instead of correlation.
While not wanting to get into a huge debate about country debt and growth models, this simple mistake shows that policy and direction can be potentially misguided or just wrong due. In my earlier blog CFOs Beware: Problems with Financial Modeling, I raise the issues of problems with Excel and how they could effect modeling and thus the expected results or strategies for a company. Excel is a wonderful tool, but must be used carefully. As I worked with companies I have noticed that many of the models are extremely complex but within that complexity there is an increased chance for error. Often many of these errors, like computer code are not realized for a long time as the models are complex and in 70%+ of the cases the error doesn't affect the outcome. However, in those cases where it does, when the error has been discovered the costs is huge.
If a company were to make a similar mistake to the one by Rogoff and Reinhart, it could launch a new product, discontinue one, implement a new pricing plan or build a new plant which was going to lose them money. It is so much easier to do either through a incorrectly used range as in the above case or some hard coded number within a formula. A good friend of mine told me that a Fortune 100 company he was working with had a $200MM hard coded number in an Excel model that no one noticed and as a result it was skewing their investment strategies.
If two Harvard professors and a Fortune 100 company are making serious mistakes like this, the odds that there are mistakes in your models are pretty high. Need to check them carefully. Good modeling and planning but be careful.
Copyright 2013 Marc Borrelli
