Whether you buy metals, plastics, energy, electronic components or other base materials or parts/components further down the supply chain, it is becoming essential to have a viewpoint on where the markets are going. But developing a perspective requires more than just observing the market and understanding your own demand drivers in isolation -- not to mention your organization’s ability and willingness to tolerate risk (from price increases to even being put on allocation, in unique market circumstances). It also requires understanding and modeling the inputs that are indicative of where the market might go next -- a month, a quarter or even a year out. I recently had a chance to sit down with Lisa Reisman, who is building a new scenario price modeling application for the metals markets, to get her viewpoint on the subject (in full disclosure, I happen to be married to her). Here's what she had to say about why we need to think about the need for such a capability today -- and the types of variables that factor into commodity price forecasting and modeling:
"We're in the process of developing a browser-based MetalMiner scenario price forecast modeling tool. The project is an outgrowth of our metals pricing forecasts which leverage similar data already. But perhaps more importantly, it is the result of us hearing the same message from multiple readers and clients -- the art of forecasting has become very challenging. Even the metal producers struggle with raw material volatility and we know industrial companies, oil and gas companies and service centers go to great lengths to better understand where the market may go. They are struggling to understand what impacts the prices of various metals. Interestingly, this is relatively new issue. When I was a metals trader fifteen years ago, we cared far less about where the price of aluminum or stainless sheet would be six months out -- primarily because price didn't fluctuate as much as it does today. But it matters now.
In other words, the price band for all metals used to trade within a much narrower band. The steel price trading band ranged from $300/$500 per ton. But in July of 2008, prices for some types of steel reached $1100/$1300 per ton (and some were higher). Since then, prices have come back down and have since increased dramatically again. The volatility story, however, came earlier. Some point to Q3 2003 as the start of this new commodity bull market. And what this has taught us is that our traditional rearview mirror analysis, relying on a few key variables no longer works. You can no longer look at a trailing 3 months of data with six variables and expect to accurately forecast out three months. Spend visibility tools -- even those that let you strip out raw material costs -- are a downright waste of money because prices change so dramatically so quickly.
What is needed is a new approach to price modeling and forecasting. But it does not need to be a crystal ball either. We need to look at the variables in a rational way and more rationally predict where prices are going. This requires understanding the new variables -- and a lot of them are new -- that impact price as well as how they interact with one and other and the weightings of these variables on the overall forecast.
Let me recount the questions we must ask. These are what matter as we think through the scenario modeling and price forecasting process for a range of metals (the answers to the questions will be different for different metals) not to mention other commodities and raw material inputs:
- What are the variables (e.g., new housing starts, country PMI, etc.)?
- How do they interact with each other?
- How do they impact the price -- and in what global markets?
- What is the best way of weighting of these variables?
- How will variables change -- and what will cause them to change? Today's variables are not necessarily constant and the weightings of these variables change over time." To be continued...