MACD
MACD
MACD is a featured indicator on virtually every computer-based technical trading program and trading platform.
MACD is, indeed, an indicator for all seasons. If monthly data is maintained, MACD may be employed for the analysis of longer term
trends. It may be applied to somewhat shorter time periods, reflected perhaps by weekly or daily data, for the analysis of
intermediate and shorter term market trends.
signals one of its strongest features, its ability to detect the conclusions of and favorable market entry junctures following serious
intermediate
market declines.
Strangely enough, in spite of its widespread use, relatively little has been written regarding MACD or of some of the best ways in
which MACD patterns may be maintained, approached and interpreted all of which we will be considering.
SCOPE OF DISCUSSION:
There will be many areas covered in our discussion of MACD.
moving average convergence-divergence indicator. The charts reviewed will represent varying periods of market history, varying
periods of time, various MACD combinations, and varying market climates. You will receive suggestions as to the best of buying,
holding and selling MACD patterns, the establishment of stop loss orders and, of course, how to apply MACD in synergy with other
market tools that you have learned.
Although our discussions will be focusing on MACD in particular, the concepts involved are
almost certainly applicable to the use of other measures of market momentum as well.
Chart 1
Chart 1 illustrates the construction and the basic underlying concepts of MACD.
The price scale is of the New York Stock Exchange Index, daily based.
averages a 12-day exponential moving average and a 26-day exponential moving average.
employed as an alternative, but exponential averages are easier to maintain and track trends more closely.
As the market declines, the shorter term average will decline more
If previous trends have been rising, the shorter term moving average may decline from
As declines continue, the shorter term average will decline further and further below the
longer term moving average cross, where they are equal, is the 0 line on the chart, the point at which shorter and longer term
trends are in probably temporary balance, the area in which relative strength relationships between short and longer term trends
often reverse.
The greater the positive distance between the short term exponential average and the longer term exponential
The greater the negative distance (short long) between the shorter and
longer term exponential averages, the more negative will be MACD readings.
indicate the differential, 12-day short term moving average minus 26-day longer term moving average.
Basic Concepts:
MACD represents the difference, short term exponential moving average minus long term exponential
average.
When market trends are improving, short term averages will rise more quickly than long term averages.
MACD lines will turn up.
When market trends are losing strength, shorter term averages will tend to flatten, ultimately falling below
longer term averages if declines continue. MACD lines will fall below 0.
Weakening trends are reflected in changes of direction of MACD readings, but clear trend reversals are not
usually considered as confirmed until other indications, which will be discussed, take place.
During the course of price movements, short term moving averages will move apart (diverge) and move
together (converge) with longer term moving averages hence, the indicator name, moving average
convergence divergence.
TREND CONFIRMATION:
MACD signals are more likely to prove reliable if shorter term MACD signals are confirmed by longer term trends in the stock market,
perhaps reflected by longer term MACD patterns.
For example, purchases made based on daily MACD lines are more likely to
succeed if weekly and/or monthly MACD patterns are favorable, indicating strength in the
sales are more likely to prove profitable if longer term market trends are down. The maintenance of multiple MACD charts, refl ecting
varying length of market cycle, is recommended.
Chart 2
Dow Industrials, 2000
Introducing the Signal Line
The signal line is an exponential average of MACD levels, not of the price of the investment
vehicle or index that is being tracked. Signal lines are usually created employing 3-day to
9-day exponential averages of MACD lines. The shorter the average the more sensitive will
be the signal line. This chart employs MACD lines based on differentials between 19 day
and 39-day exponential averages of the price levels of the Dow Industrials, and a signal line
which employs a 9-day exponential average of the MACD line.
Although changes in MACD direction (from down to up and vice versa) and crossings of MACD lines
above and below 0 carry significance, crosses of MACD from below to above its signal line (an
exponential average of MACD readings) and from above to below its signal line carry additional
significance of their own.
As a general rule, crossings of MACD from below to above its signal line may be taken as
confirmations of buy signals originally indicated when changes in direction have taken place
in MACD from down to up.
The above buy and sell signals are subject to further conditions, which will be shortly reviewed, but as a
basic rule, reversals in the direction of MACD carry signifi cance, confirmed when MACD crosses its
signal lines.
place after MACD lines change direction but usually before MACD lines have crossed the 0 line.
My research over the years has suggested that greater net gains generally occur if reversals in MACD (especially slower moving
combinations) are employed for buying and selling rather than crossings of signal lines.
MACD without signal line confirmation does produce larger numbers of trades, with attendant extra expenses.
on Chart 3 that, by and large, you would have secured slightly more favorable entries and exits during the period shown if you
had acted on changes in direction of MACD rather than awaiting signal line crossings but that crossings of the signal line did
produce more immediately productive signals than simple changes in MACD direction.
Very important supplementary buy and sell rules
Buy signals are much more reliable when MACD has crossed from above to below 0 at some time since the
most recent sell signal. MACD does not have to be below 0 at the time of the buy signal but should have
been below 0 at some time since the start of the recent decline.
Sell signals are more reliable when MACD has crossed from below to above 0 at some time since the most
recent buy signal. MACD does not have to be above 0 at the time of the sell signal but should have been
above 0 at some time since the start of the most recent advance.
During very strong market periods, usually during the early and best stages of bull markets, MACD will
retreat during market reactions, to a level just above 0, in which case you can shade the above rules a bit
as you might if MACD tops out just below 0 during a bear market or severe intermediate decline. Most
often, however, the 0 crossing condition should be respected.
Charts in this Chapter illustrate many occasions when MACD lines change direction and/or cross above and below their signal lines in
the process of tracing out declines from above to below 0 and from below to above 0, such crossings frequently of little
significance.
As a general rule, it is best to sell positions following market movements which take prices from oversold to
overbought levels and to buy new positions during periods that the market has become oversold, or at least somewhat extended to
the downside. By waiting for MACD to fall below 0 for buying and to rise above 0 for selling, you are setting up procedures by
which you buy weakness and sell strength, rather than buying and selling on every change in minor trend.
said to take place when price levels of the investment in question reach new highs but measures that refl ect upside
market momentum fail to reach new positive areas. Positive divergences are said to take place when price levels of
the investment in question fall to new lows for the cycle but measures that reflect downside market momentum fail to
decline to more negative areas.
The failure of measures of momentum to confirm new highs or new lows in price by failing to concurrently achieve new
peaks or new lows respectively indicates a diminishing of market thrust in its current direction. Divergences suggest a
more or less imminent market reversal ahead, often of at least intermediate signifi cance.
We can see on Chart 4 two peaks in the stock market between October and November.
of price and MACD have risen simultaneously.
of the New York Stock Exchange Index did, indeed, decline along with MACD, but the decline did not carry very far, was very
quickly reversed.
What happened, however, at Area 2?
The price level of the NYSE Composite Index advanced to a higher recovery high, but this
high in price was unconfirmed by MACD, which failed to reach a new high along with price.
price high was a negative divergence, suggestive of a more serious decline which, in fact, did take place.
The decline continued into Area 3. (You
following declines in both price and momentum to new MACD-price confirmed lows.
which time the NYSE Composite reached a new low area. What about MACD? At a time that price was falling to a new low,
MACD patterns failed to do so! A positive divergence hadvbdeveloped, suggesting conditions that favored a more extended
market recovery which, in fact, did take place.
A 6 19 day
exponential MACD is employed in this chart. (Different MACD parameters are employed in this chapter to illustrate the
appearance of short term, intermediate term and longer term MACD combinations. To some extent, this may appear to
contradict some of the rules of operation that I suggest be followed, but as a general rule differentials in entry and exit
levels are not considerable.)
Additional Examples
Chart 5 provides additional examples of positive and then negative divergence between price movement and market momentum.
An MACD buy signal at Area 6 was unaccompanied by a positive divergence so the failure of the subsequent advance to catch was not
all that unusual. However, a secondary buy signal at Area 7 was accompanied by a positive divergence note the confl icting trends
between price and momentum in the area and, therefore, better grounded. The subsequent advance proved very tradeable.
The advance essentially ended in Area 8 when MACD produced a well timed sell signal.
(slight new high) accompanied by a negative divergence (clear declining top in MACD) reinforced the likelihood of more signifi cant
market weakness.
The implications of this negative divergence were not, however, actualized for several weeks.
The bearish
portents of the negative divergence were, perhaps, a bit premature but it might be kept in mind that the onset of the year 2000 bear
market was only weeks away..
IMPROVING MACD SIGNALS BY EMPLOYING DIFFERENT MACD COMBINATIONS FOR BUYING AND FOR SELLING:
Chart 6 Different MACD Combinations for Buying and for Selling
More accurate MACD buy and sell signals can often be secured by using more sensitive MACD combinations for buying
purposes and less sensitive MACD combinations for selling purposes.
Two MACD combinations are often better than one
Inasmuch as stock prices do decline at more rapid rates than they rise, it is generally a good practice to employ at least two, and
perhaps even three, MACD combinations different combinations for buying and for selling.
Chart 6 employs three elements new to our discussion of MACD.
which we will employ to define market trend, actually a somewhat subjective exercise.
the prevailing market trend is at least neutral and possibly bullish if the 50-day moving average is either rising or essentially flat.
If
the 50-day average is clearly declining, then we will assume the market trend is down.
The chart has two scales in its lower half.
The upper scale represents an MACD line created with differentials between a 12-day
and a 26-day exponential moving average, along with a signal line, created via a 9-day exponential average of the MACD line.
The lower scale represents an MACD line, created from the differential between a 19-day and a 39-day exponential moving average.
This MACD is less sensitive to minor market fluctuations than the faster moving, 12 26 day MACD combination.
The use of other MACD combinations is very feasible, of course, but I have found that the use of the 12-26 day MACD
combination is generally fine for the generation of buying signals and that the use of the 19 39 day MACD
combination is generally fine for the generation of MACD sell signals, providing that the general market climate is
favorable to just moderately negative.
If market trends are very strong, then a very sensitive MACD combination, perhaps of 6 19 days, may be employed
for buying, providing more rapid entries than the 12 26 day combination. The 19 39 day combination remains as
the selling signal.
If market trends are very weak, then you would continue to buy on the 12 26 day combination but you might also
employ that combination for selling.
As you can see, however, the more sensitive 12 26 MACD combination traced out both a rising double bottom
formation (bullish) and a crossing from below to above the 0 line in advance of the slower moving 19 39 day MACD combination.
In this particular case, there was probably not much benefit to buying earlier than a few days later, but as a general rule the
earlier the buy signal at a market low the better, particularly when market trends are favorable.
confirm that the MACD combination employed for buying has crossed
Remember, however, to
Moving along now, into May, we can see that the 12 26 day combination crossed below its signal line in mid-May, producing a
premature sell signal compared to the slower moving 19 39 day combination.
levels were not significant between the faster and slower signals but we can see that the slower moving 19 39 day combination did
better in allowing the market advance to run its full course.
Remember, however, that the MACD line that you are using for selling must have crossed above 0 prior to generating a valid sell
signal.
To sum up the golden rules just one more time.
Maintain at least two MACD combinations a faster one(s) for buying, a slower one for selling.
When market trends are very positive, BUY VERY FAST! SELL SLOW!
You may employ the 6 19 combination for buying, or you may employ the somewhat more reliable 12 26 day combination.
The 19 39 day combination is employed for selling.
When market trends are neutral to somewhat positive, BUY FAST! SELL SLOW!
Use the 12 26 combination for buying. Use the 19 39 day combination for selling.
When market trends are clearly negative, you may BUY FAST! SELL FAST!
You may use the 12 26 unit MACD combination for both buying and selling, in which case you will sometimes be selling before the
slower moving 19-39 day MACD has crossed from below to above 0.
MACD lines should generally rise above 0 as a precondition for a sell to occur.
MACD during strong market uptrends (steadily rising moving averages, lasting uptrends in price trend, favorable
breadth, an ability of momentum indicators to rise to overbought levels etc.) -
This does not, by any means as we shall see, always take place,
The slower 19 39 day MACD combination did not produce a sell signal until the end of
advanced further, into the end of July, but actually a new buy signal (based on the more rapid 6 19
combination) did produce a re-entry just slightly above the level of the June exit.
enough timed, reflecting once again the benefit of using slower MACD combinations for selling.
trading against the trend, but, on the other hand, did not result in losses of consequence.
If you review the chart carefully, you may be able to identify all those areas
in which a more rapid sell, based upon the 12 26 day MACD combination, would have outperformed slower developing sell signals
shown that were based upon the 19 39 day MACD combination.
Remember, during downtrends harder to buy, easier to sell!
Taken all in all, the MACD indicator will tend to neither gain nor lose much when you are trading against a prevailing market trend,
but will allow you to lock in profits when you are trading in the direction of a prevailing trend.
positions with a certain assurance that losses are likely to be well contained even if price trends are unfavorable,
will be properly positioned as market trends change.
74.5
74.0
73.5
73.0
72.5
72.0
71.5
71.0
70.5
70.0
69.5
69.0
68.5
1.0
1.0
0.5
0.5
0.0
0.0
-0.5
-0.5
0.5
0.5
0.0
0.0
-0.5
-0.5
75.0
Modifying MACD rules to secure the most from strong market advances
Chart 9 Treasury Bonds, MACD and a Strong
Uptrend
During favorable market climates, if no negative divergences are present, you may often bypass the first sell signal
that develops following a market rally. However, the second sell signal DEFINITELY SHOULD be followed. In the
situation illustrated, the rule modification set back the sell signal from late September to late October.
We have already seen one instance where MACD patterns have produced premature sell signals, when basic buy sell rules are
followed. There is a sell rule modification that may be applied during favorable market climates which will frequently provide
superior exit levels during strongly uptrended market periods while still containing risk levels.
The stock market must be in an uptrend which may be defined by a rising 50-day moving average.
slow MACD combination is employed for selling.
Check, at the time that the first crossing of the sell MACD line from above to below its signal line takes
place, whether there have been any negative divergences, either in the MACD that is being used for
buying OR in the MACD that is being used for selling.
If there are no divergences MACD lines and price lines moving in conjunction AND if trends of the
market are favorable, prices above a rising 50-day moving average, then you can ignore the FIRST sell
signal generated by MACD. You should, however, take a second sell signal.
You may use, as a back-up exit strategy, a crossing of the price of your investment from above to below
the 50-day moving average which would probably take place AFTER the MACD sell signal that you did not
follow.
Reviewing Chart 9
We will now review Chart 9 to illustrate the conditions under which a fi rst MACD sell signal may be bypassed.
Starting on the
left of the chart, July 1989, we can see that MACD had recently generated a sell signal. Bond trends were generally favorable
at the time, so a 6 19 day MACD combination was used for buying, the familiar 19-39 day combination for selling.
Market entry supported by positive divergence An initial buy signal was actually generated in mid-July which aborted in early August neither buy nor sell noted on the
chart. A second buy signal in mid-August was well supported, by a positive divergence in the buying, 6 19 MACD, by a rising
Moving averages, MACD patterns confirm advance -Bond prices advanced into September, the 50-day moving average turning up, re-confirming the generally bullish climate.
Both the 6 19 and the 19 39 day MACD combinations reached peak areas along with the price of bonds, and both turned
down as the price decline got under way. A turndown of MACD along with prices does NOT constitute a negative
divergence, which takes place only when new price highs are unconfi rmed by MACD.
Initial sell signal not reinforced by any negative divergence -The price turndown was sufficient to produce an initial sell signal in late September when the slower MACD combination,
which had risen above 0 (Remember the stipulation?), crossed from above to below its signal line. However, inasmuch as
this was the first sell signal after the MACD buy, and inasmuch as the trend was up and inasmuch as there had
been no negative divergences, this sell signal might well have been ignored.
Secondary sell signal confirmed by negative divergence -As matters turned out, the bond market did rapidly recover, rising to a new high during October.
that time? Yes, indeed.
The new highs in October were not confirmed by MACD which did not make a new high along with
The second sell signal in October was, therefore followed (as are all second sells). As
In this case, prices did rise to new highs following the sell signal that
would have been followed, but in the end this did not result in any missed profit to speak of.
Use moving average as back-up stop signal -What if prices had not turned up quickly following that initial sell signal?
represented a court of last resort. You could have sold on a penetration of that average, as a back-up stop loss, reverting then
to the fast MACD as your re-entry trigger.
As you might surmise, this modification of sell rules draws upon the concept that sell signals that take place in the
absence of negative divergences tend to be less serious than sell signals that take place following negative divergences
hence, the assumption of some risk while still maintaining a stop-loss strategy as damage control.
In the absence of any negative divergence, and with the market in a general uptrend, short selling in this situation
would probably carry too high a level of risk to be warranted even if the 50-day moving average had been violated.
Inasmuch as the 19 39 day MACD selling combination did not rise above 0, no sell signal could be generated
based upon that MACD line. Therefore, we make use of a secondary sell rule, that is triggered when the buying
MACD combination declines, following a buy signal, to a low that is below the low that preceded the buy. The
sell-stop is not based upon price falling to below a previous low point. It is based upon momentum readings
(MACD) falling to below a previous low point.
MACD is a strong timing indicator but is, of course, imperfect. There are times when MACD entries will fail to produce any follow
through in the stock and other markets.
Chart 10 illustrates an MACD stop
that may be employed if a rise following buy signals fails to bring the slower, selling, MACD
brought prices below support levels (often taken as a stop-loss signal but one which we do not use with MACD).
As the decline
continued, the rapid 6-19 MACD combination fell to below the lows of late February early March, indicating that downside
momentum had increased to new high levels for the swing.
employed as the buy trigger, not the price pattern, created the stop-loss sell.
MACD stop-loss signals and exits are often followed by rather rapid recoveries in MACD, new buy signals, usually taking
place at roughly the same level as the sell signal.
Key point! -- MACD stop-loss signals are created by violations of support areas in the MACD indicator, not by
violations of price support. This modification of usual stop-loss strategies, often prevents
takes place when slower basing and bottom formations develop.
Chart 11 illustrates situations in which buy signals took place right in the area at which downward sloping trend lines were
penetrated and situations in which MACD sell signals were confirmed by declines in MACD down and through rising uptrend lines.
The combination of trend line violation and MACD penetration of signal lines is more powerful than either element alone.
Legend Under Chart: The stock market followed an intermediate cycle averaging approximately 110 days during
the period shown. MACD buy signals were nicely confirmed by cyclical projections, which may be based upon
momentum indicators such as MACD rather than price. T-formations matched very well with MACD sell signals
during the spring of 2001, were premature in their projections later that year.
Charts 12 and 13 illustrate the use of cyclical projections as confirmations of MACD buy and sell signals.
buy signals were well confirmed by cyclical projections during both periods, 1989 and 2000 2002.
The sell signal in May of 2001 fit the relevant T-formation (Chart 12) almost exactly but not the exact workout of the sell of
March 2002, which developed a week after the projection made by the T-formation.
Chart 13 is an
excellent example. If a market cycle takes longer than normal to complete, then the next cycle will frequently be
shorter, the average of the two defining the basic cyclical length of the cyclical wave.
in Chart 12, in which we can see how the development of a longer cycle between April September was followed
by the development of a relatively shorter market cyclical workout.
280
280
260
260
240
240
1.5
1.0
1.0
0.5
0.5
0.0
0.0 -0.5
-0.5 -1.0
-1.0
300
When MACD does NOT provide the most timely signals Chart 14 New York Stock Exchange Index, 1994 - 1995
MACD provided a fine market entry in December of 1994, carried through well into March of 1995, but then, along
with many other technical indicators, produced a premature sell signal, the advance continuing thereafter without
providing a formal opportunity for MACD re-entry.
Buy signals shown are based upon the 19 39 day MACD combination for purposes of chart clarity. Normally, a
somewhat more rapid MACD combination would be employed for this purpose.
MACD is a very powerful timing indicator, but like any indicator it does have its Achilles heel, in this case occasional diffi culty
in dealing with steadily trending, narrow channeled market advances or declines.
Chart 14 illustrates the type of formation that produces problems.
drew to a close. The initial sell signal at the end of February 1995 might have been bypassed there were no negative
divergences between MACD and price movement, but a subsequent sell signals, particularly those generated in June (which
were accompanied by a negative divergence), would generally be followed.
market declines. Prices decline steadily, quietly and persistently even as MACD rises the stock market very slow to respond to
the favorable implications of diminishing downside momentum.
I know of no ready way to deal with such situations, if decisions are based upon MACD patterns alone.
However, we have
already reviewed a number of tools that might well have identified strength in the stock market that may counteract daily based
MACD sell signals at certain times. For example, the 10-day ratio of new highs to the sum of new highs and
tracking above the 90% level, and/or weekly ratios of advances in relationship to the number of issues traded on the New York
Stock Exchange may be producing its own signals to remain in the stock market.
Money Management with MACD (and other indicators)
The stock market does advance approximately 75% of the time, shorter term sell signals, particularly during bull markets, more
often than not just as well ignored as followed on average.
This applies to daily based MACD sell signals as well, which are often
followed by relatively minor market declines and by re-entry buy signals that take place more or less in the area of
previous sell
signals.
In this regard, it is often better money management practice, to develop money management approaches that phase capital into
and out of the stock market rather than to attempt to take massive action on each and every buy or sell signal. Given the general
upside bias to the stock market, there are relatively few periods each year during which you are likely to want to be fully out of
the stock market, particularly if your portfolio includes a relatively large portion of lower volatility mutual funds that are leading in
relative strength.
An MACD configuration that suggests more active selling
The above notwithstanding, there is one MACD configuration that carries clearly bearish implications, not on each and every
signal, but based on extensive research into the history of the stock market, in the aggregate.
The 19 39 day MACD combination is employed for the stay out signal.
The 19 39 day MACD combination must be falling. If it turns up, the bearish implications are cancelled.
The 19 39 day MACD combination must also be below 0. Although the most dangerous periods for the
stock market develop when the 19 39 day MACD combination is still falling and has declined to slightly
above 0, you can, for practical purposes, take the 0 line as the crucial level that defines the completion
of the most reliable MACD sell signals.
Not every decline, even given the above conditions, will graduate into serious market damage, but on
average the sell signals described above are followed by further market decline, and should be respected.
Total timing results are likely prove slightly more efficient if all MACD sell signals are followed when they
occur, rather than waiting for the 19 39 unit MACD to decline to below 0, but at the cost of more
frequent trading, with its attendant expenses, and greater numbers of whipsaws.
One recommended money management strategy involves lightening up more rapidly with some long
positions on initial sell signals but to maintain at least some holdings until Full Scale MACD sell signals
take place.
MACD THROUGH THE YEARS LONG TERM, SHORT TERM, AND INTRA-DAY
We have pretty well covered the basic operating principles of MACD, and have discussed some ways that MACD patterns may be
integrated with other technical tools and approaches.
of MACD over decades of stock market history, employing weekly and monthly MACD patterns as well as daily, and even intra-day
patterns. This is truly an indicator that may be applied to virtually all time frames.
This is a weekly chart of the NYSE Index, 1982 1984. Familiar parameters are employed, translated from daily to
weekly data a 50-unit moving average, a 6 19 MACD combination for buying, a slower 19 39 MACD combination
for selling. (A 12 26 MACD combination might well have been employed at different times during this period as the
buy trigger.)
Following an aborted, but still profitable, buy signal during the spring of 1982, a weekly MACD based buy signal was generated in
July of 1982. Prices fell to below the purchase level in August, MACD falling right to its previous lows at the time (which could have
been taken as a sell signal) but repeat buy entries rapidly developed.
From August 1982 on, MACD performed flawlessly, both shorter term and longer term MACD combinations rapidly rising through the
0 level.
Negative divergences and crosses from above to below the signal line that took place in the short term, 6 19 week
MACD combination, did not produce sell signals inasmuch as the slower moving, longer term, MACD line advanced smoothly, until its
generation of a sell signal in July of 1983, finally signaling an end to a market advance that had lasted for nearly a full year!
This is
an excellent example of the benefits of using shorter term MACD combinations for buy signals and longer term MACD combinations
for sell signals.
An example of the MACD stop loss signal in action
The patterns of MACD and the price movement of the stock market during 1984 are instructive.
1984, stock prices drifted downwards until the start of August, moving to new lows as the spring and summer moved along.
However, MACD levels did NOT make a new low in the interim, therefore forestalling a stop-loss sell signal, the holding of positions
ultimately well vindicated by the late summer market rally.
There were two very fine positive divergences, which have been marked, that provided excellent
Intra-day price ranges are often fairly narrow so it is more important in day trading than
MACD AND MAJOR MARKET TRENDS Chart 17 MACD and Major Term Signals New York
Stock Exchange Composite, 1969 1980
Monthly based MACD patterns have had a fine, if imperfect, history of confirming reversals in major market trends, as a
general rule generating buy and sell signals approximately three months following actual major term peaks and lows
in the stock market. (During some of the period shown, the 12 26 month MACD combination would have been more
usually employed as the buy trigger than the 6 19 month combination illustrated.)
Monthly based, major term price and MACD charts generally reflect the usefulness of employing shorter term MACD
combinations for buying purposes and longer term MACD combinations for selling purposes. The 19 39 day MACD
combination did cross from above to below its signal line late in 1986, but in the absence of any negative divergences
in either the shorter term 6 19 day MACD pattern or the longer term 19 39 day pattern, the first penetration might
have been bypassed as a sell signal. It was actually touch and go as to whether monthly MACD was able to avoid the
1987 crash, though weekly MACD patterns did, indeed, clearly do so.
Monthly MACD patterns reflected the secular change that developed in the stock market between the late 1960s into the 1970s and
then into the 1980s.
The stock market was essentially neutral in its major trend throughout the 1970s, ending the decade pretty
Peaks in MACD patterns actually declined as the decade moved along, the nadir
for MACD taking place late in 1974 along with the lows of the stock market.
The stock market clearly strengthened as the 1970s came to an end, into the 1980s and ultimately the 1990s, the increasing
strength reflected in the ability of even the shortest term
a nearly six year period.
In fact, a certain liberty has been taken in the assumption of a 1984 buy signal inasmuch as the 6 19
month MACD combination never did decline to below the 0 level, generally a required pre-condition to buy signals.
There are times that stock market indicators, including the general principles involved in using MACD, produce conflicting signals and
when decisions have to be made based upon the what appears to be the weight of evidence.
in MACD, coupled with a cross of the MACD line above its signal line and just a narrow miss of reaching the 0 line, all within the
context of a rising 50-month moving average, would seem to have suggested suffi cient justifi cation
Weekly based MACD patterns, as well as monthly MACD patterns, provide meaningful information regarding longer term stock
market trends.
shorter and longer term MACDs coalesce often carry particular signifi cance.
THE AMAZING ABILITY OF MACD TO IDENTIFY SIGNIFICANT MARKET LOW POINTS FOLLOWING SEVERE STOCK
MARKET DECLINES Chart 19 New York Stock Exchange Index The February March 1980 Stock Market Decline
After producing a sell signal only three days following the significant market high in February of 1980, MACD remained
out of the stock market throughout the decline, re-entering only three days after the final lows had been established as
March came to a close.
MACD would be worth following even if only one of its abilities existed its ability to define market re-entry junctures following
serious intermediate and long term market declines.
We have already seen this ability as it relates to longer term market trends
in the ability of monthly based MACD charts to identify major bear market lows.
interest rates and rates of inflation, coupled with high levels of speculation in the precious metals sector, resulted in a severe stock
market decline which took the normally placid New York Stock Exchange Index down by approximately 20 percent in just six weeks.
And how well did MACD cope with these market gyrations?
the stock market in mid-February but did take place only three days following the ultimate peak.
of the stock market until late March.
days after the final market bottom.
MACD lines turned upwards on the first day of the recovery, crossed the signal line only three
It goes without saying that not every market cycle is so well tracked.
be observed that the success of MACD during this February cycle was not unusual.
Legend: A longer term weekly MACD turned up just one week following the final lows of a very serious market 1998
market decline, though the 19 39 week MACD did not cross its signal line until five weeks from those lows.
A sell
signal in May of 1998 was somewhat premature but the sell signal in June of 1999 did prove well timed.
Chart 20 illustrates the application of MACD, this time weekly based, to another intermediate market cycle.
longer term, 19 39 week MACD combination, normally used for selling. The use of the 12 26 week MACD for buying instead of the
39 would have produced a market re-entry one week sooner than the entry shown.
19
After a somewhat premature sell signal in May of 1998, the 19 39 week MACD remained out of the stock market through the fi nal
leg of the market advance, through the very severe summer decline of 1998, finally producing a buy signal in November, fi ve
weeks from the lows of the market decline.
(Daily based MACD patterns produced, during this period, one unusually poor and
ultimately stopped out buy signal in August of 1998, and subsequently, excellent repeat buy signals in September and October,
both just days following low points of a final double bottom formation which can be seen on the weekly chart as well.)
Weekly MACD
patterns would have kept investors in the stock market for seven months, capturing the best part of the
market advance.
A successful sequence of MACD buy and sell signals took place during 1990 and early 1991. This is a detailed chart
which illustrates many concepts relating to MACD, including positive and negative divergences, rising bottom
formations, wedges and T-formations. Readers might want to make copies of this chart and draw in chart formations
that have not been marked.
The 1990 1991 period presented many opportunities for active market trading.
of MACD, we will examine the price movement and MACD patterns of the period.
Initial rally at start of year -The stock market declined during January of 1990, an excellent 12 26 day MACD buy signal developing in early February.
MACD
patterns started strongly with the advance, rising bottoms developing in both shorter and longer term MACD lines, but between
March and April, negative divergences developed in the 12 26 day MACD pattern, warning of incipient market weakness.
Because negative divergences were in place, the sell signal produced in April by the longer term 19 39 day combination would be
followed.
Brief decline and well-timed market re-entry
The sell signal was, indeed, followed by a relatively short term though very tradeable market decline, sufficient to draw the 12 26
day buying MACD line to below 0, establishing a prerequisite for a new buy signal which took place early in May.
Rally and topping formation Prices advanced sharply immediately following the early May entry, but by the end of May, slackening momentum created negative
divergences in the 12 26 day MACD, establishing the perquisites for the sell signal that developed in early June.
There was a
quick MACD re-entry in mid-June (not shown on the chart) and a secondary sell signal (not marked) in mid-July, by which time more
serious negative divergences (marked on the chart) had developed.
The time span between the tops in early June and mid-July, approximately seven weeks, over which upside momentum clearly
diminished, is usually significant if negative divergences take place over the period.
MACD patterns were clearly lower at the highs of July than of June, though prices did make new highs in July.
mid-July MACD sell signal proved to be extremely well timed.
premature buy signal which was not stopped out though prices declined during September to new lows, because MACD failed to
confirm those new price lows.
Subsequent market action produced increasingly favorable patterns in both price movement and the development of favorable MACD
configurations.
Prices traced out, between September and October, a declining wedge formation with its bullish implications.
(See
At the same time, MACD, both short and long term, traced out a series of higher lows, positive
MACD
AND THE FOUR STAGES OF THE MARKET CYCLE
The smooth patterns of MACD frequently provide clear indications of the four stages of the market cycle.
As a final note regarding MACD, we might observe that its patterns, which are relatively smooth compared to most market indicators,
do provide suggestions as to the four stages of market cycles.
As we know, the accumulation of long positions takes place during Stage 1, generally within the latter half, or at the onset of Stage
2. Positions are held throughout Stage 2 into Stage 3, during which stage long side holdings are reduced in anticipation of the
declines that are reflected in Stage 4.
This is a daily chart that reflects a relatively short term market cycle.
The
MACD indicator may be followed for day trading as well, in which case 5-minute, 30-minute and 60-minute sets are
likely to prove very useful.
Each set should employ a short, a medium and a longer term MACD combination.
These might include, for example, a 6 19 unit MACD for rapid buy signals during more favorable market climates, a
12 26 unit MACD for buy signals during neutral market periods and for sell signals during very weak market periods,
and a 19 39 unit MACD combination for sell signals. Signal lines will usually be 6 9 unit exponential moving
averages of the MACD lines. Shorter term signal lines will often produce more timely signals but will also be more
prone to false signals and whipsaws.
In addition to MACD, the maintenance of an appropriate trend defining moving average is recommended.
The
If you do bypass, then use the 50-unit moving average as your stop, selling if that average is penetrated to the
downside. Sell stops also take place if, following a purchase, the buying MACD level falls to below its previous low
point even if the area above 0 has not been reached.
Always take the second sell signal generated by the 19 39 unit MACD combination. If longer term trends are
clearly negative, you may employ the 12 26 unit MACD for selling as well as for buying, selling on a turndown
and cross from above to below its signal line, in the area above 0.
The more confirmations you secure of the timeliness of your MACD positions, the larger the amount of positions you
should take. Remember the concept of synergy.
Oh, yes, just in case you thought I forgot
CONVERTING THE DAILY BREADTH THRUST MODEL INTO AN INTERMEDIATE ENTRY.
I promised at the end of Chapter 6 to show you how to employ MACD to convert the shorter term daily based Breadth Thrust timing
model into a more intermediate version, capable of remaining in the stock market for longer periods of time.
incidentally, employ this method to extend the lives of other shorter term timing models that you employ.
You might,
chance that you will find yourself extending profit while reducing trades with their attendant whipsaws and transaction costs.
The rule changes are actually rather straightforward.
Here are the rules for the intermediate version of Breadth Thrust.
Buy signals:
As before, you buy when the 10-day exponential average of the ratio, advancing issues/advancing + declining issues,
rises to 61.5% or higher. There are no other buy entries. Sell signals: As before, you sell when the 10-day exponential
average of advancing issues/advancing + declining issues falls to
49.0% or lower PROVIDING THAT The 19 39 day MACD of prices on the New York Stock Exchange does not lie
above 1%. BOTH CONDITIONS MUST EXIST SIMULTANEOUSLY Breadth Thrust levels at 49.0% or below, MACD
below +1.0%.
Otherwise, no sell signal takes place. The 1% level is calculated by subtracting the slower 39-day exponential
average of daily closing prices of the New
York Stock Exchange Index from the faster 19-day exponential average of daily closing prices and then dividing the
result by the level of the faster 19-day exponential average. For example, if the 19-day exponential average of the
New York Stock Exchange Index stood at 6500 and the 39-day
exponential average of the New York Stock Exchange Index stood at 6400 You would subtract 6400 from 6500
(6500 minus 6400). The result, +100. You would then divide 100 by 6400, 100/6400. The result, +.0156. Multiply by
100 to convert this into a percentage,
+1.56%. Inasmuch as MACD lies above 1.0%, no sell would take place at this time. You would sell only after MACD
has fallen below 1.0% PROVIDING that Breadth Thrust readings stand at or below 49.0% at that time. ONCE AGAIN,
BOTH CONDITIONS MUST EXIST SIMULTANEOUSLY.
Table 1 THE MACD FILTERED VERSION OF BREADTH THRUST New York Stock
Exchange Index, 1970 2004 (The method of calculating levels of the NYSE
Index was changed as of December 31, 2002. Data prior to that date has been
rescaled to conform
to current pricing procedures.)
Buy Date Price Sell Date Price Gain/Loss 12/29/70 530 05/18/71
587 +10.7% 12/03/71 565 04/28/72 634 +12.2 09/21/73 610
11/06/73 595 -2.5 01/03/74 562 01/10/74 523 -6.9 10/10/74 388
10/24/74 392 + 1.0 01/03/75 394 07/28/75 502 +27.4 01/02/76
507 04/14/76 564 +11.2 12/09/76 594 01/27/77 586 -1.3
11/10/77 549 12/06/77 542 -1.3 04/17/78 557 06/27/78 564 +
1.3 08/02/78 611 09/26/78 611 0 01/05/79 585 02/08/79 578
-1.2 08/20/82 683 07/19/83 1009 +47.7 08/02/84 936 10/09/84
989 + 5.7 01/14/85 1040 03/25/85 1089 + 4.7 05/20/85 1160
08/05/85 1165 + 0.4 11/11/85 1203 05/19/86 1423 +18.3
01/12/87 1578 04/20/87 1714 + 8.6 01/30/91 1966 05/15/91
2136 + 8.6 12/27/91 2365 03/04/92 2396 + 1.3 05/05/97 4561
10/27/97 4897 + 7.4 05/30/03 5435 08/04/03 5516 + 1.5
12/30/03 6444 03/15/04 6445 + 0.0 05/25/04 6429 06/14/04
6465 + 0.6
Summary of Results:
18 of 24 (75%) of trades were profitable, with one breakeven.
losing trades an average loss of 2.6%. Total percentage profits were 12.76 times the total percentage losses.
The
model, invested only 18.3% of the time, produced annual returns of +4.23% or more than half the total gain recorded
through 100% investment in the New York Stock Exchange over this period. Dividend payments and interest while in
cash are not included.
MACD Filtered Breadth Thrust Applied to the Nasdaq Composite Index
The Nasdaq Composite began its existence on February 5, 1971.
The summary below is based on data from May 1971, the fi rst
buy signal taking place on December 3, of that year. Apart from the initial trade, the model, applied to the Nasdaq Composite
though based on NYSE Index data, would have had the same dates as in Table 1.