As part of a series looking at technical/momentum indicators, today we're going to look at MACD.
Developed by Gerald Appel (publisher of Systems and Forecasts) in the late seventies, the rather grand-sounding "Moving Average Convergence-Divergence (MACD) indicator" is actually one of the most commonly used momentum indicators around. It is used to spot changes in the strength, direction, momentum, and duration of a trend in a stock's price.
What is MACD?
The MACD is just the difference between a 26-day and 12-day exponential moving average of closing prices (an exponential moving average or EMA is one where more weight is given to the latest data). A 9-day EMA, called the "signal" (or "trigger") line is plotted on top of the MACD to show buy/sell opportunities.
Why is the MACD useful?
The reason that traders pay attention to varying lengths of moving averages is because they want to figure out how the short-term momentum is changing relative to the longer-term momentum. If the short-term average rises faster (slower) than the long-term average, the MACD moves upward (downward). Traders use this to suggests that the buying pressure is increasing (decreasing).
One of the reasons that MACD is so popular is because its trading signals are fairly unambiguous. You can see a good video introduction to its usage here but, in summary, there are three popular ways to use the MACD: crossovers (signal line or centre-line), overbought/oversold conditions, and divergences.
a) Signal Line Crossover - The basic MACD trading rule is to sell when the "slow line" of the MACD falls below the faster 9-Day EMA line (known as a "signal line crossover") and similarly, a buy signal occurs when the MACD rises above its signal line.
b) Centre Line Crossover - It is also popular to buy/sell when the MACD goes above/below zero (known as a "centreline crossover"). Great momentum stocks stay above zero for a long period of time. Cross overs below zero are usually be ignored since the stock is weak and it is said that trends can not be predicted as easily.
Crossovers happen quite frequently. Another signal, considered more reliable but less frequent, is a pattern called bullish convergence. This is where the market price itself makes a lower low from a previous low but the underlying MACD pattern makes a higher low. This indicates that the low is weak or a “false” bottom and can resort to a turn around for a price reversal. In his book, Trading for a Living, Alexander Elder calls price divergence with the MACD, “the strongest signal in technical analysis.” A bullish divergence occurs when the MACD is making new highs while prices fail to reach new highs. Both of these divergences are most significant when they occur at relatively overbought/oversold levels. Even here, however, false signals are commonplace ("a divergence takeout"). Prices might have several final bursts up or down that trigger stops and force traders out of position just before the move actually makes a sustained turn and the trade becomes profitable.
3. Overbought/Oversold Conditions
Generally, the MACD is not used as an overbought/oversold indicator because the MACD is unbounded (the MAs can continue to diverge). When the shorter moving average pulls away dramatically from the longer moving average (i.e., the MACD rises), it might be argued that the price may be overextending. However, it is more common to use the MACD in combination with range-bound indicators like the RSI or the Williams %R.
Problems with MACD
One issue is that the MACD is a lagging indicator. Because it uses moving averages and moving averages lag price, the signal it provides can come late (the price may have reached the reversal point already before the entry signal is generated). The same can be true for exiting positions.
To try to address this, MACD charting is usually accompanied by the MACD histogram. Developed by Thomas Aspray in 1986 to anticipate signal line crossovers in MACD, this is a visual representation of the difference between the MACD and the signal line (the nine-day EMA). The histogram is positive when the MACD is above its nine-day EMA (and negative when the MACD is below). Bullish or bearish divergences in the MACD-Histogram can alert chartists to an imminent signal line crossover in MACD.
Another important point to remember is that the MACD is regarded as most effective in wide-swinging trading markets, just like any moving average crossover. Its weakness is that, when the market is trendless, the MACD tends to generate lots of false / unprofitable buy and sell signals.
Does MACD Trading work?
As with most technical indicators, there is some evidence that it does (but not much!). Terence Chong of the University of Hong Kongapparently tested MACD over a 60 year period on the UK stock market and found evidence of it being able to produce higher returns than a buy and-hold strategy.
However, Albin, Gunter and Kai did some comprehensive testing of the MACD signal for NASDAQ-100 stocks over a 10-year period. They found that it had a surprisingly low success rate of 32.73%, i.e. worse than blind chance and almost a contrarian indicator! This was mainly because its short-term nature lead to traders being whipsawed in and out of a position several times before being able to capture a strong price movement. They did come up with several refinements, known as MACD R1 & MACDR2, which substantially improved the odds.
MACD R1 - This attempts to ignore buy and sell signal if crossovers are too intensive in a short period of time. To accomplish this, the trading signal is only given three days after the crossover, provided that no other crossing has appeared in between. To overcome the lack of e a timely exit warning, they also set a predetermined target at 3% or 5% for profit taking (although the flip side of setting a predetermined profit level is that a system might miss out the greater profit potential in a steady long trend).
MACD R2 - This uses the same trading rules from MACD R1 but, in addition, there must be a pre-determined "trigger level" difference between the signal line and MACD which is greater than a percentage of the stock price. Their work tested crossing levels from 0.5% to 3.5% (above 3.5%, hardly any trading signals occur). Gunter et al claimed that with this extra factor, MACD R2 is able to capture a more significant trend in the beginning by avoiding random movement in a narrow trading range. Unfortunately, a subsequent paper, "A Test of the MACD Strategy " by Huang found that, with these adjustments, MACD trading could outperform buy-and-hold if trading costs were ignored - but the results did not hold once they were included.
How can I use this filter?
- MACD Bullish Divergence (Bearish Divergence)
- MACD Crossed Above / Below Signal Line
- MACD Crossed Crossed Above / Below Zero
- MACD-Histogram Turns Positive (& MACD negative to insure this upturn occurs after a pullback).
- MACD-Histogram Turns Negative (& MACD positive to insure this downturn occurs after a bounce).
- MACD histogram greater than 2 (or less than -2)
- MACD Histogram Consecutive day movement
These events would usually be considered in light of other confirmatory signals, eg. stocks trading above their 200-day moving average, which implies an uptrend overall.
Opinions are sharply divided on the value (or otherwise) of MACD. Anecdotally, traders do suggest that it is useful for trading trends, because it shows the (increasing and slowing) momentum of a movement. Jeff Hochman, Fidelity Investments notes that, "when I see a monthly MACD crossover that does not occur very often, say every few years, I pay attention" while Robert Colby in his “Encyclopedia of Technical Market Indicators” says that MACD backtesting shows that the indicator is most effective with longer-term trading strategies but over the short-term, is not profitable.
Overall, there is relatively little hard evidence that MACD trading on its own can provide reliable alpha, as compared with say the value effect. In part, this may be because of acknowledged deficiencies with a mechanical application of the approach, i.e. as a lagging & short-term nature of the MACD indicator, it can give false triggers, i.e. unprofitable buy and sell signals, especially when the markets / securities are going sideways. Whether modifications of the approach can effectively correct for this is open to debate. It may be that MACD is more effective when used in conjunction with other technical indicators to ensure a more accurate idea about a stock's direction, but again, we've not seen any decent empirical studies to support this claim.
- How well do traditional momentum indicators work?
- Wikipedia on MACD
- MACD - How to reduce false signals
Filed Under: Technical Analysis,