We have rarely (if ever) published any case studies let alone talked about the quality of various metal price indexes available in the market today. However, we recently received an email from a colleague, Doug Peterson, who wrote in asking if we were aware of any report or information examining the accuracy of any of the metal price indexes (we were not aware of any). Doug, who works for an industrial manufacturer, wanted to make the case to his business units that using an index (Doug’s company buys 304 stainless steel) provides a reasonably good means of understanding market changes such that the individual business units can hone their sourcing strategies based on index data. Doug set out to prove his case by comparing actual market data to the forecast data.
To that end he evaluated two different sample data sets one from 2008 and the other from 2009. His results suggested that the index he had chosen was within 4% of the actual price one month out and 10% of the actual price three months. Where the index fell down, however requires a bit more explanation. The index was not particularly accurate predicting 6 months out during the 10/08 3/09 period (but was fairly accurate in predicting the 7/09 1/10 period). We suspect the issue had more to do with the dramatic drop in stainless markets after the summer of 2008. Many forecasts examine multiple variables using three months of trailing or historic data to create the forecast. And nothing in the historical data would have predicted the dramatic drop in most metals markets during Sept/Oct 2008. But we don’t believe that serves as a reason (or excuse) for not using metals indexes.
Doug didn’t stop with that analysis. He also wanted to examine how well his company’s actual purchase prices correlated to index data. In this case, his company only did okay. 21 of the 37 times that his company’s prices changed, they were within 9% of the forecasted index price. All of the changes were directionally consistent with the index. For items purchased every month, the data had the best correlation to the index. Going forward, we would suggest Doug provide a “how to for the business units. The company can test Doug’s hypothesis by showing how one unit, using index data and adjusting sourcing strategies compares to a unit that doesn’t track to an index.
Why go through the efforts of such an analysis? Doug set out to prove two points to the business units. The first reason, according to Doug is to show the business units that using price indexes is a reasonably good tool to hone supply strategies, possibly on a quarterly basis. The second reason to conduct this kind of analysis relates to results – the company achieves a better result by using the data than either doing nothing or expecting more out of the supply base.
If you are interested in learning more about how price indexes can help your organization manage price volatility, click here to download a free report on managing commodity volatility.