THE COMMODITY CHANNEL INDEX
Writing in Commodities Magazine back in 1980, technical analyst
Donald Lambert observed that:
Many commodities exhibit some type of cyclical or seasonal price pattern. But the commodity trade still faces the problem of detecting when the regular price movements begin and end because climate and other real world conditions may affect their timing.To be useful in cyclical markets, an index must examine current prices in light of past prices but must not allow data from the distant past to confuse present patterns ...
The comparison of current prices to moving averages solves one problem by providing a moving reference point. But it leaves another problem for the trader: While some commodities typically move only a few cents each day, daily moves in others might be hundreds of cents. Rather than develop separate rules to determine each commodity's fluctuations, some standardization technique had to be found.
It was the commodity channel index, or CCI, that turned out to be the object of Donald Lambert's quest. Now a technical indicator found on virtually all technical analysis charting software packages, the CCI was developed during the height of commodity speculation in 1980. Originally intended for use with cycle analysis, the CCI has survived and prospered as a tool for technicians and chartists, whether or not they are deploying the indicator for the purpose of following cycles.
How is the CCI used? Lambert notes in his article that "The CCI doesn't calculate cycle lengths ... but is a timing tool that works best with seasonal or cyclical contracts." Here, he was attempting to use the CCI to confirm that cyclical highs and lows had been established. In best-case scenarios, the CCI peaks and troughs confirmed cyclical peaks and troughs in price action. Some optimizing led Lambert to conclude that "the data base should be less than one third of the cycle length to produce a reasonable level of theoretical efficiency."
It is telling that contemporary discussions of the CCI tend not to mention the relationship between the indicator and cycles. The boom in commodities - which often had clear, cyclical tendencies - that took place in the 1970s also encouraged a familiarity with and appreciation of cycles in price trends. With the rise of the secular bull market in stocks in the late 1970s and early 1980s, cyclical analysis slipped into disfavor - at least among the mainstream - in favor of what eventually became the momentum and "buy and hold" schools.
As part of this process, the CCI has been almost reinvented as a trend-following indicator. This reinvention is based, in part, on an examination of what really happens in trends, as Elli Gifford points out in The Investor's Guide To Technical Analysis: "... [R]eally fast markets get overbought - only to get even more overbought." And though writing about price channels and envelopes, author Alexander Elder notes that the CCI and channels do more or less the same work in measuring deviation from a moving average. Thus, when Elder notes that "an upside penetration of the upper channel line shows very strong bullish momentum," this notion is not too far removed from saying, as Lambert does, "If the CCI goes above the +100% [channel] line, that's a signal to establish a long position."
Before going too far afield, what is the "+100" level anyway? And how is the CCI constructed?
The calculations for the CCI are relatively complex - fortunately, most chart analysis software packages feature the CCI and traders are not left to calculate this indicator on their own. That said, a copy of Donald Lambert's "original recipe" can be found in the nearby sidebar. For now, suffice it to say that the CCI is constructed by developing a "typical price," which consists of the period's high, low, and close divided by three. Next, a simple moving average of n-periods is constructed using typical prices.
From here, things get a little tricky. In Steven Achelis' summary, from his book, Technical Analysis From A To Z, he writes:
For each of the prior n-periods, subtract today's Step 2 value from Step 1's value n days ago. For example, if you were calculating a five-day CCI, you would perform five subtractions using today's Step 2 value.Calculate an n-period simple moving average of the absolute values of each of the results in Step 3.
Got it? Lambert's formula then calls for a multiplier of 0.015 times the value in the fourth step, subtracting the value derived in step 2 from the value derived in step 1, and then dividing the value in the last step by the value in the step previous (the multiplier step).
Let's move into an analysis of how the CCI is used. The multiplier is used so that 70?80% of the values will be within the boundaries of a +100 % to -100% channel. This is one of the ways that the CCI resembles the use of a price channel. Beyond this, there are a number of different ways traders, investors, and market analysts have used the CCI. Lambert originally suggested using a crossing of the +100% line to signal a long opportunity, with a subsequent crossing (this time to the downside) registering a signal to exit the position. Conversely, a crossing of the -100% line signaled a short opportunity with a subsequent crossing (this time to the upside), noting that the short position should be covered.
Since Lambert's original formulation, others have deployed the CCI somewhat differently. Achelis notes that the CCI can be used to spot divergences between prices and the CCI, such as when prices make a new low while the indicator fails to do so. Somewhat unnervingly, Achelis also points out that the CCI can be used as an overbought/oversold oscillator, with values over +100% registering as overbought and values under -100% registering as oversold.
For those familiar with Lambert's original use of the CCI, this last approach is almost a direct contradiction, but this is a point that really cannot be resolved. As Gifford noted, markets that are overbought often become only more overbought - thus selling a market that appeared overbought because a CCI value over +100% could be a mistake.
At the same time, everyone who has ever looked at a price chart knows that markets that are topping and ripe for reversal tend to do so from an overbought condition - one that would be reflected by the CCI with a value in excess of +100%. Yet another variation on this theme is the one provided by Richard Bauer and Julie Dahlquist in Technical Market Indicators: Analysis And Performance. There, the authors suggest that "a long signal is generated whenever the CCI goes below -100 and turns upward. Whenever the CCI goes above +100 and then turns downward, a bearish short signal is generated."
Let's look at a few examples, keeping each of these uses of the CCI in mind. Figure 1 is of the July crude oil futures contract, a closely watched commodity in early June. Using Lambert's original formulation, which called for a long position when the CCI rose above +100 and an exit when the CCI moved back below -100, we can spot a number of opportunities that the indicator would have pointed to.
Figure 1: The Bauer-Dahlquist variation is an excellent complement to Lambert's original methods for using the CCI.
One such opportunity occurred in early January. The CCI rose to +163.72 by the close of trading on January 5, with July crude at $31.26. When the CCI moved back under +100 on January 14, the closing price for July crude was $31.68, a gain of 42 cents. During the bull move in late February and March, the CCI would have had the trader moving in and out of positions at least three (if not four) times, rather than simply buying on the first signal and riding the trend. And while it is true that a trader who bought on the first CCI buy signal of the period on February 17 and held until the last sell/exit signal of the period on March 22 would have done better than the trader who moved in and out with each signal, the CCI trader would not have fared badly in the end (approximately 1.71 points compared to a buy and hold approach that would have netted 2.5 points).
Using the same chart, let's look at the other two strategies I've mentioned for using the CCI. The divergence method discussed in the Achelis book would have noted two significant negative divergences - one in early January and the other in early March - that would have alerted traders to a pair of corrections. Perhaps what is most unfortunate about his approach in the example of July crude presented is that, at the end of those corrections, there was no positive divergence in the CCI to help traders get back into the same market that the previous negative divergence in the CCI warned them to avoid. With regard to the notion of using the CCI as an overbought/oversold oscillator, the previous examples involving Lambert's original methods cast a shadow on those prospects, insofar as it was precisely the arrival of an overbought condition per Achelis that was the profitable buy signal per Lambert!
Finally, let's take a look at the CCI variation provided by Bauer and Dahlquist. Here, Bauer and Dahlquist call for short positions when the CCI moves above +100 and falls back with long positions being in order when the CCI moves below -100 and bounces back. Looking at the chart of July crude, there are two long signals generated per this method: one at the end of January/beginning of February, and the other late in March. In both cases, assuming the entry rule called for entering above the high of the price bar during which the CCI makes its specific move, the following trades would have been quite profitable.
I should point out that this variation is most effective when used to take advantage of short-term weakness in a bull market or short-term strength in a bear market, even though the short signals generated in the otherwise bullish chart of July crude are moderately profitable (with a few scratches or nontrades).
David Penn is Technical Writer for STOCKS & COMMODITIES.
Learn more about the commodity channel index!
Achelis, Steven B. [2000]. Technical Analysis From A To Z, McGraw-Hill.
Bauer Jr., Richard, and Julie Dahlquist [1999]. Technical Market
Indicators: Analysis And Performance, John Wiley & Sons.
Bulkowski, Thomas [2003]. "The CCI Trade," Technical Analysis of
STOCKS & COMMODITIES, Volume 21: November.
Davies, D.W. [1995]. "Defining The Commodity Channel Index," Technical
Analysis of STOCKS & COMMODITIES, Volume 13: January.
Elder, Alexander [1993]. Trading For A Living, John Wiley &
Sons.
Gifford, Elli [1995]. The Investor's Guide To Technical Analysis,
Financial Times/Pitman Publishing.
Hartle, Thom [1993]. "Calculating An 11-Period CC," Technical Analysis
of STOCKS & COMMODITIES, Volume 11: September.
Hill, John R., George Pruitt, and Lundy Hill [2000]. The Ultimate
Trading Guide, John Wiley & Sons.
Lambert, Donald [1983]. "Commodity Channel Index: Tool For Trading
Cyclic Trends," reprinted from Commodities magazine, Technical
Analysis of STOCKS & COMMODITIES, Volume 1.
Schwarz, Jeanette [1995]. "The Commodity Channel Index Revisited,"
Technical Analysis of STOCKS & COMMODITIES, Volume 13: December.
Star, Barbara [1992]. "The Commodity Channel Index," Technical Analysis
of STOCKS & COMMODITIES, Volume 10: February.
Wu, Amy [2002]. "Commodity Channel Index," Working Money, April.
--eSignal
Originally published in the October 2004 issue of Technical Analysis of STOCKS & COMMODITIES magazine. All rights reserved. © Copyright 2004, Technical Analysis, Inc.
Return to October 2004 Contents