After studying this chapter, you should       •  Understand what the term sentiment means       •  Understand the concept of contrary opinion       •  Be familiar with methods for measuring sentiment of uninformed and informed market players       As a general rule, it is foolish to do just what other people are doing, because there are almost sure to be too many people doing the same thing. (William Stanley Jevons [1835–1882], as quoted in Neill, 1997, p 13)The focus of this chapter is market sentiment. Market sentiment refers to the psychology or emotions of market participants. At times, investors are acting on feelings of fear and pessimism. At other times, hope, overconfidence, and greed characterize investor psychology. Investors react emotionally to the market, and these reactions affect the market. Thus, investor psychology is both influenced by and an influencer of market activity.       From a simplistic standpoint, consider a bull market in which stock prices have been advancing. Investors see their portfolio values increasing. Those who have been sitting on the sidelines hear how their friends have made money in the stock market. Not wanting to miss out on these returns, they join in. The average investor is hopeful and confident that the trend of rising stock prices will continue. Of course, as these investors place more and more money in the market, stock prices do rise; in economic jargon: As the quantity of investors in the marketplace increases, the demand for stocks increases, driving stock prices higher. The optimistic view of the market participants drives prices even higher. Seeing that they were correct, investors eventually become overconfident and greedy, and purchase even more stocks irrespective of “value.” At the peak of optimism, investors have placed most of their available money in the stock market. At this point, there are dwindling amounts of money available to fuel the demand that has been driving price upward. There is no more fuel to keep stock prices rising, and the stock market reaches a peak.       Conversely, when investors are pessimistic and fearful, they begin to sell stock. As the level of pessimism rises in the market and more investors sell, stock prices fall. These falling prices lead more and more investors to feel fearful and to sell their shares. When investors are the most pessimistic and fearful, they have withdrawn much of their money from the market. The downward trend exacerbated by investors leaving the market ends, and the market reaches a bottom.WHAT IS SENTIMENT?Sentiment is defined as the net amount of any group of market players’ optimism or pessimism reflected in any asset or market price at a particular time. When a stock or commodity is trading at a price considerably above or below its intrinsic value, something we will not know until considerably later, the difference or deviation from that value often will be accounted for by sentiment. It is the collective emotion and other intangible factors that come from the human interaction involved in determining a price over or under the supposed value. It is the subject of study by behavioral finance departments, which are interested in the ways that human cognitive bias and brain activity affect financial decisions, and it is a staple in technical analysis, for technical analysts have long held that prices are a combination of fact and emotion. When emotion becomes excessive and prices thereby deviate substantially from the norm, a price reversal is usually due, a reversion at least to the mean and sometimes beyond. It is, thus, important for the technical analyst to know when prices are reflecting emotional extremes.       BOX 7.1 THE THEORY OF CONTRARIAN INVESTING       Whenever nonprofessional investors become “significantly” one-sided in their expectations about the future course of stock prices, the market will move in the direction opposite to that which is anticipated by the masses.           Suppose the overwhelming numbers of investors (call them “nonprofessionals”) become rampantly bullish on the market. The logical extension of highly bullish expectations results in the purchase of stocks right up to the respective financial limits of the masses. At the very moment when the masses become most bullish, they will be very nearly fully invested! They won’t have the financial capacity to do more buying. Who then is left to create demand? Certainly not the minority of investors we call professionals. It is that group which recognizes over-valuations and presumably has been the supplier of stock to the nonprofessionals during the time that both prices and the optimism of the masses were rising. Thus, when the crowd has become extraordinarily bullish, a dearth of demand exists. The nonprofessionals are loaded with stocks and are cash-poor, while the professionals are liquid, but in no frame to buy. Demand is saturated, and even minor increases in supply will cause stock prices to tumble. At this point, prices are a strong bet to go (nowhere) but down. (Marty Zweig in the foreword to Ned Davis’ The Triumph of Contrarian Investing, McGraw-Hill, New York, 2004.)MARKET PLAYERS AND SENTIMENT       The appropriate corresponding timing strategy is to follow informed trader sentiment, act against positive feedback trader sentiment, and ignore liquidity trader sentiment. (Wang, 2000)Remember from Chapter 5, “An Overview of Markets,” that there are three types of players inany market: the informed, the uninformed, and the liquidity players. The estate that needed tosell stock to raise cash in the discussion of market players in Chapter 5 was an example of a liq-uidity player. Liquidity players have only a cursory interest in the markets and do not have animportant role in determining price trends. They affect the market minimally. On the other hand,informed and uninformed players are the market. Because the interactions between the informedand uninformed players determine prices, we will focus our discussion on those two groups.       The uninformed players are those participants who, being ruled by their emotions andbiases, act irrationally. They tend to be optimistic after a market rise and buy, thus creating mar-ket peaks, and to be pessimistic during a market decline and sell, thus creating market bottoms.Though the uninformed players are often called the “public,” even professionals can be part ofthis group. It is not simply the profession or career standing of a market player that classifies anindividual as an informed or uninformed player, it is the timing of the player’s optimistic buyingand pessimistic selling relative to market highs and lows. Research has found that even profes-sionals such as mutual fund managers, Wall Street strategists, and investment-advisory newslet-ter writers often behave as uninformed participants. In other words, the majority of marketplayers are uninformed players.       The informed market players tend to act in a way that is contrary to the majority. That is,the informed market participants tend to sell at the top, when the majority is optimistic, and buyat the bottom, when the majority is fearful and selling. Just as uninformed players need not beamateurs, informed players need not be professionals. They can be corporate insiders or daytraders sitting in their dens in the Caribbean.       By and large, the uninformed players have considerably more money than the informedplayers. While day to day the informed players stabilize the markets by spotting and acting uponsmall anomalies in prices or as contrarian investors invest in undervalued assets, over longerperiods, the uninformed tend to overwhelm the price action with their positive feedback, in manyinstances, forcing the informed to ride with the trend of emotion.

Because maximum optimism and pessimism tend to occur at market price extremes, andbecause this emotional background is principally the provenance of the uninformed player, if thetechnical analyst can determine how each group is acting, some knowledge of the future direc-tion of prices can be gained. Presumably, the informed professional will act correctly, and theuninformed public will act incorrectly, especially at emotional extremes. If we know that amajority of those participating in the market are extremely optimistic about stock prices contin-uing on an upward trend, we can conclude that these investors are near fully invested in the mar-ket and that stock prices are close to a peak. The sentiment indicators that we discuss in thischapter are designed to measure the extent of investor optimism or pessimism. By using senti-ment indicators, the technical analyst is attempting to separate the opinions and actions of theinformed players from the uninformed players. The analyst wants to make investment decisionscontrary to those that the uninformed group is making and wants to mimic the actions of theinformed players.      BOX 7.2 NEUROCHEMISTRY AFFECT ON HUMAN THINKING      Neurotransmitters affect emotion and behavior. At present there have been discov-      ered over 108 different neurotransmitters that interact, stimulating and inhibiting the      activities of each other. “Five of these neurotransmitters act throughout most of the      brain: Histamine, serotonin, dopamine, gamma-aminobutyric acid (GABA), and      acetylcholine.” (Peterson, 2007 [Inside the Investor’s Brain: The Power of Mind      Over Money], p. 48) “Additionally, local actions of opiods, norepineprhine, stress      hormones, and omega-3 fatty acids affect behavior and decision making. And if that      weren’t enough, common medications, street drugs, and foods also should be con-      sidered for the neural effects on judgment.” (Peterson, p. 50)          “Many pathological mood states (such as depression, mania, anxiety, and obses-      sion), neurological conditions (such as Parkinson’s disease and Alzheimer’s dis-      ease), and impulse-control disorders (such as kleptomania, compulsive shopping,      and pathological gambling) are known to affect financial decision making: depres-      sion is associated with risk aversion, mania with investing overconfidence, anxiety      with “analysis paralysis,” and compulsions with overtrading. Interestingly, the finan-      cial symptoms of these illnesses can be reduced by medications.” (Peterson, p. 47)HOW DOES HUMAN BIAS AFFECT DECISION MAKING?Remember, the Efficient Markets Hypothesis (EMH) suggests that enough investors are actingrationally at any particular point in time to make it impossible for a technical analyst to profitfrom security mispricing due to the emotions of the uninformed players. However, the field ofbehavioral finance has defined numerous ways in which investors act less than rational. Thesebiases are common not just to the occasional investor or uninformed public but to professionalsas well. Just look at how many professional securities analysts were caught in the late 1990s stock market euphoria. These were not stupid, irrational people, but their inherent biases, thosecommon to all humans, overcame their ability to reason, and they became caught up in the opti-mism of the time, to tragic effect.       BOX 7.3 INVESTORS ARE THEIR OWN WORST ENEMIES       From Zweig (2007)           • Everyone knows that you should buy low and sell high—and yet, all too             often, we buy high and sell low.           • Everyone knows that beating the market is nearly impossible—but just about             everyone thinks he can do it.           • Everyone knows that panic selling is a bad idea—but a company that             announces it earned 23 cents per share instead of 24 cents per share can lose $5             billion of market value in a minute-and-a-half.           • Everyone knows that Wall Street strategists can’t predict what the market is             about to do—but investors still hang on every word from the financial pundits             who prognosticate on TV.           • Everyone knows that chasing hot stocks or mutual funds is a sure way to get             burned—yet millions of investors flock back to the flame every year. Many do             so though they swore, just a year or two before, never to get burned again.          …our brains often drive us to do things that make no logical sense—but make          perfect emotional sense.       Those who study behavioral finance attribute some of the biased behavior of financial mar-ket players to crowd behavior. These researchers have found that crowd opinions are formed byseveral biases. People tend to conform to their group, making the taking of an opposite opinionsometimes difficult and dangerous. People do not like rejection or ridicule and will stay quiet toavoid such pressure. People often meet hostility when going against a crowd. Another bias is thatpeople gain confidence by extrapolating past trends, even when doing so is irrational, and, thus,they tend to switch their opinions slowly. Also, people feel secure in accepting the opinions ofothers, especially “experts,” and tend to believe the establishment will take care of them.       Understanding that investor emotion and bias affect investment decisions is important fortwo reasons. First, understanding the links between emotions, investment behavior, and securityprices can help the technical analyst profit by spotting market extremes. Second, technical ana-lysts must remember that they are subject to the same human biases as other investors. This set ofhuman biases is so strong that even those who recognize them still are affected by them and mustconstantly fight against them. Successful traders and investors often say that the worst enemy ininvestment is oneself. Technical analysts hope to profit from understanding how human bias cancause people to pay prices greater than the intrinsic value for a stock, but if they are not careful,their own biases may cause them to do the same.For example, the behavioral finance principle of “representation” suggests that peopleoften recognize patterns where they do not actually exist. Although it is the technical analyst’sstrategy to attempt to recognize patterns, an analyst must be certain not to “see” patterns that donot really exist. Therefore, an investor or trader must not only understand our human foibles butmust also find a way to either fight against them or avoid them.        At times, emotional excess leads to extraordinary rises in prices (and sometimes to extraor-dinary declines, called crashes or panics). These periods of extraordinary price increases,whether in the stock market, gold, or tulip bulbs, are called bubbles. During a bubble, stock mar-ket returns are much higher than the mean, or average, return. Bubbles are part of that fat tailmentioned in the discussion of the nonrandomness of prices in Chapter 4, “The Technical Analy-sis Controversy.” Although bubbles occur infrequently, they occur considerably more often thanwould be expected under an ideal random walk model.        For the current discussion, the existence of bubbles is proof that prices are not always deter-mined rationally; emotion can get hold of the market and, through positive feedback, run prices farbeyond any reasonable value before reversing. This type of bubble is visible in Figure 7.1. Duringthe late 1990s, security prices were rapidly increasing. By 2000, security prices were extremely high,especially in the technology sector. The price earnings ratios for many companies were at recordhighs. For some companies, the price earnings ratios were infinite because there were no earnings atall. In fact, investors would have to assume that earnings would grow at an astounding 100% per yearfor 20 years to justify the stock prices using traditional stock valuation models. According to invest-ment analyst David Dreman, “This seems to be a classic pattern of investor overreaction” (Dreman,2002, p. 4). Nevertheless, the bubble occurred, indicating that investors of all kinds can become blindto reality when greed and other psychological biases influence decision making.

FIGURE 7.1 The late 1990s bubble (S&P 500 Index: 1990–2004)

BOX 7.4 BOOKS ON THE HISTORY OF MANIAS AND PANICS     A number of excellent books have been written about the manias and panics that     characterize the financial markets. For further information about this phenomenon,     you can read the following:         Ahamed, Liaquat. Lords of Finance: The Bankers Who Broke the World. New         York, NY: Penguin, 2009.         Allen, Fredrick Lewis. Only Yesterday. New York, NY: First Perennial Classics,         2000.         Amyx, Jennifer. Japan’s Financial Crisis: Institutional Rigidity and Reluctant         Change. Princeton, NJ: Princeton University Press, 2004.         Bruner, Robert F. and Sean D. Carr. The Panic of 1907: Lessons Learned from the         Market’s Perfect Storm. New York, NY: John Wiley & Sons, Inc., 2009.         Galbraith, John K. A Short History of Financial Euphoria. New York, NY: Pen-         guin House, 1994.         Kindlelberger, Charles P. Manias, Panics, and Crashes: A History of Financial         Crises. New York, NY: John Wiley & Sons, Inc., 2005.         Mackay, Charles. Extraordinary Popular Delusions and the Madness of Crowds.         Petersfield, Hampshire, UK: Harriman House, 2003.         Reinhard, Carmen M. and Kenneth Rogoff. This Time is Different: Eight Cen-         turies of Financial Folly. Princeton, NJ: Princeton University Press, 2009.         Sobel, Robert. Panic on Wall Street: A History of America’s Financial Disasters.         New York, NY: Macmillan, 1968.         Wicker, Elmus. Banking Panics of the Guilded Age. UK: Cambridge University         Press, 2008.CROWD BEHAVIOR AND THE CONCEPT OF CONTRARY OPINION     The art of contrary thinking may be stated simply: Thrust your thoughts out of a rut.     In a word, be a nonconformist when using your mind.     Sameness of thinking is a natural attribute. So you must expect to practice a little to     get into the habit of throwing your mind into directions that are opposite to the obvi-     ous.     Obvious thinking—or thinking the same way in which everyone else is thinking—     commonly leads to wrong judgments and wrong conclusions.     Let me give you an easily remembered epigram to sum up this thought:     When everyone thinks alike, everyone is likely to be wrong. (Neill, 1997, p. 1)

When individuals think by themselves, they can be very logical and reasonable, but when joinedwith a crowd, they tend to let certain cognitive biases affect their decision making. History isreplete with examples of financial manias, those periods when, in retrospect, the crowd ofinvestors became overly irrational. During those periods, the irrationality is met with a new hys-teria, the belief that “things are different this time.” We have seen this emotional excess justrecently in the Internet stock-price bubble in the late 1990s and in the real estate bubble in theearly 2000s. At those times, it was very difficult to argue, much less invest, against the prevailingtrend of emotion. Too many were making too much money regardless of their reasoning. Ofcourse, such times eventually reverse and return to normal and often decline to an oppositeexcess. Not believing that they are personally caught in a mania, when prices reverse, peopleblame others for their own irrationality. Denying their own responsibility for being duped by theemotions of the crowd, they demand new laws be passed to prevent “evil” corporations or gov-ernment laxity or fads as derivatives from causing another bubble. Such behavior is not limited tofinancial markets. Manias occur in politics, religion, philosophy, education—almost everyhuman endeavor. They are often man-made, as in either the Tulip Bulb mania or as in politicsthrough propaganda and “spin.” The Theory of Contrary Opinion is an attempt to teach individ-uals how to recognize and profit from such excesses in emotional fervor and to look at all sidesof a belief before committing to it or rejecting it.       A “crowd” thinks with its heart (that is, is influenced by emotions) while an individ-       ual thinks with his brain. (Neill, 1997, p. 3)       Contrary opinion is a “way of thinking…It is more of an antidote to general forecastingthan a system for forecasting. In a few words, it is a thinking tool, not a crystal ball” (Neill, 1997,p. 9). To be a contrarian, an investor must sell (be pessimistic) when the overall market mood isgrossly optimistic and buy (be optimistic) when most investors are pessimistic and in a panic.Although this might sound easy enough, the problem with implementing a contrarian strategy isthat it is indefinite. Remember that one of the basic tenets of Dow Theory is that prices trend.When prices are trending upward, we want to be in a long position, riding the trend. The goal ofunderstanding sentiment is to discern when that trend is losing energy and will reverse. There-fore, the task of the contrarian player is to find a way in which to quantify which direction themajority of market players is headed and to question whether there is enough remaining energyto keep the market moving in that direction. Remember that so long as players still have moneyto invest in the market, their optimism will drive prices higher. It is only when players are fullyinvested that their optimism will not be accompanied by security purchases. At this point, themarket is at an excess, and the trend often ends. To quantify these excesses, the technical analystuses publicly available data to construct indicators of emotional excess. Now that we havelooked at some of the theoretical underpinnings, let us look at how these sentiment indicators aretypically constructed and evaluated.HOW IS SENTIMENT OF UNINFORMED PLAYERS MEASURED?       A top in the market is the point of maximum optimism, and a bottom in the market is       the point of maximum pessimism. (Davis, 2003, p. 9)

Sentiment indicators are data series that give the technical analyst some feeling for how muchprices are at an excessively emotional level. With that information, potential future reversals intrend can be better anticipated. Generally, sentiment indicators are more useful in analyzing mar-kets than individual issues. Individual issue prices have their emotional component, of course,but ways to measure that component are much less reliable than those of measuring overall mar-ket sentiment. Therefore, we focus our discussion on indicators that reflect overall market opti-mism.       Remember, we are interested in the two broad categories of players—the uninformed andthe informed. Most sentiment indicators focus on the uninformed. These uninformed players areusually wrong at major market turns. Therefore, if we know what the uninformed are doing, wehave a clue about what not to do. On the other hand, some sentiment indicators attempt to meas-ure the action of informed players, who generally are accurate in their assessment of marketprospects. These indicators are based on watching professional traders and corporate insidersand following their lead.       Fear and greed are not mirror images of one another. Emotional excess is often the sharpestat market bottoms when panic has occurred. On the other hand, optimism can last for a longwhile. Most sentiment indicators are, therefore, useful in determining market bottoms when thefear reaches its highest level. These indicators can often be deceiving on the rise in prices, how-ever, because the extreme in greed, the converse of fear in markets, will place them at high levelsduring which the market will continue to rise. A sell signal generated by a specific sentimentindicator is, thus, less likely to be as valid as a buy signal.       Sentiment Indicators Based on Options and VolatilityTo glean some information about what uninformed traders are doing, analysts often consideroption trading activity and volatility measures. Option trading can be a sign of market specula-tion, and volatility can be an indication of the anxiousness of market players. Let’s look at someof these measures.       Option Trading and SentimentTraditionally, odd lot statistics were reliable indicators of the sentiment of uninformed, smallinvestors. That small investors, who did not have enough capital to purchase round, 100-sharelots, were heavily buying stocks was an indication that the uninformed public was overly opti-mistic. When small, uninformed investors were highly pessimistic, they would short sell oddlots. The odd lot figures represented a measure of uninformed, public speculation, which tendedto be highest at market turning points.       Today, listed options data has replaced the old odd-lot figures as one of the best measuresof public speculation. A call option is an option to buy an asset, usually a stock or commodity, ata fixed price for a specific period. A put option is an option to sell an asset at a fixed price for aspecific period of time. Some options, by expanding on the basics of time and price, can becomevery complex. However, the standard call and put option is the most widely traded and has thehighest volume of any option type. The option market, by its very nature, is a speculative market.

It depends on leverage for maximum gains, and positions can close worthless on the expirationof options. As such, it has become a speculative vehicle for the uninformed public.         Let us look at how the options market can measure sentiment. Let us assume that Jerrythinks that the price of stock XYZ will increase above its current level of $20 per share. Jerry canpurchase a call option in which he has the option to buy 100 shares of XYZ at a price of $20 pershare anytime in the next three months. The option price and premium—say, $2 per share—ismuch less than the outright purchase price of the stock. If the price of XYZ rises above $20, Jerrycan exercise his option and purchase shares at the guaranteed, and now very favorable, $20. Ifinstead, the price of XYZ declines or remains flat during the three-month period, Jerry will allowthe option to expire and he will lose his investment. Thus, the option market gives Jerry, an unin-formed player, a way to speculate about the movement of the price of a stock by paying a smallfee for the option. When investors think that stock prices will rise, they speculate by purchasingcall options. When investors are bearish, they speculate by purchasing put options. Wheninvestors are very bullish, they buy out-of-the-money call options—those that have a strikingprice above the current stock price—because they trade at very low prices.         Owners will exercise their call options when they correctly project price increases andtheir put options when they correctly anticipate price decreases. When investors incorrectly pre-dict market moves, exercising their options is unprofitable. If the owner of an option does notexercise the option by the expiration date, then the option expires worthless.1         Because the purchase of a call represents one who believes the stock market will rise and aput reflects a bearish opinion, the ratio of calls to puts represents the relative demand for optionsby speculators and, thus, is a hint as to their disposition toward the market. The more call buyersrelative to put buyers, the more optimistic are the speculators.         Using Put/Call Ratios to Gauge SentimentThere are several ways to calculate a ratio between puts and calls. Some have used a ratio of theaverage premium paid for calls versus the average premium paid for puts. In theory, the premiumrepresents the anxiousness of the option buyer and the reticence of the option seller. Statistically,however, this has not been reliable for indicating sentiment. Some analysts have added the priceof all options traded each day multiplied by the volume of each trade to arrive at a dollar volumeratio between calls and puts. Not only does this calculation require accurate data and significantcomputing power, but the information provided by this calculation also has not seemed to be par-ticularly useful. Others have calculated a ratio based on the open interest in calls and puts. Unfor-tunately, this has also turned out to be a mediocre indicator of contrary opinion.1. It has long been thought that most options expire worthless, indicating that most people purchasing options have incor-rectly predicted the direction of market moves. However, recent research indicates that more are exercised than had beenthought. In the November 2004 issue of Technical Analysis of Stocks and Commodities, Tom Gentile reports on a studyof 30 years of option data conducted by Alex Johnson of the International Securities Exchange who found that only 30%of options expire worthless. Roughly 10% are exercised, and the remaining 60% are closed through offsetting transac-tions. The percentage expiring worthless, nevertheless, is large and still suggests that many option buyers are unin-formed.

The final, simplest, and most consistent method of calculating puts to calls is to calculate aratio of the total volume of puts traded in a day versus the total volume of calls (McMillan,1996). For the stock market, the raw volume statistics as well as the ratio are available inMicrosoft Excel format on the Web site of the Chicago Board Options Exchange(www.cboe.com), known as the CBOE, the largest options exchange in the world according tothe Futures Industry Association. One unique way of measuring option-player sentiment is theInternational Securities Exchange (ISE) Sentiment Index. The ISE (www.ise.com) is the largestelectronic options exchange in the United States, and offers trading in over 2,000 equity, ETF,index, and foreign exchange products. Their Sentiment Index, based on the idea that investorswill often buy call and put options to express their market view of a stock, uses opening longposition volume only. Because they are not considered representative of true market sentimentdue to their specialized nature, firm trades and market markers are excluded from the calculation.        Option volume ratios have changed over the 35 years that options have been traded inexchanges. It is, thus, imperative to smooth the raw volume data using moving averages toreduce the effect of this long-term relationship change. Ned Davis Research, Inc.(www.ndr.com), using the ISE Sentiment Index, discovered a technique to take advantage ofshort-term stock market moves. They calculated a ratio of the 13-day simple moving average tothe 20-day simple moving average of the index. When the ratio rose above 1.02, a buy signal wasgenerated, and when the ratio declined below 0.98, a sell signal was generated. Figure 7.2 dis-plays and summarizes the results of these signals. They are high-turnover signals with an averageof 12.5 trades per year.

FIGURE 7.2 ISE Sentiment Index and the S&P 500 (daily: May 17, 2002–July 16, 2010)

 Ken Tower, CMT, chief market strategist at Quantitative Analysis Services, Inc., uses aratio of the 10-day moving average of the put/call volume to the 126-day moving average,roughly equivalent to a 2-week versus a 26-week moving average. Deviations between these twoaverages determine the extremes in option emotion. A high ratio suggests more put buyers thancall buyers, indicating that the uninformed players are pessimistic. Because this ratio is a contrar-ian indicator, a high put/call ratio is generally favorable for the future market direction. By com-bining futures premium/cash ratios on the NASDAQ and S&P 500 futures with CBOE put/callvolume, Ned Davis Research, Inc., found another superior method with excellent trade perform-ance, as shown in Figure 7.3. This combination of contrary opinion indicators produced a 26.6%gain per annum when favorable and a 17.8% loss when unfavorable. It appears that option vol-ume and futures premiums are excellent methods of measuring speculator opinion.

FIGURE 7.3 Combining put/call ratios with futures premium to gauge sentiment (December 31,2003–July 19, 2010)        Volatility and SentimentAnother strategy for analyzing the behavior of the uninformed market participants is looking atvolatility. Volatility is a measure of the amount by which a security price oscillates, usually about

its mean, without regard to its trend over a specified period. The most common calculation forvolatility is the standard deviation about the mean. Historical volatility is the standard deviationof prices in the underlying security about its mean over some past period. The 100-day volatility,for example, is the amount by which a security oscillated over the past 100 days about its mean.       Volatility is mean reverting. Therefore, when it gets out of alignment with what it has beenon average, we can assume that it will return to its mean. As in security returns, however, this isnot an absolute. Just as there are fat tails in the distribution of price returns, fat tails also occur involatility distributions. Another common assumption is that volatility is independent of pricereturn. In other words, adherents to this assumption claim that the ability to predict the volatilityof a security will not aid in predicting the future price direction or return. Some evidence refutesthis hypothesis. Volatility is often a measure of the anxiousness of the players in the security mar-ket, increasing as they become nervous and decreasing as they become complacent. Because theplayers act as a crowd and are often uninformed, volatility can be a predictive factor in markets.Let us look at some of the ways to measure volatility.       Using Volatility to Measure SentimentVIX is the exchange symbol for a percentage indicator of implied volatility in Standard & Poor’s500 options. Volatility in the NASDAQ Composite and the S&P 100 Index are represented byVXN and VXO, respectively. VIX, VXN, and VXO are traded on the CBOE as futures andoptions. Instead of measuring historical volatility, these indicators measure what is known asimplied volatility. Historic volatility is past volatility and generally oscillates with past anxious-ness. By looking at implied volatility, the analyst hopes to measure market participants’ anxious-ness about the future.       Implied volatility is a figure derived from the Black-Scholes option formula. The Black-Scholes option-pricing model suggests that the price of an option is a function of the spreadbetween the underlying security price and the strike price of the option, the time remaining in theoption, the prevailing interest rate, and the volatility of the underlying security. If we know theprice of an option, the option strike price, the price of the underlying security, the interest rate,and the time remaining in the options, we can calculate the only missing variable—the impliedvolatility. Thus, implied volatility is the volatility implied by the option traders in their pricing ofthe options in the marketplace. Implied as well as historic volatility correlates to some extentwith market prices. High volatility tends to occur at periods of stress, emotion, uncertainty, fear,and nervousness, most often peaking at a panic bottom. On the other hand, low volatility seemsto occur during market rises and market peaks when emotions are calm, content, and relaxed. Bylooking for extremes in implied volatility then, because implied volatility expresses the expecta-tions of those option traders, we can determine market emotion. VIX is the most common meas-ure of implied market volatility, and its predictive ability is demonstrated in Figure 7.4.FIGURE 7.4 The S&P 500 and VIX (December 1996–July 2010)         Combining Put/Call Ratio and VolatilityOne final method of using option data is using the CBOE put/call volume ratio for all stocks andthe VXO, the implied volatility index for the Standard & Poor’s 100 Index, the OEX. By calcu-lating the ratio of a 10-day moving average to a 65-day moving average for each figure to nor-malize the data and averaging it by summing the two calculations and dividing by two, Kaeppel(2004) developed an option sentiment index that signaled intermediate market turns during theperiod from January 1997 through February 2003. As with many other indicators, the long, sus-tained rise into 2000 was partially missed with early sell signals. Option sentiment and othermeasures of sentiment tend to work more reliably at bottoms when investors panic than during anadvance when greed develops more slowly.         PollsOne way to measure the sentiment of market participants is simply to ask the players if they arebearish or bullish. Although this might appear to be the most straightforward way of gathering information about expectations, sampling problems and other biases associated with poll takingexist. Despite these biases, poll results, if measured over a constant time interval, can give someidea of the public mood. Poll results are contrary indicators because they express optimism atmarket tops and pessimism at market bottoms. Polls, thus, gather information and measure thesentiment of uninformed investors. Several different companies collect and publish sentimentinformation based on polls. Let us look at a few of these.       Advisory OpinionInvestors Intelligence (www.investorsintelligence.com), a wholly-owned U.S. subsidiary ofStockcube Plc, a U.K. company, located in New Rochelle, New York, provides sentiment infor-mation in its Advisory Service Sentiment survey. Since 1963, the company has read approxi-mately 120 independent (“not affiliated with brokers or mutual funds”) investment-advisorynewsletters every week and determined the percentage of those that are bullish, bearish, orexpecting a correction. Intuitively, it seems that newsletter writers would be more sophisticatedand, thus, more in tune with the market than the public they advise, but the numbers over the past40 years instead show a tendency to be incorrect, especially at market extremes. Thus, this sur-vey provides information about the uninformed players and works as a contrarian indicator.What they have found is that when the percentage of bearish advisors is greater than 50 and thepercentage of bullish advisors is less than 25, a buy signal occurs in the general stock market. Onthe other hand, when the percentage of bearish advisors declines below 20 and the percentage ofbullish advisors exceeds 55–60, a sell signal occurs. They neglect showing any tests of these lev-els, and have derived them purely from observation of the statistics over 45+ years.       The profitability of using this information to make trading decisions is questionable. Soltand Statman (1988) found no statistically significant relation between the sentiment of invest-ment newsletters and stock returns: The raw numbers and several ways of looking at them havenot proven to be informative in the past. Colby (2003) found no profitable results in the crossingof advisory data exponential moving averages between 1 and 1,000 weeks.       However, several studies by others have shown that with certain modifications, the advi-sory sentiment in the past has been a somewhat reliable indicator of future stock market priceaction. A standard calculation of advisory sentiment is the ratio of the percentage of bullish advi-sors to the total of the percentages of bullish and bearish advisors. This is then plotted (see Fig-ure 7.5) and the signal levels are determined. Ned Davis Research, Inc., used a ten-week simplemoving average of this ratio and determined that a rise above 69% in the ratio resulted in aloss/year of 0.6%, and a decline below 53% produced a gain/year of 11.5% over the period Sep-tember 18, 1970 through February 12, 2010. These are credible results.       To improve these results, they added a monetary background component, as displayed inFigure 7.6. They based the monetary background on a 26-week rate of change in three-year U.S.Treasury bill yields and Moody’s BAA bond yields from January 1965 through February 12,2010, and produced a gain/annum of 13.1% versus a buy-and-hold gain of 5.8%.

FIGURE 7.5 Advisory opinion—Percentage bullish/[percentage bullish + percentage bearish] (July1968–July 2010)

FIGURE 7.6 Advisory opinion with monetary component (January 1965–July 2010)

Colby (2003) also suggests that when a large percentage of advisors are bearish, marketprices will rise. He suggests using advisory sentiment to find these periods of extreme pessimismby using an optimistically skewed decision rule. In this decision rule, investors take a short posi-tion whenever the percentage of bearish newsletters is greater than the 54-week exponentialmoving average of bears plus ten percentage points. Following this strategy over the 1982–2001period would have resulted in a net profit of 70.3% over the profits of a buy-and-hold strategy.      American Association of Individual InvestorsThe American Association of Individual Investors (AAII; www.aaii.com) compiles a daily pollfrom its 150,000 members on what they believe the stock market will do over the next sixmonths. DeBondt (1993) found that the members polled by the AAII tended to forecast as thoughthey expected a continuation of the past stock returns. Ned Davis Research, Inc., found that overthe July 1987–February 2010 period, when a two-week moving average of the bullish investorpercentage divided by the sum of the bullish and bearish percentages rose above 65.0 (that is, theAAII investors were too optimistic), the stock market had a tendency to decline 2.8% per annum.When the ratio declined below 59.5, the stock market had a tendency to rise 9.9% per annum (seeFigure 7.7).

FIGURE 7.7 American Association of Individual Investors bulls and bears (August 1987– July 2010)

Consensus, Inc., (www.consensus-inc.com) of Independence, Missouri, draws from a mix ofbrokerage house analysts and independent advisory services to compile the Consensus BullishSentiment Index. The data covers a broad spectrum of approaches to the market, including thefundamental, technical, and cyclical. Consensus, Inc., considers only opinions that have beenpublicized.       Market VaneMarket Vane Corporation (www.marketvane.net) of Pasadena, California, polls 100 leadingcommodity trading advisors every day for their opinions of the futures markets, principally:stock indexes, T-bonds, gold, silver, Yen, crude oil, soybeans, live cattle, sugar, and others. Thisdata is then used to construct the bullish consensus statistics published in Barron’s every week.       The Sentix Sentiment IndexOriginated in February 2001, the relatively new Sentix Index (www.sentix.de) is a comprehen-sive poll of German investors about their opinion of the markets, including the U.S. stock andbond markets. The poll is taken every week on Friday, and the results are published each Mondaymorning in Germany. About 3,100 people (among them more than 690 institutional investors) areasked about their opinion on 12 different markets: DAX-Index, TecDAX (German technologystocks), EuroSTOXX 50, S&P 500, NASDAQ Composite, Nikkei-Index, Bund-Future, T-Bond-Future, EUR-USD currency, USD-JPY currency, gold, and oil. The poll includes the investors’expectations for one month (short-term) and six months (medium-term). Huebner (2008) hasdescribed several uses of the Sentix data for anticipating market direction, and van Daele (2005)used the Sentix data for his PhD thesis on why “noise traders” act the way they do.       Consumer Confidence IndexThe Conference Board (www.conference-board.org), producers of the index of leading eco-nomic indicators and the help-wanted index, reports on consumer confidence each month. TheConsumer Confidence Index is based on a representative sample of 5,000 U.S. households. Thesurvey is based on consumer expectations for the U.S. economy. Like most other opinion polls,the survey has been a contrary indicator to the stock market. As seen in Figure 7.8, Ned DavisResearch, Inc., found that between 1969 and 2010, when the survey number rose above 110,demonstrating that consumers were overly optimistic, the stock market remained relatively flat(-0.2% per year). However, when consumers were predominately pessimistic and the surveynumber declined to below 66, the stock market rose on average 14.6% per annum.

FIGURE 7.8 The Consumer Confidence Index (February 1967–June 2010)           Other Measures of Contrary OpinionThe poll-based measures of market sentiment that we have just discussed are based upon whatmarket participants say their market opinions are. Of course, we are not so much interested inwhether market participants say they are optimistic; what were are really interested in is howmuch is their level of optimism resulting in buying and security price increases. Now we con-sider some other measures of contrary opinion that are based upon movement of money withinthe markets.         Buying and Selling ClimaxesInvestors Intelligence uses the term climax to describe a specific event that occurs over a 1-weekperiod. A buying climax occurs when a stock makes a new 52-week high but then closes belowthe previous week’s close. A selling climax occurs when a stock makes a new 52-week low andthen closes above the previous week’s close. “The reason for such a rigid definition for climaxes is that this enables us to classify accurately and consistently what is and what isn’t a climax. Thisis important as we maintain historic records of the climaxes generated each week and have notedthat important market turning points are often accompanied by a sudden rise in the number ofbuying or selling climaxes,” states Investors Intelligence (www.investorsintelligence.com). Fig-ure 7.9 shows the buying and selling climaxes from March 2009 through the beginning of Febru-ary 2010. Their work shows that sellers into buying climaxes and buyers into selling climaxes arecorrect in direction about 80% of the time after four months.

FIGURE 7.9 Buy and sell climaxes (March 2009–February 2010)        Mutual Fund StatisticsBecause mutual fund investors are mostly from the uninformed public sector, mutual fund statis-tics can often be useful in determining what the uninformed is thinking and doing. The most reli-able statistic is the cash reserves in stock mutual funds as a percentage of the assets and adjustedfor interest rates.       Mutual Fund Cash as a Percentage of AssetsIt has long been known that mutual fund cash holdings are contrary indicators for the stock mar-ket. There are many reasons for mutual funds to hold cash, but the bottom line is that high levelsof cash usually occur at stock market bottoms. Jason Goepfert (2004), in his Charles H. DowAward paper, building on earlier work by Fosback (1976) and Ned Davis Research, Inc., foundthat adjusting mutual fund cash for the interest rate is an even more reliable indicator than thecash percent position by itself. He found that when mutual fund cash, adjusted for interest rates, during the period from January 1981 through June 2004, declined to below its lowest threshold,the stock market rose on average 8.1% over the following year. When the cash level was at itshighest, the stock market declined by an average 6.1% over the following year.         Ned Davis Research, Inc., found essentially the same relationship with mutual fund cashpercentage, adjusted for interest rates, and the stock market (see Figure 7.10). By measuring thedeviation from the 13-month average of the stock mutual fund cash/assets ratio, adjusted forinterest rates and during the period from 1965 through February 2010, a level above 0.1 pro-duced an annual gain of 8.4%, and a level below –0.7 produced a gain of only 0.3%.         In neither instance are the results exceptional. Although some watch the mutual fundcash/assets ratio, it appears to be only a marginal indicator of uninformed sentiment. 

FIGURE 7.10 Mutual fund cash positions adjusted for interest rates (January 1965–February 2010)         Rydex FundsMutual fund management companies within the past few years have presented both style andleverage to their public offerings. Rydex Global Advisors (www.rudexfunds.com) has been par-ticularly inventive in its styles. Not only do these include standard long-only stock mutual funds,but also funds that replicate market averages, such as the S&P 500 and the Russell 2000, and oth-ers that add leverage to the portfolios. These are called bull funds because they increase in value

 when the stock market rises. Contrarily, Rydex offers inverse funds in the same style that areshort the averages and other indexes. These are called bear funds because they increase when thestock market declines. If the public expects the market to rise, they will purchase the bull fundsand sell the bear funds, and vice versa. The ratio of the assets held by these two funds, thus, pro-vides an indication of what direction the uninformed investors in the funds expect the market willtravel.         Ned Davis Research, Inc., found that when these investors become optimistic, invariablythe market performs oppositely. Indeed, from January 1994 through July 2010, when the ratiorose above 82.5, when more people were purchasing the bull funds than the bear funds, the stockmarket declined 16.4% per annum, and when these investors loaded up on the bear funds, thestock market advanced 51.7% (see Figure 7.11). These results show an outstanding relationshipbetween sentiment and future market direction.FIGURE 7.11 Rydex Global Advisors bull and bear mutual funds (January 1994–July 2010)         Margin BalancesEach week, Barron’s reports the NYSE margin debt for the previous month. Traditionally, ana-lysts have considered margin balances evidence of what the uninformed speculator is doing,especially at market peaks. Remember that when uninformed investors are most optimistic, they

 have placed most of their capital in the market and, thus, may buy stocks on margin to leveragetheir position. More recently, margin debt reflects professional speculators and might not be asuseful as before. Taking away from the usefulness of margin debt for market forecasting is theability through derivatives of holding positions outside the Federal Reserve requirements formargin, which only apply to banks. Part of the risk incurred by the Long Term Capital Manage-ment (LTCM) operation was over a trillion dollars in derivative contracts, most of which requiredvery little margin. Thus, Barron’s weekly report of margin debt, which at one time was a veryreliable indicator, is no longer an accurate gauge of market sentiment. The relationship betweenmargin debt and the S&P 500 from 1970 through 2010 is shown in Figure 7.12.

FIGURE 7.12 Margin debt and the S&P 500 (January 1970–May 2010)          Money Market Fund AssetsWhere margin debt looks at speculators who are borrowing money to leverage their positions,money market funds are the repository of funds when uninformed players decide to pull backfrom the markets and hold cash equivalents. As a contrary opinion indicator, we would expectthat money market fund assets would increase as investors become more pessimistic and thatwould, thus, be a sign that the market is bottoming. Ned Davis Research, Inc., found exactly that relationship in the percentage of money market assets to total stock market value (see Figure7.13). When the percentage increased above 10.8 and uninformed investors were, thus, bearish,the subsequent market rise was 10.9% per annum. When the percentage declined to and below10.8, the annual loss per annum was 12.5%.

FIGURE 17.13 Percent money market assets to total stock market value (October 1980–June 2010)          Relative VolumeAnother reliable indicator of uninformed sentiment is the ratio of NASDAQ to NYSE volume(see Figure 7.14). This ratio gradually increases as public enthusiasm for speculative stocks onthe NASDAQ increases, and NASDAQ volume increases relative to NYSE volume. The peak ina trend seems to occur when the ratio peaks, and the bottom of the trend occurs after the bottomin the ratio. Ned Davis Research, Inc., during the period August 1998 through January 2010 (seeFigure 7.14), found that when the volume ratio increased above 1.4, the S&P 500 lost 20.9% perannum, and when the ratio declined below 1.1, the S&P 500 gained 30.0% per annum, makingthis an easy-to-calculate indicator with a profitable history.

FIGURE 7.14 Ratio of NASDAQ volume to NYSE volume (August 1998–July 2010)         Uninformed Short SellingHistorically, short selling has been predominately a professional activity. It is even more so todaywith the many derivative securities traded. The modern use of derivatives requires short selling toreduce risk. Thus, the old relationship of short selling having to do solely with opinion about theprospect for companies has diminished. On the other hand, the total amount of short sellingseems to increase with an increase in the market direction and is, thus, a contrary opinion indica-tor.         The Short Interest Ratio is calculated from data provided by the major exchanges, tradi-tionally the New York Stock Exchange, on a monthly basis, and reported in Barron’s and otherfinancial papers. It is calculated by taking the total amount of stocks sold short as of the specificday of the report divided by the average volume for the month. Colby (2003) reports that over the69 years of data from 1932 to 2000, using a buy signal when the current ratio was greater than its74-month exponential moving average, and a sell signal when the 74-month average was broken,a sizable return resulted but it still underperformed the buy-and-hold strategy. This signal workedonly for long positions and was out of the market for 457 months, more than half the time.Ned Davis Research, Inc., (see Figure 7.15) found that the ratio of the 3-month simplemoving average to the 45-month simple moving average of short interest was a useful indicatormarket performance during the 1945–2010 period. When the ratio rose above 1.5 and investorswere generally pessimistic, the market advanced 15.2% per annum, and when the spreaddeclined to and below 1.04, the market only declined 4.4% per annum. Interestingly, these resultsoccurred during the recent periods of derivative use when short selling was an active part of thehedging process.FIGURE 7.15 Short interest ratio and the S&P 500 (December 1945–June 2010)          As for the usefulness of short selling data on individual stocks and determining the poten-tial for a “short squeeze,” which is the rapid rise in a stock’s price as short sellers scramble tocover, data on individual stocks is available but is often clouded by many variables. To get atinformation that is more useful on a company basis, considerable digging and filtering must bedone to eliminate the influence of derivative transactions that may have little to do with theprospects for the firms. Phil Erlanger (www.erlanger.com) accomplishes a considerable amountof work in this area and publishes results periodically on his Web site. He has found several fil-ters that must be applied to individual stock short sale data: (1) The data must be adjusted forsplits—not only price adjustments, but also volume and short interest; (2) it must be normalizedto adjust for short-term volume fluctuations; and (3) it must be normalized to adjust for historicvolatility. A ranking is established, placing the stock within its smoothed, historical context overa five-year period. This ratio establishes the potential attractiveness of the stock. This number should not be used as a mechanical buy signal, however, because the short sellers may be correctin anticipating the stock to decline.         Odd-Lot Short-SellingEarlier in this chapter, we discussed how odd-lot statistics have generally fallen out of favor as ameasure of uninformed sentiment because the option markets have now replaced small stockorders as a means of speculation. The only odd-lot statistic to have some validity today is theodd-lot short sale data. Traditionally, short selling is considered a speculative endeavor. Mosttrading in the stock markets, but the commodity markets, is done on the long side, either initiat-ing an invested position or selling that position. A short sale is a trade that anticipates a down-ward price swing, requires margin, and has infinite risk of loss. It is, thus, used mostly byspeculators. Odd-lot short sellers are, therefore, speculators and their position sizes suggest alack of sophistication. It turns out that, indeed, they are an excellent indicator of uninformed sen-timent and, thus, an excellent contrary indicator of market direction.         Ned Davis Research, Inc., found that odd-lot statistics still have some predictive ability.Figure 7.16 shows the results from combining two odd-lot ratios with a put/call ratio into a con-trary indicator of market direction. For the period January 2002 through mid-July 2010, the indi-cator when showing a preponderance of optimism, resulted in a 6.2% decline per annum, andwhen showing a preponderance of pessimism, resulted in a 27.3% advance. These are impressiveresults and show that the odd-lot trader is still with us and still acts in an uninformed manner. 

FIGURE 7.16 Odd-lot statistics (January 2002–July 2010)

Over the years, analysts have watched a number of developments in the society around them inan attempt to gauge the overall mood, emotion, and sentiment of market participants. Many ofthese indicators are qualitative and not quantitative. Although these indicators are not easilyquantifiable and do not lend themselves to traditional statistical testing, they still provide impor-tant information to the technical analyst.       One of these unquantifiable indicators is the magazine covers theory. The media covers thenews but with a strong bias. It is selling news to those who are interested. If the stock market ishigh and ready to decline, the media would be unlikely to report the danger even if they know it.Instead, they will emphasize the fact that the market has risen and is strong. They want people tolisten, to subscribe, and to read their output, and they will not get business if they report contraryto the popular beliefs of the day. Their business is providing their subscribers with what theywant. Thus, when major news magazines such as Time, Newsweek, U.S. News and World Report,Barron’s, the Economist, or BusinessWeek include on their cover an article on the stock market,up or down, they are emphasizing what the public believes and already knows—and as has beenshown previously, the public is generally wrong, at least at extremes. For this reason, these sto-ries usually occur at or before major turning points in the stock market.       Paul Macrae Montgomery, currently with Universal Economics, has studied magazinecovers back at least to 1923. He has observed that after a positive major magazine cover story onthe stock market, 60% to 65% of the time the market has gained about 30% per annum over thefirst one to eight weeks. Eighty percent of the time, however, the market has then reversed withina year and sustained significant losses (Baum, 2000).       In 2007, University of Richmond Professors Arnold, Earl, and North (2007) published inthe Journal of Finance a study on the market action of the stock of companies featured as coverstories in BusinessWeek, Fortune, and Forbes between 1983 and 2002. They found that the coverfeature article usually followed the stock performance rather than the reverse. A negative story,for example, tended to occur after the decline in the company’s stock, and a positive storyoccurred after a stock rise. They did not find any statistically significant results in postarticle per-formance, either momentum or contrary (with or against the previous trend). They concludedthat if one has a position in a stock that had a good run up or down and a cover story appeared inone of those magazines explaining the reason for the price move, it was probably time to closethe position.       Not only does the media report about the market (which provides some idea about the sen-timent of the market players), but the reports of the media also impact the mood and emotions ofinvestors. A study commissioned by the Wall Street Journal (Klein and Prestbo, 1974, asreported in Kaufman, 1998), found that 99% of financial analysts read a newspaper regularly.Ninety-two percent of these analysts considered the newspaper the “most valuable” publicationthey read. Obviously, the news is important. However, rapid and correct interpretation of facts isdifficult. Sometimes factual news is immediately interpreted by the market incorrectly. Forexample, when Saddam Hussein was captured, the stock market opened up with a large gap justfrom the joy of the news. When investors thought about the consequences of that news, they real-ized it did not change anything, and the stock market closed down that day. Informed traders in amethod called “event trading” likely sold into this emotional opening. It is a method of rapidlygauging the sentiment produced by a news announcement, determining whether the market is

 overacting, and if it is, acting contrarily. Another aspect of event trading is gauging whether themarket or a stock is acting as it should on particular news, and if it is not, perhaps the news wasalready discounted in the price and a change in direction is due. Event trading or news trading isa very short-term use of contrary opinion.      BOX 7.4 ECCENTRIC SENTIMENT INDICATORS      Over the years, stock market followers have developed a number of “eccentric indi-      cators” to predict stock market movements. Not based on economic or financial data,      these indicators are “feel-good” or “hype” indicators that attempt to measure the      overall morale of the investing population. One of the oldest feel-good indicators,      first suggested by either the late Ralph Rotnem of Harris, Upham & Company (now,      after a long string of mergers, Citibank) or Ira Cobleigh & DeAngelis (1983), fol-      lowed women’s hemlines: As hemlines rose, so did the stock market, and as hem-      lines fell, so did the stock market. Consider the Roaring 20s, when women wore      short flapper skirts and a stock market rise followed. When the stock market crashed      during the great depression, long, modest skirts followed. The hemline index implies      that as people become more exuberant, stock prices rise and clothing becomes more      daring, and as society becomes more pessimistic, people become more conservative      with their clothing and investment choices. Market and economy watchers have also      considered beer versus wine sales (people drink more beer when the market is down      and approaching a low), sedans versus coupes (people buy more sedans and fewer      coupes when the market is down), lipstick sales (when the market is down, women      buy cheaper brands), aspirin (a rise in sales correlates with distress about the mar-      ket), and the number of golf balls left at the driving range (people don’t leave balls      when the market is declining).          It must be emphasized that none of these indicators has an effect on stock prices.      If a direct relationship exists, it exists only as correlation without a direct link to the      markets. Indicators, to be truly useful, must have a rationale for their existence. Cor-      relations may be purely accidental and, thus, meaningless. For a discussion of how      some of these offbeat indicators have performed, go to www.Forbes.com/2001/06/      28/exotics.html.      Historical IndicatorsThere are several indicators that technical analysts have used historically that you may see dis-cussed in the literature. Although these indicators have little relevance today, at one time theyplayed a prominent historical role in the measurement of market sentiment.      The first is NYSE member and nonmember statistics. The advent of off-board trading andof electronic trading, complicated by the use of derivatives for hedging, marginalized the useful-ness of this data. At one time, the various ratios (nonmember short sale ratio, public to specialist short sale ratio, and specialist short sale ratio) had some useful, predictive meaning in the stockmarket. No longer, however, do these figures mean anything. Because the marketplace itself haschanged so drastically, the member figures have gone out of use and are considered unreliable.       The second historically important sentiment indicator to have fallen from favor is the Bar-ron’s Confidence Index. This index, developed in 1932, measures the ratio of yields on high-grade bonds versus yields on speculative bonds. Although it is still published today, it seems nolonger to have relevance for measuring stock market sentiment.HOW IS THE SENTIMENT OF INFORMED PLAYERS MEASURED?Thus far, we have focused mainly on the sentiment of the uninformed market players. Rememberthat these market participants often make incorrect market decisions, especially at marketextremes. Therefore, the sentiment of the uninformed players is used in a contrarian strategy.Now, we center our attention on the sentiment of informed players—those most likely to makecorrect market decisions.       InsidersThe ultimate informed player, at least in individual stocks or commodities, is the insider. Aninsider is anyone who is a knowledgeable member of a firm that either trades in the commodityunderlying the future most important to the firm’s business, such as oil to an oil company orcocoa to a candy company, or who has knowledge of a company’s internal business prospectsand results and is a stockholder. Naturally, these people will act for their own benefit, hopefullywithin the law, and buy and sell based on their knowledge. Under SEC regulations, corporateinsiders must report any stock transactions they make within a month, and in turn the SECreports these transactions weekly. Because insiders are not allowed to profit from transactions intheir company’s stock for six months, their actions are a long-term indicator of prospects for thecompany beyond six months. Investors Intelligence and Vickers Stock Research have found thatthe compilation of all insider transactions is useful for forecasting the stock market a year outfrom the reports.       Sell/Buy RatioThe Sell/Buy ratio is compiled by Vickers Stock Research Corporation, a subsidiary of ArgusResearch Group (www.argusgroup.com). It takes into account the total number of insider buyand sell transactions for each company, the percentage of change in insider holdings, the unanim-ity of the transactions within each company, reversals in transaction patterns, and very largetransactions. A plot of the Sell/Buy ratio for 2006 is shown in Figure 7.17.       Vickers considers a ratio under 2.25 a portent of a higher stock market, and a ratio greaterthan 2.25 as a sign of impending market problems. Colby (2003) found that between 1971 and2000, the ratio averaged over five weeks produced for longs only when the smoothed ratiodeclined below 2.25, a 29.2% profit over buy-and-hold. Bjorgen and Leuthold (2002), using onlylarge insider block transactions, found that from 1983 to 1999, only a small percentage of the time do insiders transact their shares in a one-sided direction, but the three times that theyshowed excessive buying, the stock market bottomed within several weeks. On the sell side,when excessive selling occurred, the market peaked, and it either declined or consolidatedapproximately a year later, thus confirming the observations of Investors Intelligence outlinednext.FIGURE 7.17 Insider Sell/Buy ratio (October 2005–October 2006)       Investors Intelligence MethodInvestors Intelligence (www.investorsintelligence.com) follows insider activity in two ways.First, for over 1,600 stocks they calculate the total insider buy and sell transactions over the pre-vious nine months. This figure is then compared with the same figure three months prior. Thisgives them an estimate of the recent direction that insiders are taking with their stock. Second,they break down insider activity into industry groups and compile group rankings that they findoften precede the directional movement of stocks within each group. An example of this analysisis shown in Figure 17.18.Secondary OfferingsAs the stock market rises and long-term interest rates increase, companies tend to offer second-ary offerings of stock. Sometimes insiders wanting to sell stock cause this, and sometimes it iscaused by a desire to raise cheap capital for expansion. Whatever the reason, increased second-ary offerings of stock should increase as the market rises and give warning of a peak. Conversely,when insiders are staying away from the secondary market, the stock market is likely at or closeto a bottom. Ned Davis Research, Inc., looked into this phenomena and found a weak correlationthat confirmed the previous thesis. By calculating a ratio of the 5-month simple moving averageof secondary offerings to the 45-month moving average, they found that when the ratio roseabove 1.47, the percentage gain per year following was 0.7, whereas when the ratio declined to109 and below, the gain per year increased to 12.9% (see Figure 7.19). The Entertainment Industry showed a modest                                                                                                                                              17 Feb 2010               ranking retreat. There was a much larger                                                                                                                                        13               percentage increase for the total sells compared                                                                                                                                12               with the increase for the total buys. Only one of                                                                                                                               11               the twelve stocks shows favorable insider buying.                                                                                                                               10                                                                                                                                                                                                9                                                                                                                                                                                                8                                                                     Last quarter                                This quarter                                   Change                                                                                                                                                                                                7               Buying decisions                                                                       9                                 12                                    +3                6               Selling decisions                                                                26                                     49                                +23                    5               Surplus buying                                                                  -17                                    -37                                -20                                                                                                                                                                                                   4               That positive stock was Sinclair Broadcasting               SBGI (4 buys, 1 sell). They own 58 television                                                                                                                                       3               stations in 35, mostly, medium sized markets.               Revenues have begun to rebound but poor recent                                                                                                                                      2               results forced them to increase their debt load to a               very risky level. The chart shows Feb and Jul 09                                                                                                                                    1               lows below $1 and then a great percentage move                                                                                                                                          Chartcraft semi-log scale, H/L Prices               up to $5 in Oct-09 and $5.50 this month. That is                                                                                                                                        www.investorsintelligence.com 2010 Stockcube R               the area of the long-term down trend line,               extended from the April 2007 peak at $17.50. We               suggested longs three months ago on the upturn to               $4, after the pullback on the first test of resistance.FIGURE 17.18 Example of Investors Intelligence industry insider analysis, February 16, 2010

FIGURE 7.19 Secondary offerings (November 1974–June 2010)

Large blocks tend to be transacted on behalf of professionals. There are several ways that largeblock data is used. The first is the use of the total number of large block volume relative to thetotal volume traded. This figure gives an indication of when the large block trader is transactingthe most number of shares relative to the market as a whole. Colby (2003) found that when thelarge block ratio crossed above its 104-week exponential moving average, a buy signal was gen-erated that was profitable 70% of the time with a net profit of 511% over the period from 1983 to2001. This strategy was only successful on the long side. The short side, which was triggered bythe ratio declining below its 104-week average, ended with a loss.       The anxiousness with which stocks are traded is shown by whether buyers trade on upticksor downticks. Aggressive buyers anxious to get a position in a stock will buy large blocks onupticks. The ratio of blocks traded on upticks to those on downticks is, therefore, an indicator ofthis professional interest in owning stocks.       Ned Davis Research, Inc., found that when large blocks transact predominately ondownticks and then reverse direction, a profitable signal was generated in a study from January1978 through February 2010 (see Figure 7.20). This seems to suggest that although large blockstrade at a market bottom on downticks during the latter stages of panic, when that panic is overand substantial investors begin to take offers on upticks, they know what they are doing and thestock market is invariably at a bottom. The results of the study showed an annual gain of 15.0%over the buy-and-hold of 9.9% per annum on the long side only. Using large block tick data onthe short side was unproductive, thus confirming the directional bias observed by Colby (2003).

FIGURE 7.20 Big block trades from January 1978–July 2010

Art Merrill studied transactions of large blocks of 50,000 or more shares. Merrill dividedthe blocks into categories of upticks, downticks, and flat, smoothing each category’s data. He rana ratio of the uptick average to the downtick average, smoothing that ratio over 52 weeks and cal-culating the running standard deviation from the smoothed average. This provided significantdirectional signals of 66%, 81%, and 76% over the next 13 weeks, 26 weeks, and 52 weeks,respectively (Colby 2003).       Commitment of Traders (COT) Reports       This study examines whether actual trader position-based sentiment index is useful       for predicting returns in the S&P 500 index futures market. The results show that       large speculator sentiment is a price continuation indicator, whereas large hedger       sentiment is a contrary indicator. Small trader sentiment hardly forecasts future mar-       ket movements. Moreover, extreme large trader sentiments and the combination of       extreme large trader sentiments tend to provide more reliable forecasts. These find-       ings suggest that large speculators possess superior timing ability in the market.       (Wang, 2003, p. 891)       With regard to S&P 500 Index futures, we find that large speculator sentiment is a       price continuation indicator, whereas large hedger sentiment is a weak contrary indi-       cator. Small trader sentiment does not forecast returns. We show that extreme levels       and the combination of extreme levels of sentiments of the two types of large traders       may provide a more reliable tool for forecasting. Our result suggests that large spec-       ulators may be associated with superior forecasting ability, large hedgers behave like       positive feedback traders, and small traders are liquidity traders. (Wang, 2000)In 1974, Congress created the Commodity Futures Trading Commission (CFTC; www.cftc.gov)to do the following: (1) “protect market users and the public from fraud, manipulation, and abu-sive practices related to the sale of commodity and financial futures and options” and (2) to “fos-ter open, competitive, and financially sound futures and option markets.”       Each week, the CFTC reports on the large positions held in 22 different futures markets,including stock and bond futures, metals, currency exchange rates, and agriculturals. Thereports, called the Commitments of Traders or COT, are for positions held as of Tuesday’s closeand are published on Friday. Only those markets with 20 or more traders holding positions largeenough to meet the CFTC requirements are included in the reports.       The “public” position is then taken as the difference between total open interest in eachfuture less those positions held by the traders required to report. The trading positions are dividedinto two major categories: commercial and noncommercial. This nomenclature is an outgrowthof the agricultural origins of the reports. In the financial markets, the commercial traders, indi-vidual or institutional, are those who operate in the cash market and are thus called hedgers. Thenoncommercial participants take speculative positions, change positions more frequently, andare called large speculators. Traditionally and empirically, in the stock market, the large specula-tors have a better record of anticipating market moves, whereas the hedgers tend to lag behindand follow the trend (Wang, 2000). Thus, an indicator should consider the spread between the large speculator and the hedger. The small speculators tend to be dysfunctional, and their statis-tics are of little value.       With respect to the S&P 500 futures, Ned Davis Research, Inc., considered only the com-mercial (hedger) positions and found a correlation between their changes in position and the sub-sequent market gain or loss (see Figure 7.21). They took the net position of commercial tradersas a percentage of the 78-week range, smoothed over six weeks. Later, you will be exposed to anoscillator called the stochastic. This NDR calculation is a long-term stochastic. When the sto-chastic advances above overbought at 55%, i.e., when the commercials have large positions, theS&P 500 futures tended to rise 16.0% per annum. When the commercials were bearish and thestochastic declined to 31.5% and below, the market declined 9.0%. This method is the most reli-able one of understanding what the professional, informed traders are doing.

FIGURE 7.21 Commitment of Traders (COT) and the S&P 500 Index futures (September 1984–July2010)      Because the futures market in the stock market is fractionalized by hedging between mar-kets and other financial instruments, the COT figures for any one market might not be reliable.Tom McClellan, editor of the McClellan Market Report (www.mcoscillator.com), combines allthe stock futures data into one series of indicators on a dollar-weighted basis and then watchesthe commercials (hedgers) net long positions as a percentage of the total. He finds that it hasrecently had a three-week lead to cash stock prices.

A number of tests have used COT data in stock futures, as well as data reported by theCFTC. The most workable systems appear to use smoothed data to normalize the longer-termtrends and find that the relationship between commercials and noncommercials are different overtime and between futures contracts. It thus behooves the technical analyst to experiment with thedifferent methods for each future contract to see what works best and to update that work contin-ually to expose any changes in relationships between the major players in each market.SENTIMENT IN OTHER MARKETSGiven the subject matter of this book, our focus thus far in this chapter has been on the stockmarket. However, we complete our discussion of sentiment by presenting a few major measuresof sentiment in other markets.         Treasury Bond Futures Put/Call RatioThe advent of an options market in futures has created a whole world of new sentiment indicatorsfor these futures markets. The most widely traded are the options of Treasury Bond futures. Thechart in Figure 7.22 shows the latest study by Ned Davis Research, Inc., of the predictive ability ofthese options using the standard put/call volume ratio as a proxy for speculation in the TreasuryBond market. What they found was that when ratio advanced above 1.03, the market had too muchpessimism and that the subsequent advance per annum averaged 5.7%. On the other hand, when theratio declined below .66, the market was too euphoric and subsequently declined 9.1% per annum.

FIGURE 7.22 Treasury Bond futures put/call ratio (July 2007–July 2010)

The spread between large speculators and commercial hedgers is positively correlated with bondprices and inversely related to long-term interest rates (see Figure 7.23). Ned Davis Research,Inc., found that in the period between August 1992 and February 2010, when large speculatorswere net long, the bond market rose on average 4.8% per year and declined 4.7% per year whenthe relationship inverted.FIGURE 7.23 Large speculators’ positions and U.S. Treasury Bond futures (September 1992–July 2010)          Treasury Bond Primary Dealer PositionsContradicting the preceding relationship between commercials and the future for the bond mar-ket is the relationship between primary dealer inventories and the future for bond prices. Onewould think that primary dealers in bonds, those who can deal with the Treasury Departmentdirectly, would have hedged inventory positions in long-term bonds and would thus be consid-ered part of the commercial hedger designation by the CFTC in the COT reports. Further, thiswould suggest that the dealers would be net long at bond market bottoms and net short at tops.The opposite seems to be the case. Primary dealers have tended to have the most long positions

 at tops and the most short positions at bottoms (see Figure 7.24). The reason for this logical dis-parity is likely that dealers must anticipate customer demands. They buy issues from the auctionand then sell them to customers. If customers have been bullish, dealers must have an inventory.Thus, they tend to be long at the top when they believe their customers are bullish and willing topay extra for bonds. Likewise, when pessimism reigns, dealers are hesitant to build inventory andinstead hold net short positions, believing that the pessimism will cause customers to sell tothem. Thus, at bottoms, dealers become net short.

FIGURE 7.24 Positions of Treasury primary dealers from November 1996–July 2010         Ned Davis Research, Inc., found that when dealers were net short, bonds advanced 3.4%per annum, when dealers were essentially neutral, bonds advanced 6.1% per annum, and whendealers were net long, bonds declined 9.1%.         T-Bill Rate Expectations by Money Market Fund ManagersThe money market fund business is highly competitive. Money managers, in order to compete onyield, tend to anticipate future short-term interest rates by lengthening or shortening the durationof their T-bill positions. Longer maturity positions suggest that money managers believe thatshort-term rates will decline, and shorter positions indicate a belief that short-term interest rates will rise. This has turned out to be a contrary indicator for the T-bill market yield. Money man-agers have tended to be generally incorrect in their assessment of the future for short-term rates.As seen in Figure 7.25, when money managers increase the maturity of their positions in antici-pation of lower rates, the rates generally rise instead, and vice versa when they shorten their posi-tions in anticipation of a rise in rates.

FIGURE 7.25 Average portfolio maturity of money managers and U.S. T-bill yields (March 1978–May 2010)       Figure 7.25 shows the results of a Ned Davis Research, Inc., study of money fund man-agers and found that when the average maturity in days rose above its six-month simple movingaverage, the 91-day Treasury bill rate advanced 102 basis points per annum. This measure thenis a contrary indicator because when rates are expected to rise, managers should be shorteningtheir maturity length to await the higher rates. Instead, when they believe the rates will go down,it appears they lengthen their maturity, and the T-bill market does just the opposite of theirexpectations. When the maturity length declines below its six-month simple moving average, therate tends to advance, making this calculation a good contrary indicator of Treasury bill rates.

Mark Hulbert publishes a newsletter, the Hulbert Financial Digest, a subsidiary of MarketWatch,that follows the performance of other investment newsletters. He has been doing this since 1980.His methods are similar to those of Investors Intelligence. A number of these newsletters discussthe price of gold. Using this information, Hulbert calculates an index of gold market sentiment.As with other measures of investment-advisory newsletters, the performance results prove to bean excellent contrary indicator of the market’s future direction. Ned Davis Research, Inc., lookedat this data and concluded that when the percentage of gold bulls advanced above 59, the price ofthe XAU, a gold mining stock index, plunged 47.6% per annum, and when the percentagedeclined below 7, the XAU advanced 91.9% per annum. The results are shown in Figure 7.26.

FIGURE 7.26 Hulbert Newsletter Gold Sentiment Index (January 1997–July 2010)CONCLUSIONIn this chapter, we have focused on the idea of market sentiment—the overall psychology of themarket players. Emotions play an important role in determining the actions of market partici-pants. Market participants demonstrate periods of both extreme optimism, when bubbles occur,

 and periods of extreme pessimism, when crashes or panics occur. The uninformed market play-ers tend to be most optimistic as the market reaches a peak. These same individuals tend to bemost pessimistic when the market is at its lowest point in a downturn. In other words, mostinvestors are fully invested just at the time they should liquidate their holdings and out of themarket just at the point when they could be buying stocks at a low price. Sentiment indicatorshelp the technical analyst pick these market extremes. By following a contrarian strategy, thetechnical analyst hopes to act opposite of the uninformed, majority of market players.REVIEW QUESTIONS  1. How would you define the term sentiment as it relates to the financial markets?  2. Warren is searching for a good trading rule to follow. He says, “I would be just as happy to      get information from someone who always makes the wrong investment decision as some-      one who always makes the right investment decision to use in devising my trading strat-      egy.” Explain why Warren would find it helpful to have information about someone’s bad      trading decisions.  3. Explain why extremely high investor optimism is associated with market peaks.  4. Sandra thinks Microsoft (MSFT) is currently overpriced, while Tony thinks MSFT is      underpriced. Which of these two investors would be more likely to buy a put, and which      one would be more likely to buy a call? Explain your answer.  5. You hear a report that the ratio of put-to-call volume is extremely high. How would you      interpret this high put/call ratio? What would you conclude about investor sentiment given      this high ratio? What investment strategy would you want to follow given this high ratio?  6. Explain what is meant by a contrarian investment strategy. What are some market signs      that the contrarian investor might watch for?  7. What information might polls give you about sentiment? What are some sources of poll      data, and what general conclusions can you make about how to use poll data?  8. What type of relationship is generally seen between news reporting and market sentiment?