Technical Analysis Explained

by Martin J Pring

  CHART 23.7 U.S. Government Bond Yields versus Commodity Prices, 1860–2012

  The relatively close, but certainly not perfect, correlation between them is self-evident. What is striking is that commodity prices led yields in four of the five secular turning points shown on the chart. In 1920, the two reversed more or less simultaneously. Clearly, the lead times varied, and one could certainly argue the point that the mid-1990s commodity peak was higher than that of 1980. Nevertheless, the record shows that commodities lead interest rates at secular as well as at cyclical turning points. Unfortunately, the leads for each turning point are varied, starting from the simultaneous reversal in 1920 to a 10-year lead time in the 1932–1946 period. Even so, the strong secular commodity rally in the 2001–2011 period, coming after a 31-year decline in yields, suggests that a secular reversal in favor of inflation may well be in the cards as we approach the middle part of the decade.

  Techniques That Help Determine the Direction of Secular Trends

  Background

  When we are trying to spot changes in primary trends associated with the business cycle, it is occasionally possible to identify reversal signals that take place within a matter of a couple of months of the final turning point. Secular trends extend over many business cycles and, therefore, are much longer in duration. This means that it may take many years, or indeed several business cycles, before a reversal signal can be identified. However, the patience and discipline required to track down these changes are well worth the trouble. First, such signals do not develop very often and are likely to remain in force for one or more decades. Second, the direction of the secular trend has a huge influence on the character of the primary trend. Bull markets in uptrends last, on average, much longer than bull markets in downtrends and so forth. Understanding the direction of the secular trend can, therefore, put us ahead in the process of allocating assets around the business cycle.

  The explanation that follows does not offer all the answers we might like, but it does represent a starting point.

  One of the problems we face is that the recorded history of U.S. financial markets does not go back very far when we consider that a secular trend often extends for 25 years or more. This means that there are not that many turning points to consider. All we can do is apply some of the trend-following principles and tools that might be used for identifying reversals in short-term trends and see how well they work. Specifically, I have found that momentum offers the most accurate and timely signals when confirmed by trendline breaks, whereas moving-average analysis plays a less substantive role. Let’s start our analysis with stocks.

  Identifying Secular Equity Trend Reversals

  The average secular equity bull market since 1900 has lasted 12 years, whereas 18.5 years is the number for bear markets. If the 1921–1929 bull market outlier is ignored, the average length of just over 17 years is more in line with the average secular bear. Once a new secular trend has been identified, one starting point is to relate the time that has already elapsed to the average in order to see how much that trend may be expected to extend into the future. Another benchmark would be the Shiller P/E Ratio to see where it stands relative to its 22 times to 5 to 7 times extreme benchmarks. Chart 23.8 shows that previous secular bears have undergone five to seven price swings in excess of 25 percent. Working on the assumption that all secular bears will be subject to a similar experience, that, too, could be used as a benchmark for discerning the maturity of any existing decline.

  CHART 23.8 Secular Bear Markets Are Deeply Cyclical Affairs, 1900–2011

  Secular bulls are completely different, as primary-trend bear markets that develop under their context rarely experience declines in excess of 25 percent. The 1920–1929 and 1950–1966 inflation-adjusted bulls averaged around 400 percent, and the 1982–2000 experience was close to 700 percent.

  Charts 23.9 and 23.10 compare the U.S. stock prices to a price oscillator using a 3-year moving average (MA) divided by a 12-year span. Since annual data are being used, precise timing should not be expected, but the peaks and troughs in this indicator nevertheless do offer some useful benchmarks of the market’s long-term temperature. Secular accumulation points are indicated in Chart 23.9 when the oscillator bottoms out from at or below the –10 percent level. Only one of the seven signals since 1800 has proved to be a whipsaw and that was the one given in the late 1930s. It will be interesting to see whether the 2012 signal will turn out to be a valid one or more of an accumulation indication as it was in the 1930s and 1940s. Since markets spend more time rising than falling, the negative-signaling benchmark has been raised from 10 percent to 40 percent. In this exercise, peaks are signaled when the oscillator crosses below the +40 percent level. They have been flagged with the downward-pointing arrows. Often, the actual market peak is signaled when the oscillator reverses direction, so the negative overbought crossover is a more conservative approach. In some instances, these peaks are followed by multiyear trading ranges rather than actual declines, but in all instances, nominal prices had a hard time advancing for many years after the signal was given.

  CHART 23.9 U.S. Stock Prices and a Trend Deviation Indicator 1800–2012 Showing Peaks

  CHART 23.10 U.S. Stock Prices and a Trend Deviation Indicator 1800–2012 Showing Bottoms

  Trendlines are often a very useful secular identification tool. Chart 23.11 shows how trendline violations or price pattern completions have reliably signaled major reversals over the last 100 years or so. The problem, of course, is that it is not always possible to construct lines against fast markets, such as the 1929–1932 drop. Alternatively, lines can be constructed but their violation comes well after the turning point. That does not happen to any of the lines drawn on Chart 23.11, but would have for, say, the one joining the 1911 to the 1915 top had it been included.

  CHART 23.11 Inflation-Adjusted Equity Prices 1890–2012 Showing Trendline Applications

  An alternative is to combine trendline violations with an oscillator. In this case, a useful secular span is to divide a 60-month (5-year) by a 360-month (30-year) moving average, as shown in Chart 23.12. Apart from the disastrous 1930s signal contained in the ellipse, when every signal developed completely out of kilter with the price, this approach worked well in the 1900–2013 period.

  CHART 23.12 Inflation-Adjusted Equity Prices 1890–2012 Showing Oscillator Signals

  Identifying Secular Commodity Trend Reversals

  One method is to run a long-term moving average through the data. The problem is that we need to extend the time frame to eliminate whipsaws, but the signals often develop well after the new trend is under way. Chart 23.13 shows a 156-month (13-year) MA for U.S. commodities. It works reasonably well, and crossovers are reliable enough to provide a hint of a reversal, but certainly not enough on which to bet the mortgage. Note that prior to 1860, annual prices are used in the commodity index.

  CHART 23.13 U.S. Commodity Prices, 1800–2012 and a Rate of Change Indicator

  The chart also includes a momentum indicator—in this case, a 240-month (20-year) rate of change (ROC). The smoothing is a 72-month (6-year) MA, which is very good at identifying parabolic tops. Reversals in the smoothing often give reliable signals at bottoms. In that respect, the up-pointing (solid) arrows show when the moving average of the momentum series reverses to the upside. Often, these signals develop some time ahead of the final turning point in commodity prices, so some of the arrows slant to the right to indicate when the price series confirms with a moving-average crossover. The downward-pointing (dashed) arrows indicate secular peaks. In this case, the signals develop when the ROC crosses below its 72-month moving average, not when the average reverses direction. This is because bottoms tend to be rounded affairs, whereas peaks typically take the form of a spike. Note the two whipsaw signals that developed in the 1980–2001 trading range. At the end of 2005, the moving average for the ROC moved back above its 156-month moving average. This was the fifth confirmed buy signal in almost 150 years of dat
a.

  Another useful technique is to adopt the 60-month/360-month price oscillator approach used earlier for stocks. This is shown in Chart 23.14, where 48-month MA crossovers of the oscillator are used as momentum buy/sell alerts. Notice the two arrows at A and B, which indicate the only whipsaws in nearly 200 years of history—okay, a couple of the signals were late, but not a bad overall performance.

  CHART 23.14 U.S. Commodity Prices, 1800–2012 and a Price Oscillator

  Trendline analysis can also be adopted for commodity prices. Some examples are shown in Chart 23.15. Note the dashed up trendline that has its roots in the 1930s. If it is ever violated, expect to see a major commodity decline or extended trading range follow.

  CHART 23.15 U.S. Commodity Prices, 1800–2012 and Trendline Application

  Identifying Secular Bond Yield Reversals

  Many of the same techniques used in commodity analysis can be adopted for bond yields. For example, Chart 23.16 shows that the 240 ROC/trendline combination works quite well. The series in question uses the 30-year yield since its inception in the 1990s but is also spliced to the 20-year government yield prior to that.

  CHART 23.16 U.S. Bond Yields, 1865–2012 and a Rate of Change

  The path of bond yields tends to be smoother than that of stocks and commodities, so a useful combination is to compare a 9-month exponential moving average (EMA) with that of a 96-month series. This is shown in Chart 23.17.

  CHART 23.17 U.S. Bond Yields, 1865–2012 and Two Moving Averages

  The small arrows show the very few whipsaws that have taken place in the last 150 years or so. Note that the yield remained below its 96-month MA for most of the course of the post-1981 secular bear. Since the MA and trendline are in the same vicinity and the line has been touched or approached on numerous occasions, their joint penetration should prove to be a very reliable secular trend-reversal signal whenever that comes. Chart 23.18 shows a more recent history. Note how the 96-month EMA and trendline were almost indistinguishable between 1990 and 2013, thereby reinforcing each one as a resistance barrier. Note also that the yield did not reverse on a dime at either of the two secular turning points shown on the chart. Instead, it experienced an extended trading range in both instances.

  CHART 23.18 U.S. Bond Yields, 1928–2012 and Trendline and Peak-and-Trough Analysis

  Peak-Trough Progression

  Peak-trough progression is another technique that can be applied to the process of identifying secular reversals in bonds yields. It is not a perfect approach, but seems to work on a timelier basis than most. The idea is that a valid uptrend develops when each successive peak is higher than its predecessor, as is each successive trough. In this instance, a peak is a rally high associated with a specific business cycle and a low is a low associated with the contraction or slowdown. When the series of rising peaks and troughs gives way to one of lower peaks and troughs, a trend-reversal signal is given by this technique. The magnitude and duration of the new trend, however, are not indicated. That would be nice to know, but a warning on the direction is not to be sneezed at. Downtrend reversals are signaled in exactly the opposite way, with a series of rising peaks and troughs replacing a declining trend. It should not be assumed that this technique will work in every situation, but it is surprising how effective it can be, especially when used in conjunction with moving-average crossovers and trendline violations, etc.

  The solid wave forms represent movements in excess of 12.5 percent and are used as a basis for objectively measuring what constitutes a legitimate peak or trough. The first signal at A is actually a reconfirmation of the secular downtrend that began in 1920. The series of declining peaks and troughs had been interrupted in early 1932 with a higher high. Since the 1931 low was slightly below its predecessor, the declining troughs were still intact. The break below it at A reconfirmed the secular downtrend. Point B shows the reversal of this decline in the late 1940s. The yield then continued to trace out a series of rising peaks and troughs until point C in the early 1980s. As the chart closes in 2012, the downward peak tough progression continues.

  Secular Trends Dominate the Characteristics of Primary Trends

  In an earlier chapter, we learned that the primary trend determines the characteristics of shorter-term price movements. During a bull market, short-term uptrends have greater magnitude than short-term uptrends that develop in a primary bear market and vice versa. The same is also true for the relationship between the secular and business cycle (primary) associated trend. This is fairly self-evident if you look at Chart 23.2. You can see that primary-trend bear markets that developed, say, in the 1949–1966 or 1982–2000 secular bull market are far more benign those that developed in the secular bearish periods between 1966 and 1982 or 2000 and 2012. An understanding of the direction of the secular trend clearly puts you in a very powerful position. For example, if you correctly conclude that equities are in a secular bull market, it is likely that prices will be much more sensitive to an oversold reading. On the other hand, if the very long-term trend is a downward one, oversold readings would have far less power. Moreover, it is very probable that the magnitude and duration of a primary trend rally will be less in a secular bear market and more likely to run into resistance rather than register a sustainable new all-time high.

  There is an old saying that surprises come in the direction of the main trend. Since the secular trend is really the more dominant, this means that during the secular uptrend, any surprises are likely to come on the inflationary side. Commodity prices rise much faster and further than most people expect. The same would be true of bond yields. The opposite set of surprises develops during a deflationary secular trend. Having said that, these “surprises” typically occur as the trend is in a more mature phase. When it is starting off, commodity prices and interest rates often experience a trading range or transitional period lasting around 5 to 10 years. It is only toward the end of the up wave, when distortions are beginning to evolve, that scary and unexpected rises in commodity prices and yields materialize.

  Tables 23.2 to 23.5 show the actual movements during the 1946–1981 up phase and the 1981–201?down phase for Moody’s Corporate AAA yields.

  TABLE 23.2 Cyclical Yield Rise in a Secular Uptrend

  TABLE 23.3 Cyclical Yield Decline in a Secular Uptrend

  TABLE 23.4 Cyclical Yield Decline in a Secular Downtrend

  TABLE 23.5 Cyclical Yield Rise in a Secular Downtrend

  We have noted that 2012 was the low for the secular trend, but in early 2013, there is insufficient evidence to draw a firm conclusion on this, even though several indicators were suggesting that that could be the case.

  During the secular rise in yields, the average bull part of the cycle lasted around 30 months and took yields approximately just under 40 percent higher; bear markets in yields were shorter, at 19 months, and smaller, as they averaged 13 percent. During the down wave between 1981 and 2012, the bear markets lasted much longer, at 42 months, and took yields down an average 29 percent. Bull markets were shorter, averaging 15 months, but still took the yield up an average of 25 percent. Not every bull move in a secular advance is greater than every bull move in a secular decline and vice versa. However, the average figures indicate that if you can make a correct interpretation about the direction of the secular trend, you have already come a long way in the investment battle.

  Summary

  1. The Kondratieff wave describes the long-term interaction between inflation and deflationary forces, but its rigid, almost predetermined, interpretation has meant that many financial events have not transpired as expected.

  2. Since the nineteenth century, stocks, commodities, and bonds have alternated between secular bull and bear markets, generally lasting about 15 to 20 years.

  3. Secular bear markets for stocks are determined by long-term psychological swings. They are influenced by structural economic problems and unusually volatile commodity prices.

  4. Secular trends can be analyzed with regular t
echnical tools such as momentum, trendline, and moving-average analysis.

  5. Surprises typically come in the direction of the secular trend, which determines the characteristics of the primary or business cycle–associated trends.

  24 TIME: CYCLES AND SEASONAL PATTERNS

  Basic Principles

  Time is represented on the horizontal axis of most technical charts. It is normally used in conjunction with price, volume, and breadth, the other three dimensions of psychology involved in determining trends in the stock market. These latter three, though, are measured on the vertical axis. Time can also be assessed independently through the analysis of cycles.

  Discussions of the importance of time have, so far, been limited to the idea that the significance of a reversal in trend depends upon the length of time needed for a distribution or accumulation to complete. The longer the period, the greater the magnitude and duration of the next move are likely to be. Removing the speculative excesses of a trend requires a commensurately large corrective movement, just as the discipline of a long period of accumulation provides a sound base from which a substantial and lengthy advance can take place. The very long (8-year) bull market between 1921 and 1929 was interrupted by corrective reactions, but the substantial increase in stock prices during this period resulted in a considerable amount of excess confidence and speculative excesses, which were only erased by a sharp and lengthy decline.