Last week we ran a post on a Procurement Leaders conference we recently attended and some of the interesting nuggets of discussion that arose. We also mentioned an interesting debate going on at our sister site, SpendMatters. That debate involves price forecast modeling where we received a number of comments about the role of “advanced analytics, potential company cost scenarios (based on the price forecasts), and cost modeling vs. price modeling. The second part of that series will come out this week on SpendMatters. The debate about the effectiveness of these types of tools and applications centers around a few key principles.
The first principle or rather, subject of debate goes like this, “the actual prediction of potential price movements is great in theory but not in practice, especially for purchasing. There are legions of hedge funds, investment banks and others hiring the top mathematical and industry experts to try to model and predict price movements, with mediocre success at best. The argument continues, “If anyone could actually build a successful predictive model, the real value would not be from better purchasing, but from trading commodities futures (and printing money), or selling the model. It just does not seem like purchasing should be spending their time trying to predict commodity costs, but instead minimizing exposure to potential swings. The argument has some validity but we don’t believe it tells the full story. So let’s back up a minute and identify the value of price forecasting. First, it is not to “beat the market, though we have previously written about that subject at length. Companies use price-forecasting tools for two main purposes for planning and budgeting and second, for risk management, as the author states.
A forecasting tool has to allow for price modeling or scenario modeling. Companies need to have the flexibility to better understand potential price increases and decreases and what those fluctuations will do to the bottom line. That’s the planning and budgeting argument. By having greater visibility to the myriad of variables impacting a particular metal’s price, commodity managers can more effectively perform budgeting and planning activities.
But perhaps more important price forecasting modeling tools allow companies to develop more sophisticated risk mitigation strategies. For example, if a supplier “pads a price of a part or material based on a future delivery (and anticipated price increase), by understanding the direction and degree the input variables move, a sourcing manager can more effectively develop a hedging strategy to lock in prices (and hence take risk off the table). A whole range of sourcing strategies becomes available when companies can model the various scenarios.
Another interesting discussion around price forecasting tools involves the subject of transfer pricing (or the price charged for goods/services bought and sold within an organization, perhaps between a holding company and a subsidiary or between two subsidiaries etc). Determining the “transfer price of the goods has its own set of tax ramifications and costing models. Some have argued that the notion of transfer pricing may include a combination of both price modeling/forecasting and cost modeling (primarily because in transfer pricing no market mechanism exists to set the price). We would suggest price-modeling tools have a role to play in transfer pricing.
What do you think?