Understanding Turns: Building a Better Forecasting Model at Dow
In my last post, I explained how when it comes to forecasting, Dow Chemical Company has recently embraced the notion that understanding the timing behind economic upswings is as important as forecasting accuracy.
So what did Dow do next? To start building a framework that would connect recessionary cycles to business cycles, Dow looked to tap sources of existing economic data.
“We first thought, Well, who is doing this type of work? And how do we use it to a different advantage?” said Tim Rey, the manager of Dow’s advanced analytics group, situated in Dow’s Six Sigma Expertise Center. According to Rey, Dow first turned to the Chicago Fed National Activity Index.
“What we tried to do initially is look at the five basic sectors that are in the Chicago Fed index and find like variables in the other country databases to build a country-specific Chicago Fed index,” explained Rey, who said that the search for indexes capable of better revealing the timing behind cycles eventually led to the Economic Cycle Research Institute (ECRI).
“So we said, Now we have two sources. And just as these guys were doing their work to characterize recessions and get a better handle on when they begin, we realized that these concepts and techniques can be used to look at business cycles, not just large recessionary cycles,” explained Rey, who recently gave a presentation on the subject at F2010, a business forecasting conference hosted by analytics developer SAS.
According to Rey, the audience response to Dow’s new forecasting approach was twofold.
First, people were empathetic to a one-index approach. “There is just tons of information out there,” said Rey.
Second, they quickly recognized the value behind being able to leverage the upward part of a cycle. ###









