The following commentary comes from an independent investor or market observer as part of TheStreet's guest contributor program, which is separate from the company's news coverage.
NEW YORK (
) -- We have often been asked why we do not use the MACD indicator more in our analysis, and instead use the stochastic oscillator.
One reason for this question is that the MACD indicator appears on the default chart of most of the software packages. The other is that the stochastic is an older indicator, and not as widely understood. In order to fully understand my rationale for this we first need to discuss some features of oscillators and also some definitions.
First, there are two key types of oscillator. One is bounded, and the other is unbounded. A bounded oscillator generally moves between two points, most often "0" and "100". The Stochastic is a bounded oscillator, and another example of a bounded oscillator is Welles Wilder's RSI. The MACD is an unbounded oscillator, which means the rank can keep moving in a direction for as long as a trend lasts.
Let's look at the MACD indicator. MACD stands for Moving Average Convergence Divergence. This means that essentially the indicator measures the difference between moving averages. This is interesting because most people use this indicator as an overbought/oversold indicator when it is actually more of an indicator for strength of trend.
This is one of the main problems with the indicator for me, and I prefer to look at the moving averages themselves to evaluate trend strength. Look at the chart of the SPY with our 5- and 20-day moving averages, and the MACD (created with the 5- and 20-period averages) below. You can see how the MACD has crossed negative while the moving averages themselves have stayed positive. This is another problem with the indicator--- the mathematics of the indicator can cause it to give a signal when, in my opinion, none should be given.
Now, let's discuss the Stochastic. At The FRED Report, we use the Stochastic to find buy and sell points in the trend, or alternatively to identify areas of risk and reward within the trend. This indicator is a pure overbought/oversold indicator. There are two lines that comprise the indicator -- %K and %D.
One of the reasons this is my favorite of the standard overbought/oversold indicators is that %K is derived from an equation that analyzes the whole trading range of the look-back period. Most indicators, including the MACD, only look at the closing price, and I believe that analyzing the entire trading range of prices is important, especially during periods of volatility.
In the chart of the SPY and Stochastic, below, notice how the Stochastic signal is clearly defined, i.e. goes below, and then above 20 on the buy side, or conversely above and below 80 on the sell side. You can see the horizontal lines on the chart at 20 and 80. This is a clear and unambiguous signal.
The stochastic is not a trend following indicator, the moving averages on the chart take care of that, and they cross over soon after, CONFIRMING the stochastic signal. Here at
, the idea of confirmation is vital.
Also important is that one can see that price bottomed at a higher low on the same sort of stochastic signal (point "B" vs. point "A"). This provides added support for the idea of a price uptrend. When moving averages and stochastics are used together it can be a powerful tool, providing new information at several stages in the market cycle.
Now, scroll up to the chart of the SPY and the MACD. Notice how buy points came in at similar spots. Yet, without a signal line (like the 20% line on stochastics) it is much more difficult to tell how low is low, or how high is high. Indeed, from a purely "oversold" perspective the buy signal in June at point "A" looks better than the signal at "B", when in fact the opposite is the case.
The stochastic pattern shown above was one reason why we had our real buy signal in July, by the way. The other problem with not having a signal line is that in order to use the indicator one must take all crossover signals - and notice how early the signal near point "C" has been so far.
Also notice that the moving averages remained positive (5 above 20), and the market continued to advance, even as the MACD went negative at point "C". This is another reason why I prefer to look at the moving averages themselves, rather than a derivative calculation.
To summarize and conclude, we prefer the stochastic in conjunction with our favorite moving averages, the daily 5 and 20, because it measures something different than moving averages, yet can give an idea of where you are in the cycle of a trend.
The MACD is so related to moving averages (it is basically the difference between the 5- and 20-period moving averages in this example), that it really does not add much to moving average analysis. In addition, because it is an unbounded oscillator it is difficult to say how high is high and how low is low a move is within the context of the trend.
Fred Meissner is founder and publisher of
. Fred is a CMT and past President of the Market Technicians Association (MTA). He recently left Merrill Lynch's Market Analysis Department and Sector Strategy Department to form The Fred Report. A detailed bio is here: