Print Share Home

Purchasing Price-Volatile Commodities: Risk Control And Forecasting


John D. Baker, Ph.D.
John D. Baker, Ph.D., President John D. Baker and Associates, Bethesda, MD 20817, 301/469-7954.

80th Annual International Conference Proceedings - 1995 - Anaheim, California

Abstract. Significant volumes of price-volatile materials in the cost structure of a business can pose a risk-control and cost-savings challenge to purchasing. This paper addresses the various risks that are involved, how to determine what is worth doing, and several ways to devise a purchasing strategy for such materials. Focus is placed on major factors that impact the decision-making processes of purchasers.

Background. The last 20 years has seen a rapid acceleration in knowledge and in the use of computers for technical analyses, testing, and decision making. Artificial intelligence systems, neural networks, and genetic analogs are widely in use, although more so in the financial markets. There has been significant expansion in the sheer number of exchanges internationally, and in the variety of commodities traded. Methods and markets for off-floor trading are increasingly available. Such developments create a parallel rise in purchasing opportunities but add to the speed with which price-making forces cause markets to change, thus making forecasting ever more difficult. Conditions like these call for managers to develop high levels of knowledge and expertise in risk control, price forecasting, decision making, and in the use of technology for purchasing. Even so, basic principles of decision making still apply. Many simple approaches to purchasing remain quite effective and frequently offer an advantage when combined with the more sophisticated forecasting systems.

Key Factors. There are three key factors to consider when developing a purchasing decision-making system. (1) A company's objectives and philosophy, which taken together, is strongly influenced by the size and character of the business, as well as by the attitudes of management. (2) The commodity's importance to profit or loss. (3) The price-volatility of the commodity.

Risk. When a volatile commodity is a major cost component, risk identification is required. Nearly always, the risk of price-change from the time of raw material purchase and the sale of the finished product is the criterion used. This risk factor is essential to processors of basic commodities such as flour millers, soybean processors, country grain dealers, or any manager making a contract for the future when costs are not fixed.

For consumer products, it may be necessary to consider the trading practices and market positions of the competition and what that will mean in terms of finished goods prices. Here, the risk of losing market share comes into play. The risk may not be what price change might occur, but the coverage and inventory price of the principle competitors. It might be that hedging against price change increases, rather than decreases, risk. The no-risk base may be more difficult to determine if one or two companies comprise the real competition, particularly if they are risk takers and have significant variations in the length of their commitments. Should it prove necessary to determine a no-risk base, a competitive cost estimating system can be established to utilize available information.

Should the risks to the bottom line be significant, then, with the approval of management, critical resources of knowledge, information, and technology should be obtained, and organizational processes put in place to facilitate optimum use of these resources to meet this purchasing challenge. Measurement of Risk. If the company is not significantly attempting to reduce risks and costs, then changes are needed in purchasing or in the management policies of the company. Factoring in price-volatility, the company must draw comparisons between its buying practices and the purchasing practices of the industry in general, though more specifically with those of the company's major competition. One simple method for measuring volatility in an agricultural commodity is to calculate the annual average percentage in price change from the low to the high over a period of five or ten years. This can be compared to the estimated average cost to competition. Estimates can be made indicating how hedging and more advanced price forecasting and decision-making systems would benefit the firm. Some tools for doing so are discussed in the Performance Measurement section, below.

Management's View Of Risk. The above process and its results may hinge on management's attitudes, philosophical orientation, and policies. If management does not recognize and appreciate the importance of purchasing, then it is incumbent upon purchasing personnel to educate management.

In addition to conducting a basic risk analysis, the buyer or purchasing manager might utilize a table of purchasing options to make the case to management. This tool presents management with objectives the firm might like to achieve against the various purchasing options available. The utility of this method is discussed by Howard Madsen in Economics and Management of Food Processing (W. Smith Grieg, editor). The following is an example of such a table.

Purchasing Options
Objectives Buy at
market Weekly
Buy and
Protect profit plan No Yes Some Some
Reduce cost volatility No Yes Yes Some
Reduce costs No Possibly Possibly Possibly
Protect inventory values No Yes No No
Assure supply No Some Yes Yes

Another technique is to ask management to recognize and rank their various kinds of risks. Included are such things as (1) highly fluctuating material costs; (2) product vulnerability to competition; (3) plant construction costs for a new product; (4) introduction of new products; and (5) changes in economic conditions.

If management is risk-averse with regard to volatile materials, then the buyer must make purchasing commitments in close accordance with a risk neutral position. On the other hand, should management prove willing to take risks based on adequate purchasing safeguards and the probability of price movements, then the buyer can attempt to improve the bottom line by reducing costs of materials.

Methods Of Controlling Risk When Maximizing cost Savings. This section assumes management approves the goal of cost reduction, only those materials impacting significantly on the cost structure of the business are included, and risk-protection systems are operating in market phases of uncertainty.

The initial coverage or commitment decision, is very important, Good forecasting systems and good information databases are required for determining both price and the expected difference between delivered price and the futures. However, other safeguards are required because of the serious consequences on the financial well-being of the business and on the buyer (in terms of the severe mental and emotional stress that this kind of decision making can bring). One or more of the following control methods might be utilized:

  1. Coverage or commitment limits. These might be in duration of coverage or in money values, with exceptions being made only by top management.
  2. Organizational safeguards. These might take the form of a review committee, or oversight by an individual in upper management experienced in the buying and risk control measures relative to the particular material, or two buyers of the same material from separate divisions of the company agreeing on a coverage strategy. It is imperative to bring to bear substantive experience, differing points of view, and to ask pointed questions worthy of a devil's advocate.
  3. Monitoring by finance - A high-level financial officer should be educated as to the impact of risk on the business. This individual should review paperwork on all positions of actuals and futures.

Once the decision to commit is made, the means whereby this is accomplished should be left to the buyer. Various means to choose from include buying the actual material; using a buying hedge in futures; using options; cross-hedge purchasing (if the material does not have a futures market); or trading derivatives, as in the case of financial and currency transactions. The buyer must make informed trade-off decisions when executing the buying strategies.

As soon as the planned needs for materials have been covered, the objective is to protect the commitments from adverse price movement. Should the situation change or should conditions become uncertain, prompt action is required. For example, the buyer might hedge in the futures market, which is common, or use options. Other risk-limiting tools should be examined and implemented in response to the need.

Methods of Determining Price-Risk Probabilities. It is critical to determine whether a price is high or low, and what the probability of price change will be. There are several ways to do this effectively.

Where statistical models are being used, probabilities can be calculated directly from those models. However, models are usually based on a limited number of variables, as well as on historical data. To the extent the present situation differs from the past, the probabilities will have to be adjusted through extrapolation and/or in-depth analysis, especially in the case of deviations that might occur due to the current situation.

Charles Schroeder, in the 1994 CRB Yearbook, presents a method based on adjusted historical prices. Monthly inflation-adjusted prices for a commodity are taken for a ten-year period and ranked from the lowest to the highest. Percentiles are assigned based on the percent of time prices were below the indicated price level. These percentiles can be converted to percent chances for ranked prices to decline or advance.

A long-term price chart can be used. Simply draw two lines placing the lower one-fourth of prices below the bottom line, and the higher one-fourth of prices above the top line. (Other proportions may be chosen, and several technical systems have rationales and rules governing their use.) It is essential that a sufficient number of years be used so that seasonals, other cycles, and trends become evident and can be accounted for.

Semi-logarithmic charts can be used for this purpose of determining whether a price is high or low. These charts have the advantage of automatically filtering out many of the influences to changing price levels over time.

Price Forecasting. Before choosing or developing a price-forecasting system or model, it is important to define the company's objective. This fairly obvious task is frequently neglected.

If the company desires the highest possible increase to its bottom line, the objective is to determine the level and timing of lows and highs, or short periods covering these.

If the company is content with "following the crowd", then a buying strategy can be devised based on expected deviations from quarterly and annual averages.

The tools of price forecasting fall into two categories: fundamental and technical. Fundamental tools involve supply and demand data inputs. Technical tools involve analysis that looks at and interprets price movements as representing a synthesis of all relevant forces. Analyses of volume of trading and open interest conditions are frequently included in this synthesis. It is now generally agreed that fundamental and technical tools must jointly be used in order to make superior purchasing decisions.

Usually, discussion among buyers focuses on price-forecasting models, these being tangible and concrete tools for arriving at decisions. However, because of constant market flux, it is generally better to think in terms not of particular models, but of decision-making systems. Listed below are several brief descriptions of various decision-making systems relating to volatile materials or commodities. This list is by no means a complete one.

  1. Intuition. It is not uncommon in smaller companies to find purchasing managers using a minimum of statistical and other formal tools. Reliance is placed on years of experience and sources of information proven to be reliable. There is usually full awareness of the impact of the decision on the overall business. The drawback of this system is that it is difficult to pass on to a successor.
  2. Statistical or Econometric Models. The focus may become the forecasting of deviations in the model to better fit the current market situation. Models can then be combined with technical analyses to determine timing.
  3. Comparable-Year and Seasonal Analysis. Years are classified depending on the ratio of past or projected usage to the new supply. Additional factors, such as the rate of inflation, can be added. Seasonals are then calculated for the large, medium, and short supply years. This is really a combination of fundamental (supply-demand) analysis and price-movement analysis. If semi-logarithmic charts are available, price-level forecasts and timing can easily be made for many commodities with reasonable confidence.
  4. Bullish-Bearish Analysis. This simply entails listing all opposing forces in the market, ranking them, estimating their strength and the probabilities of their occurrence, and producing a price forecast giving direction, level, and time. This form of analysis can be combined with technical analyses and serve as a supplement or check to other decision-making systems.

Performance Measurements. If properly conducted, performance measurements can indicate possible areas of improvement and identify valuable opportunities for cost reductions. Below are listed several performance measures and comments concerning each.

  1. Cost vs. Hand-to-Mouth. This performance standard, for example, can be set by comparing weekly purchases to the actual buying price. The standard needs to include all the costs of delivery to the plant for comparability. If the buyer is making no improvement over this naive system, then the entire buying process must be brought into question.
  2. Cost vs. Competitive Costs. If the competition is represented by the market in general, then it would be the same as the above hand-to-mouth standard. Known seasonal buying patterns not represented in this standard can be used for adjustment. Prices can be weighted by a volume index to achieve a more refined standard. The presence of only one or two major competitors requires additional information or assumptions to define a standard against which performance can be judged.
  3. Cost vs. Perfection. This standard assumes that buying is done at a major low sufficient to last through to the next major low, hedging the remaining commitment and inventory at the top. This is not a good performance standard as such, though it is a measure of total potential. Estimates can be made as to what proportion of the potential is being realized. Additionally, estimates based on this standard might indicate the value of expanding, changing, or improving the forecasting and decision-making system. These techniques can be applied to determining the potential regarding basis, spreads, or other elements relating to the materials being purchased.
  4. Decision Analysis. As a learning tool, coverage decisions with accompanying rationale and supporting forecasts can be recorded in a written log, or on the computer. Actual prices can also be entered as inputs so that differences between actions taken and subsequent market behavior are automatically obtained. A frequent review of these is likely to be useful in improving the system and in improving the learning curve.

Conclusion. Tools are available for evaluating the cost-saving and risk-reducing potential in the purchase of price-volatile materials. A studious analysis and presentation of costs and risks may be required to secure the approval of management to embark on a systematic approach of increasing the company's bottom line through risk-control purchasing; this is possible even in today's complex and sophisticated environment.

Purchasing managers cannot avoid the responsibility of supplying the business with adequate raw materials or products. Buyers are not awarded the freedom of traders or speculators who can choose to be in or out of the market. Indeed, purchasing can be very much a "sink or swim" proposition. Therefore, the work must be approached with vigor, sound analysis, and creativity in the use of forecasting and risk-control tools in order for it to be both satisfying and successful.


  • Herbst, Anthony F. Analyzing and Forecasting Futures Prices. New York: John Wiley & Sons, Inc., 1992.
  • Krokll, Stanley, and Michael J. Paulenoff - The Futures Markets. Homewood (IL): Business One Irwin, 1993.
  • Madsen, Howard C. "Managing Commodity Price Risks in the Food Industry. 11 In Economics and Management of Food Process -.1 edited by W. Smith Grieg, 355-390. Westport (CT): AVI Publishing, Inc., 1984.
  • Schroeder, Charles. "Universal Commodity Price Analyses." In The CRB Commodity Yearbook, 1994, 33T - 37T. New York: John Wiley & Sons, Inc., 1994.

Back to Top