CHART 26.10 Columbia Energy, 1999–2000, and a Volume Oscillator
It is fairly obvious by now that volume oscillators move between bands of extremes just as price momentum does, but with one important difference.
When a price index is overextended on the upside, it usually indicates an overbought market, with the implication that a reversal to the downside is due. An unusual rise in activity is also associated with an imminent trend reversal, but it is essential to relate the movement of the volume oscillator to the prevailing movement in price.
Apart from this important difference, volume oscillators should be interpreted in the same way as price oscillators. One of the strong points of the volume oscillator is that reversals from overbought readings often signal exhaustion, either from the buy or sell side. Other indicators should then be consulted to make sure that the volume action is being confirmed.
One possibility is the use of volume with a price oscillator. This relationship is not always exact. If volume and price changes are not closely related for the stock or market being followed, this approach should not be used.
Chart 26.11 shows the Brazilian Index, the Bovespa.
CHART 26.11 The Brazilian Bovespa 2006–2008, a Price versus Volume Oscillator
At point A we see the higher price preceded by a series of declining volume peaks. These indicated vulnerability, which was finally confirmed when the price violated its dashed up trendline. I would have expected to see a more prolonged decline, but a selling climax developed at point B as an oversold price oscillator combined with an overbought volume oscillator. The price oscillator is not telling us much at point C, but the volume series is, as it violates a very significant down trendline. Since the price breaks to the upside, this turns out to be a bullish period as the index rallies to its final bull market peak. Things get a bit complicated in early September 2008 as the volume oscillator rallies above a down trendline. This indicated that volume was going to expand, and since the price broke below the horizontal support trendline at D, that indicated weaker prices. However, volume soon expanded to selling climax proportions at E and that should have been bullish. In fact, the price fell below the low of the selling climax very quickly and that meant that a lot of people were trapped at higher prices. While a selling climax does not necessarily mark the final bottom, there is usually an interregnum before prices register a new low. In this case, that did not happen and therefore provided a clue that things would get much worse.
The following are the main rules for interpreting the volume oscillator:
1. When the oscillator reaches an extreme and starts to reverse, it is indicating the potential for a reversal of the prevailing trend.
2. Volume oscillators occasionally lend themselves to trendline and pattern construction.
3. Expansion in price associated with a contraction in the volume oscillator is bearish.
4. An expansion in the oscillator associated with a contraction in price is bearish, except when volume reaches an extreme, in which case, a selling climax is usually signaled.
5. The volume oscillator usually leads the price oscillator.
Remember, this is by no means a perfect indicator; therefore, you should first make sure that it is acting consistently with the price trend being measured, and also that there is corroborating evidence from other indicators.
The Demand Index
The Demand Index was developed by Jim Sibbet as a method of simulating upside and downside volume for markets and stocks where such data are not generally available. It combines price and volume into one indicator with the objective of leading market turning points. The Demand Index is based on the premise that volume leads price, and is included in many charting packages. Unlike the volume ROC and volume oscillator, the Demand Index always moves in the same direction as the price, in that high readings indicate overbought conditions and vice versa. I find this to be a very useful indicator when used in the following ways:
1. A divergence between the indicator and the price points up an underlying strength or weakness, depending on whether it is a positive or negative one.
2. Overbought and oversold crossovers often generate good buy and sell signals. Since the level of the Demand Index is affected by the volatility of the security being monitored, optimum places to plot overbought and oversold levels will vary and should be determined on a case-by-case approach. That said, the + and –25 levels appear to be a good compromise for most markets.
3. The index sometimes forms price patterns and trendline violations, which normally represent a reliable advance warning of an impending price trend reversal.
Chart 26.12 featuring Citrix shows some interesting features. First, at A, both the price and demand index break trendlines for a good sell signal. Later on, the process is reversed as the Demand Index completes a base along with the price. The resulting rally can only be described as a whipsaw. Does this mean that the Demand Index is no good? No, this is a normal phenomenon in a bear market. What looks to be a very valid breakout turns out to be a whipsaw—it can happen with any indicator. The remedy is to try to get a fix on the direction of the main trend before any short-term analysis is attempted. Point C shows a positive divergence between the Demand Index and the price, and point D a small double trendline break. Finally, at E we see a breakout by the Demand Index from a reverse head-and-shoulders pattern, but it fails. Once again we need to use some common sense because the breakout came from an overbought condition. As we learned in the momentum section, breakouts from such high levels typically result in whipsaws.
CHART 26.12 Citrix Systems, 2000–2001, and a Demand Index
Chaikin Money Flow
The Chaikin Money Flow (CMF) indicator is based on the principle that rising prices should be accompanied by expanding volume and vice versa. The formula is as follows:
CMF = SUM (AD, n) / SUM(VOL, n)
where n = period
AD = VOL * (CL – OP) / (HI – LO)
where AD stands for accumulation distribution.
The formula emphasizes the fact that market strength is usually accompanied by prices closing in the upper half of their daily range with increasing volume. Likewise, market weakness is usually accompanied by prices closing in the lower half of their daily range with increasing volume. This indicator can be calculated with any time span—the longer the period, the more deliberate the swings. Money flow indicators calculated with a short-term time frame, such as 10 periods, are, therefore, much more volatile.
When prices consistently close in the upper half of their daily high/low range on increased volume for the period under consideration, then the indicator will be positive (i.e., above the zero line). Conversely, if prices consistently close in the lower half of their daily high/low range on increased volume, then the indicator will be negative (i.e., below the zero line).
It’s possible to construct overbought and oversold lines and use these as buy and sell alerts, but the indicator really comes into its own with divergence analysis. In Chart 26.13 featuring National Semiconductor we can see some good examples in practice. In early 1994, the Chaikin was falling sharply as the price ran up to its final peak, and at the time of the actual high was barely above the equilibrium line. This showed that the quality of the last few weeks of the rally left a lot to be desired. At the end of 1995, the divergence was more blatant since the indicator was barely able to rise above the zero line at a time when the price was making a new high. Both these examples were followed by long downtrends.
CHART 26.13 National Semiconductor, 1993–1997, and a CMF
Positive divergences also work quite well, as we can appreciate from observing the mid 1996 bottom. See how the price makes a marginal new low but the oscillator is hardly below zero. This compares to the late-1995 bottom, where it was at an extremely oversold condition. Of course, this is merely a positive momentum characteristic—we still need to witness some kind of trend reversal in price to confirm this event. Divergences are not uncommon
in momentum indicators. What sets the money flow indicator apart from the rest is that the divergences are usually far more blatant than, say, the relative strength indicator (RSI) or ROC series. As a result, it can provide clues of probable trend reversals that may not be apparent elsewhere.
One of the ways in which I like to use the indicator is to study trading ranges and then compare the price action to the oscillator to see if it is giving a clue as to the direction of the eventual breakout. Chart 26.14 shows American Business Products with a 20-period CMF. The price was experiencing a sideways trading range in 1987. During the formation of the rectangle, the money flow indicator violated an up trendline and diverged very negatively with the price in the September/October period. This combination indicated vulnerability, so it was not surprising that the price experienced a nasty decline.
CHART 26.14 American Business Products, 1986–1991, and a CMF
Later on, the price traced out another trading range, but this time the CMF rallied the moment the trading range started.
Choice of time span also influences the character of the indicator. A short one, such as 10 days, returns a volatile series, whereas the 45-day span used in Chart 26.15 for Reliance Communications, an Indian stock, offers a more deliberate path. In this instance, you can see how the indicator deteriorated substantially as the price peaked at A.
CHART 26.15 Reliance Communications, 2006–2009, and a CMF
Note the final few sessions formed a kind of trading range even though the price moved slightly higher. However, the Chaikin was deteriorating throughout this period until the price confirmed with a double trendline break. The opposite was true at B, where we see a series of positive divergences culminating with a joint trendline break to the upside.
Volume in the Stock Market
Upside/Downside Volume
Measures of upside/downside volume try to separate the volume in advancing and declining stocks. Using this technique makes it possible to determine in a subtle way whether distribution or accumulation is taking place. The concept sounds good, but in practice, volume momentum based on ROC or trend-deviation data appears to be more reliable.
The upside/downside volume data is published daily in The Wall Street Journal and weekly in Barron’s and provided by many data services. Upside/downside volume is measured basically in two ways.
The first is an index known as an upside/downside volume line. It is constructed by cumulating the difference between the daily plurality of the volume of advancing and declining stocks. Since an indicator of this type is always started from an arbitrary number, it is a good idea to begin with a fairly large one; otherwise, if the market declines sharply for a period, there is the possibility that, the upside/downside line will move into negative territory, which unduly complicates the calculation. If a starting total of 5,000 million shares is assumed, the line will be constructed as shown in Table 26.1.
These statistics are not published on a weekly or monthly basis, so longer-term analysis should be undertaken by recording the value of the line at the end of each Friday, or taking an average of Friday readings for a monthly plot. The appropriate MA can then be constructed from these weekly and monthly observations.
TABLE 26.1 Calculation of the Upside/Downside Volume Line
It is normal for the upside/downside line to rise during market advances and to fall during declines. When the line fails to confirm a new high (or low) in the price index, it warns of a potential trend reversal. The basic principles of trend determination discussed in Part I may be applied to the upside/downside line.
When a market is advancing in an irregular fashion, with successively higher rallies interrupted by a series of rising troughs, the upside/downside line should be doing the same. Such action indicates that the volume of advancing issues is expanding on rallies and contracting during declines. When this trend of the normal price/volume relationship is broken, a warning is given that one of two things is happening. Either upside volume is failing to expand sufficiently or volume during the decline has begun to expand excessively on the downside. Both are bearish factors. The upside/downside line is particularly useful when prices may be rising to new highs and overall volume is expanding. In such a case, if the volume of declining stocks is rising in relation to that of advancing stocks, it will show up either as a slower rate of advance in the upside/downside line or as an actual decline.
The upside/downside line from 1986 to 1987 is shown in Chart 26.16 together with its 200-day MA. For most of this period, the line remained above its MA despite some fairly large short-term corrections in the S&P Composite. In mid October it fell below its 200-day MA just before the crash. An important up trendline was violated earlier for both the price and the upside/downside volume line which had the effect of emphasizing the bearish signal.
CHART 26.16 S&P Composite, 1985–1987, and an Upside/Downside Volume Line
Also worth noting were the positive divergences that occurred in October 1986, when the S&P made a new short-term low, which was not confirmed by the cumulative upside/downside line.
Chart 26.17 compares the line to the S&P between 2010 and 2012.
CHART 26.17 S&P Composite, 2010–2012, and an Upside/Downside Volume Line
The dashed arrows represent negative divergences, where the line failed to confirm new S&P highs, and the dashed lines represent joint confirmation. There were no noticeable positive divergences during this period. Chart 26.18 shows the same period, but this time both series are plotted with a 100-day MA. Note how the most reliable crossovers, flagged with the large arrows, develop when both series confirm. The arrows at A, B, and C indicate where the S&P break was unconfirmed by the upside/downside line. C was a complete failure of this system as both series whipsawed.
CHART 26.18 S&P Composite, 2010–2012, and an Upside/Downside Volume Line
There are several ways to construct oscillators of this data. One possibility is to use the following formula:
Oscillator = M (U – D)/(U + D)
In this case, U = upside volume, D = downside volume, and M is the MA time span. This formula is plotted in the bottom panel of Chart 26.19 using a 30-day MA.
CHART 26.19 NYSE Composite, 2010–2012, and an Upside/Downside Volume Oscillator
Notice how this setup allows negative divergences and, more importantly, joint trendline breaks that signal both buy and sell signals. In Chart 26.20 we see a 30-day MA of the oscillator in the previous chart together with its 12-day MA plotted as a dashed line. This series comes more into its own as it reverses from an overextended level. The solid arrows flag successful signals, and the three dashed ones the two failed sells, as well as the failed buys signal in August 2011.
CHART 26.20 NYSE Composite, 2010–2012, and a Smoothed Upside/Downside Volume Oscillator
Another possibility is to plot an oscillator of up and down volume and overlay them. Chart 26.21 features such an exercise using two daily Know Sure Things (KSTs), but a stochastic, smoothed RSI, etc., could have served the purpose equally as well. The idea is that when the KST of upside volume crosses that for downside volume, a buy signal is triggered and vice versa. The various signals have been flagged in the chart by the up and down arrows. The dashed arrows indicate whipsaw signals. As usual, there is no way of knowing whether a signal will be accurate or not, and that is the reason why we need to monitor the price action for some kind of confirmation, as well as base the decision on the position of other indicators.
CHART 26.21 S&P Composite, 2000–2001, and Two Upside/Downside Volume Oscillators
Major Technical Principle When the upside/downside volume line fails to confirm a new high (or low) in the price index, it warns of a potential trend reversal.
The Arms Index
This indicator was developed by Richard Arms and is constructed from breadth and upside/downside volume data. It is sometimes referred to as the TRIN or MKDS. It is calculated by dividing the ratio of advancing and declining stocks by the ratio of volume in advancing issu
es over volume in declining issues. In almost all cases, daily data are used, but there is no reason why a weekly or even monthly series could not be constructed. Normally, the Arms Index is used in conjunction with NYSE data, but its principles can be applied to any market situation, such as the NASDAQ, where upside/downside volume and breadth data are available. There is one important thing to note, and that is that movements in the Arms Index are inverse to those of the market. This means that oversold conditions appear as peaks and overbought conditions as valleys. Since this is contrary to virtually all other indicators described in this book, the chart examples are presented inversely in order to be consistent with the other indicators.
The concept behind this indicator is to monitor the relative power of the volume associated with advancing issues to that of declining ones. Ideally, it is important to see a healthy amount of volume moving into rising issues relative to that associated with declining stocks. If this is not the case, the indicator will diverge negatively with the market average and vice versa.
This momentum series can be calculated for any period. For example, the quote services and the number appearing on the CNBC ticker represent an instant in time and are based on the volume and number of issues experiencing an up or down tick. Unless you are fortunate enough to be able to chart this indicator on a continuous basis through a real-time service, isolated quotes of this nature are limited to gauging whether the market is intraday overbought or oversold. In this respect, 120 or higher is regarded as oversold and 50 or below overbought (remember, these numbers are inverse to the other momentum indicators).
Technical Analysis Explained Page 38