The Bullwhip Effect

Tallow? Isn't that what candles were made of in the 18th century? How can the price of tallow, the rendered fat of cattle or sheep, signal whether the modern global industrial economy is trending up or down? You may be surprised by the answer.

It turns out that in this day and age, tallow is an important raw material in the production of soaps, detergents and a wide variety of chemicals. The price of tallow, a raw material in industrial production, is very sensitive to changes in demand.

That's why tallow is one of the 18 industrial commodities that are tracked by the index that is compiled every weekday by The Journal of Commerce and the Economic Cycle Research Institute. The JoCECRI Industrial Price Index has a remarkable record as a barometer that can forecast broad trends in the world's industrial economy. When he was Federal Reserve chairman, Alan Greenspan checked it daily as a signal of changes in demand for industrial raw materials.

The economic theory behind the index is called the "bullwhip effect," for lack of a more academic title. "This index is designed to be a leading indicator; if demand starts to slow, the index starts to turn down in advance," said Anirvan Banerji, director of research at ECRI.

It is based on an economic theory first explained 50 years ago in a study of the cycle of shoe leather and hides by Ruth Mack, an economist with the National Bureau of Economic Research. "There is a very simple reason for the way commodities like shoe leather behave," Banerji said. When Mack did her study, shoes were expensive products that the average consumer would only replace in good economic times. They were not an impulse buy, as they often are today.

"If consumers felt things were a little tight, then they would get their shoes repaired and postpone the purchase, so that meant that shoes were moderately cyclical in terms of demand," Banerji said.

Shoemakers caught with too much inventory of shoe leather when demand started to slow would cut back on production and on orders for leather, which is made from cattle hides. But even as the demand for shoe leather slowed, the production of beef did not because it continued to be eaten by consumers, and so the surplus of hides continued to pile up at the slaughterhouses because they were a byproduct of beef production. The result? Hide prices plunged, Mack's study showed.

The study showed how small shifts in demand growth at the consumer level are amplified through the supply chain into big swings in price at the wholesale level. "It's called the bullwhip effect because a little flick of the whip at the wrist will produce a big arc at the end of the whip," Banerji said. "This bullwhip effect is responsible for the big swings in commodity price, but it also gives us a tool because demand shifts are manifested in the movement of prices of commodities where the supply doesn't change very much in the short term. This is the information we use to form an early-warning indicator."

The JoC-ECRI Industrial Price Index is a predictor of broad trends in industrial demand. When charted against the Baltic Exchange Index of dry bulk freight rates over the period from 1985 to 2002, the JoCECRI index has proved to be an accurate indicator of the trend in those rates (See chart, Page 18). It has foreshadowed up- or downturns in the Baltic index by an average lead time of eight months.

In the years since 2002, the Baltic Exchange Index has been skewed by what Banerji calls the "once in-a-lifetime" secular trend of rising demand from China, which has overwhelmed some of the smaller cyclical fluctuations in the Baltic index that could have been predicted by the JoC-ECRI index. That has not affected the ability of the commodity price index to predict underlying changes in overall demand, even though Chinese demand has tended to offset any downturn elsewhere.

The JoC-ECRI Industrial Price Index was created in 1985 by Geoffrey H. Moore, the renowned economist who pioneered the study of business cycles at NBER, where a committee of economists that he established is still the official judge of when the U.S. economy goes into and out of a recession.

Banerji and ECRI's managing director, Lakshman Achuthan, worked for Moore at NBER, which was then part of Columbia University. In the 1980s, The Journal of Commerce, which had long maintained its own commodity price index, asked Moore to update the index, which at the time was still using 1967 as the base year for its prices. Moore and his team, including Banerji and Achuthan, developed a new index of prices based on a rotating group of 18 commodities that was launched in 1986. Moore and his team of researchers left NBER in 1996 following a change in its direction by Columbia University and founded the Economic Cycle Research Institute.

The index is updated every few years to adjust the weights of key industrial commodities to reflect their importance in today's U.S. economy and prune those that no longer play an important role. It was revamped most recently on Jan. 2. The new index eliminated polyester, whose use in the U.S. has plunged, and added the price of natural gas, an increasingly important energy source. The index has been revamped twice previously, in 2000 and 2004, since it was introduced in September 1986.

The JoC-ECRI index differs from other commodity indexes in several ways. It is not a weighted average, but a pure reflection of prices. It does not include agricultural commodities or precious metals such as gold or silver, but only materials that are used in industrial production, such as nickel, tin, aluminum, plywood, benzene, cotton, burlap and crude oil.

In addition, half of the commodities in the index are not "exchange-traded" commodities, which means that their prices are not established on the floors of any of the world's commodity exchanges. This serves as a reality check on the prices of the other half, so that the index cannot be skewed if a hedge fund or a speculator tries to manipulate the price of a commodity, as the Hunt brothers tried to do in 1980 when they bought half of the world's deliverable supply of silver.

"Why would an index like this, which is only updated once a day, and is of no use in futures trading, survive? Because it is structured as a composite leading indicator, not as a weighted average, which is what all the others are," Banerji said. "It is a leading indicator of the turning points in demand. It leads those turning points 94 percent of the time."