Commodity prices have emerged as a symbol of inflationary expectations, a management tool for the Federal Reserve Board in setting monetary policy and a guide for traders to be long or short the bond market.
While economists and investors have become preoccupied with their favorite commodity index, the analytical gap between commodity price movements and subsequent changes in the Consumer Price Index seemingly has not been bridged. In view of the compositional differences and biases of the various commodity indexes, how accurate is any one index in predicting price inflation?
Some might argue that merely monitoring the current trend in commodity indexes is sufficient to get a sense of inflationary expectations. Another approach would be to use commodity price trends to forecast the CPI and then compare current bond market expectations with the projected CPI trend. Before undertaking such a comparison, however, we must first evaluate the statistical relevance of traditional commodity indexes to future CPI experience and, then explore the value added by using a new index, The Furman Selz Commodities Composite.
The statistical correlation between various commodities indexes and the CPI is strong enough to warrant consideration.
Because of compositional differences, certain indexes are better indicators than others. For the 1984-1993 period, we derived a correlation between four major commodities indexes and the CPI, on the premise that commodity prices lead the CPI by about one year. Percent changes were measured year-over-year for both the commodities indexes and the CPI. The four indexes examined were: Goldman Sachs Commodity Index; Dow Jones Spot Commodity Index; Commodity Research Bureau Spot Index; and Journal of Commerce Industrial Index.
The Goldman Sachs Commodity Index has the best statistical correlation with the CPI. The index explains 72% of the change in consumer prices a year later. The index is biased toward energy with a 48% weighting, causing it to fluctuate with wide swings in energy prices. On balance, however, this index produces a more accurate representation of each commodity's importance.
The Dow Jones Spot Commodity Index, with a 65% correlation to the CPI, is a close runner-up to the Goldman index. The Dow Jones Spot Index, which is published daily in the Wall Street Journal, is an equal-weighted index of 12 representative commodities, giving the index a uniquely neutral weighting characteristic. Its primary shortfall, however, is the omission of energy prices.
Perhaps the most widely followed commodities index is that of the Commodity Research Bureau, which ranked third in our analysis with a 60% correlation factor. The CRB index consists of 21 commodities, but it is biased heavily toward agriculture, which has a 48% weighting. Soybeans in particular are overweighted.
Finally, the Journal of Commerce Index is, in our opinion, the least reliable gauge of future inflation. This index of 18 commodities explains only 51% of the movement in the CPI in a year hence. Underlying the JOC's poor predictive properties is its overweighting of base metals, underweighting of energy and exclusion of food.
In summary, the traditional commodities indexes have built-in biases and statistical shortcomings. Although conclusions as to the direction of inflation can sometimes be drawn by scrutinizing a particular index, inflation signals of the various indexes are neither uniform nor coincident. In addition, precise CPI forecasts, as well as quarterly modeling, are rendered less creditable.
In an attempt to neutralize biases inherent in the various commodities indexes and produce one index with good statistical and predictive properties, we combined components of the indexes to create what we term the Commodities Composite.
The analytical procedure was identical to that used for individual indexes, with the composite lagging the CPI by one year during the 1984-1993 period. The composite not only resulted in an improved predictive capability, but also helped quarterly modeling, which is beneficial in projecting trends and detecting inflation turning points. The annual model explains 89% of the change in the CPI (vs. 72% for the Goldman Index) and displays better correlation results overall. The quarterly model, which covered the period from the first quarter of 1984 to the first quarter of 1994, explains 64% of the CPI's movement. The Commodities Composite/CPI relationship is depicted in the accompanying chart for the annual series.
What insights does the composite offer regarding the future course of inflation and interest rates? In the short run, reported CPI numbers are expected to be generally benign. This assertion is predicated on meager composite growth of 0.1%, on average, in 1993. Specifically, the CPI, which averaged 2.6% higher through October vs. the year-ago period, should be up about 2.8%, on average this year. Importantly, however, trailing 12-month and quarterly composite changes point to a CPI rate of change of 4% by mid-1995. (Third-quarter composite growth was a robust 11.5% year-over-year). Moreover, the pattern portrayed by the Commodities Composite is a creeping inflation trend that will be both gradual and protracted. This suggests there will be no sharp turning point in inflation, but rather, a tedious twisting of the knife.
For both bond and stock investors, a legitimate question might be, "Are inflation fears overdone relative to prevailing bond yields?" One way to answer this question would be to compare the projected CPI with current bond market expectations. It would seem inflation fears are not overblown and that long bond yields eventually would have to reach 8.25% to reflect the mid-1995 4% inflation projection proposed above, as well as tighter Fed policy.
To the extent monetary policy decisions are guided by commodity prices, the short end of the yield curve apparently is subject to further upward adjustment as well. Based on the rise in commodity prices experienced thus far in 1994, the Fed likely will target Fed funds at 6.25% during the next six months or so.
Given the volatile nature of commodities prices, the foregoing analysis requires ongoing evaluation to, it is hoped, stay ahead of the inflation curve.