The cyclical nature of equities trading is one of the eternal truths of the financial markets, with bull and bear markets having been empirically documented for more than 100 years. Charles Dow,1 co-founder of The Wall Street Journal and a pioneer in technical analysis, compared market movements to ocean waves influenced by tides. Value-investing legend Benjamin Graham2 took a different approach, one emphasizing fundamental analysis. Graham believed that his fundamental approach would be rewarded in the long run, because mispriced securities would ultimately revert back to their fair values. As he so eloquently characterized it, “The market is a pendulum that forever swings between unsustainable optimism and unjustified pessimism.” The emerging field of behavioral finance supports Graham’s point of view, providing a strong theoretical foundation for the overreaction bias that is often a driving factor in cyclical markets.
One of the fascinating aspects of market cyclicality is its fundamental relationship to risk and return. Figure 1 shows realized investor returns over the past 20 years compared with the performance of a wide variety of indicators. The astonishing result is that the average investor badly underperforms across the board because he or she is prone to chasing performance near market tops and panicking near market bottoms. For this reason, an effective measure of market sentiment may be of significant value in navigating risk and enhancing returns. The purpose of this paper is to introduce the Acertus Market Sentiment Indicator (AMSI), which is designed to provide an enhanced means of assessing market behavior, and to discuss how it may add value to the investment process.
The Purpose Of The AMSI, A New Multifactor Sentiment Index
While no one can predict the future, it is important for investors to understand where we are today and to be able to view current conditions in a historical context. The purpose of any index or an indicator is to measure the current level of something. There is a formidable list of indicators for a range of market conditions, and, when appropriately constructed and widely followed, they have the potential to add value when important investment decisions are being made.
We chose to focus on constructing an indicator of “market sentiment,” which, in conjunction with its component parts, can provide important insight into current investor behavior and place it within a historical context. Understanding investor behavior becomes particularly important during the cyclical phases of markets, especially during market extremes. “Sentiment” refers to a feeling, emotion or attitude about something, and, of course, it can have a range of values. With respect to financial markets, fear represents one extreme, while greed represents the other end of the spectrum. We view sentiment as a continuum, with anxiety and complacency representing less extreme and nuanced forms of fear and greed, respectively. In sharply rising markets and near market tops, greed and overreaching clearly come into play, but the authors believe that complacency is the more insidious and potentially destructive sentiment because it can lead investors into a failure to manage and monitor risk.
Reference.com defines complacency as “… a feeling of quiet pleasure or security, often while unaware of some potential danger, defect, or the like; self-satisfaction or smug satisfaction with an existing situation, condition.”4 Rather than the excessive desire for wealth that characterizes greed, we believe that being unaware of dangers accurately describes the reason for many investor losses. “Anxiety” represents a less extreme form of fear that corresponds to distress or a lack of peace of mind. Mild to moderate investment losses cause most investors anxiety, but do not necessarily drive them toward abject fear. Figure 2 expresses this range of investor emotions in the context of Graham’s pendulum analogy.
An effective indicator of market sentiment could help investors avoid the pitfalls associated with the investment performance displayed in Figure 1. A number of sentiment indicators already exist, such as the put-call ratio, the CBOE Volatility Index (VIX), as well as various surveys that attempt to take the “investment pulse” of individual and institutional investors. Each of these indicators has its shortcomings, most often caused by the omission of one or more important factors that may influence market sentiment. The challenge in designing an effective indicator is to make it relatively simple and understandable, yet robust enough to reasonably explain the levels of and changes in market sentiment. A model that is too simple will provide insufficient information, while a model that is too complex may be difficult to understand and is therefore not useful.
Our approach in selecting a mix of fundamental, technical, short-term and long-term elements for the AMSI was to include variables that make intuitive, logical and empirical sense without becoming unwieldy. The result is an indicator that uses the following five factors:
• Price/earnings (P/E) ratio
• Price momentum
• Realized volatility
• High-yield bond returns
• Treasury eurodollar (TED) spread
As described below, all these factors provide some measure of market sentiment, but each adds a unique element to the overall index.
Elements Of AMSI
S&P 500 Price to Earnings (P/E) Ratio
Since stocks represent ownership in a business, it is clear that valuations are what will ultimately matter in the long run. In the short to intermediate term, however, market psychology and other factors may cause prices to deviate substantially from their fundamental values. Hence, we believe that P/E is one of the important elements for a reasonable assessment of market sentiment.
P/E ratios have a long history of being followed by the analyst community, and for many decades they have been published in the stock tables of leading daily newspapers. The AMSI focuses on trailing P/E, rather than forward P/E, due to the former’s longer time history and the latter’s reliance on ever-changing and usually overly optimistic analyst estimates. In 1960, Francis Nicholson5 was the first among many researchers to find that low-trailing P/E levels are generally a predictor of higher stock market returns ahead, while high P/Es usually correspond to lower levels of future returns—findings consistent with the investment maxim, “Buy low and sell high.” Similarly, for the AMSI, low levels of P/E may be indicative of anxiety or fear, while high P/E levels may signal complacency or greed.
Sometimes prices move in advance of the fundamentals. The movement may be due to shifting market psychology, unanticipated events or other factors. For example, investors often legally act on information before it has passed the checks and balances of the published news cycle. Some investors uncover news through original research and make it available to reporters before the news is officially disseminated. Other times, investors may trade in anticipation of a rumored policy change such as a Fed action or an event such as a military attack that may or may not come to pass. This dynamic behavior often explains the “buy on the rumor, sell on the news” pattern observed in security price movements. At other times, momentum can be driven simply by investors not wanting to be left behind the crowd. For the AMSI, high levels of price momentum are indicative of complacency or greed, while low levels may be indicative of fear or anxiety. We define price momentum as the percentage of S&P 500 Index stocks trading above their 200-day moving averages.
Momentum is important because it may cause stock prices to deviate from the underlying fundamentals, as Graham alluded to in his pendulum analogy. High returns in the recent past may lure investors into performance-chasing out of complacency, greed or the often-mistaken belief that the past is prologue to the future. At the other end of the spectrum, low recent historical returns or market movements caused by triggered stop-losses may cause investors to sell near market bottoms out of anxiety or fear. In extreme cases, momentum may completely overwhelm the fundamentals, as was the case in the frenzied rise and subsequent fall in the prices of technology stocks during the Internet bubble from the late 1990s to early 2000.
In the short run, volatility begets more volatility. Complex statistical models are often used to measure this phenomenon. We chose to include the 30-day realized standard deviation of S&P 500 returns as our measure of volatility because of the long-term nature of its data series and the relative simplicity of its calculation and interpretation. By way of contrast, VIX relies on implied or estimated forward volatility, and its data series only goes back to the late 1980s, limiting its usefulness in longer-term analysis.
Volatility may beget volatility due to margin calls, crowd psychology and trend-following strategies like portfolio insurance. Volatility, like momentum, may also help explain why prices deviate from fundamentals. High levels of persistent volatility are generally reflective of anxiety or fear, while low levels may signify complacency or greed. But over long periods, volatility tends to revert to the mean. For example, the standard deviation of U.S. stock returns averages about 20 percent per year. Abnormally low or high periods of volatility are unlikely to persist for very long because the market ultimately recalibrates the relationship between risk and return.
High-Yield Bond Returns
Interest rates are an important macro indicator because they determine the cost of capital for nearly all individuals, businesses and governments. Theoretically, a rise in interest rates should result in a declining present value of future cash flows and lower asset prices. The base level of interest rates is largely determined by the U.S. Treasury curve. Credit spreads are added on top of U.S. Treasury (or other highly rated sovereign debt) rates in order to determine the appropriate yield to maturity or cost of debt capital for an issuer. When sentiment is low and risk is high, credit spreads widen. Credit spreads generally fall as economic fundamentals and sentiment improve.
We believe the returns on high-yield bonds represent a good measure of sentiment in the fixed-income markets because they often correspond to changes in the economy and interest rates. U.S. Treasurys can benefit from a “flight to quality” effect during times of market distress and may be artificially depressed when, for example, the Federal Reserve or other central bank is executing a quantitative easing policy. Also, during times of complacency and greed, credit standards often become lax. For these reasons, we view the returns on high-yield bonds as a “purer play” on interest rate and credit risk, and have included the Credit Suisse High Yield Bond Index as the fourth component of the AMSI. The Credit Suisse index was established in late 1985, and, we believe, provides an effective barometer of returns for this asset class.
Treasury Eurodollar (TED) Spread
We included the TED spread as a measure of systemic financial risk. The TED spread is based on the difference between the three-month Libor and the three-month U.S. Treasury bill interest rates. Banks serve as financial market conduits for nearly all businesses and governments through their capital-raising, market-making and savings activities. As we observed during the credit crisis, when banks have a problem, it ultimately becomes everyone’s problem because of their enormous global scale and substantial leverage.
Global financial intuitions today rely not only on customer deposits to fund their operations, but also on short-term loans from other banks. Therefore, if there is a crisis of confidence among banks, the TED spread is likely to expand dramatically. It historically ranges between 10 and 50 basis points, but it spiked to an astonishing 488 basis points around the time of the Lehman Brothers bankruptcy announcement. The TED spread also provides value to the AMSI outside of times of financial crises. For example, a less-than-dramatic rise in the TED spread may portend a credit crunch or slowdown in the economy. Therefore, we believe that the TED spread adds a unique measure of market sentiment to the index.
Architecture Of AMSI
Figure 3 provides a summary of the five factors that comprise the AMSI. The weights of the individual AMSI factors are proprietary and dynamic, adjusting each time the index is updated. To facilitate the interpretation of the AMSI, we converted the indicator to a percentile value, ranging from 0 (extreme fear) to 100 (extreme greed). The importance of each element with respect to the significance of their weights in the AMSI, in descending order, are P/E, momentum, realized volatility, high-yield returns and TED spread. AMSI spans the period from January 1986 to the present, i.e., the longest common period over which data for all component elements is available. The absence of data prior to the mid-1980s for high-yield bonds and Libor, which is a component of the TED spread, precludes AMSI from starting at an earlier date.
Figure 4 is a table showing the weights of the different factors at different times, including maximum and minimum weightings. To determine the dynamic weights, we find the percentage ranks of the individual components’ historical data series and run correlations against rolling three-month, six-month, nine-month and 12-month trailing and forward-looking S&P returns. For each rolling data set, we calculate a weight for each component by dividing the absolute value of its correlation by the sum of all of the components’ correlations to that data set. We continue this process for each set of rolling returns. The final component weight is a simple average of all of these weights across the four periods.
A graph of AMSI and of the AMSI Six-Month Moving Average from AMSI’s January 1986 inception through November 2013 is shown in Figure 5. The dynamic and cyclical nature evokes the ocean wavelike movement described by Charles Dow and the pendulum analogy used by Benjamin Graham. Figure 5 also calls to mind the comments made by prominent present-day investor Howard Marks6 of Oaktree Capital Management. Marks said, “I believe strongly that (a) most key phenomena in the investment world are inherently cyclical, (b) these cycles repeat, reflecting consistent patterns of behavior, and (c) the results of that behavior are predictable.”
A moving average is designed to smooth volatile data series, thereby providing a more stable pattern. Moving averages of data series often enable trends to come into clearer focus, while original data series may often “flip flop” in its direction, providing mixed signals. The weakness of moving averages is they are lagged and may be delayed in identifying the turning points of an indicator. We tested a number of moving averages and found the six-month moving average provided the best combination of trend exposition and timeliness. In Figure 5, both line graphs illustrate a similar pattern, but the moving average line is more stable.
The Value Of AMSI In The Investment Process
Indicators do as their name implies—they indicate. As such, they need not be predictive to add value. For example, the Dow Jones industrial average is a barometer of U.S. blue chip stock market performance that is generally not used to forecast future stock market returns. While neither AMSI nor any other indicator is a silver bullet able to forecast future stock market returns with a high degree of accuracy, as we noted earlier, a well-designed indicator of market sentiment can add significant value in the investment process.
In our view, AMSI adds value from several perspectives. Tracking AMSI on a regular basis may provide a more robust measure of market sentiment than the VIX, put-call ratio or other indicators. AMSI, along with a concurrent analysis of its component parts and moving average, should help investors gain additional perspective and insight into the current relationship of market levels of risk versus potential return. Lastly, sentiment readings can also play an important role with respect to adherence to investment policy statements or structuring portfolios with proper risk management controls. This is the case because the temptation to abandon well-thought-out long-term plans is typically highest at market and sentiment extremes.
Figure 6 provides information on the historical distribution of AMSI and forward-looking S&P 500 returns. In general, an AMSI level of 0-25 indicates fear, 25-50 indicates anxiety, 50-75 indicates complacency and 75-100 indicates greed.
AMSI levels at the extremes are somewhat rare, typical of most bell curve distributions. AMSI levels of 80-100, indicative of high levels of greed (an extreme version of complacency), occurred roughly 1 percent of the time since our measurement period began in January 1986. Similarly, AMSI levels of 0-20, signifying high levels of fear (an extreme version of anxiety), were also uncommon, occurring only 6.0 percent of the time. However, market returns for the six- and 12-month periods ahead of these extreme periods are striking and aberrational. High complacency/greed levels following periods of extreme market performance portended lower-than-average returns, with six- and 12-month gains of 1.7 percent and 8.7 percent, respectively. Conversely, high anxiety/fear levels following significant market sell-offs historically indicated high future returns, delivering on average 8.3 percent and 10.2 percent for the six- and 12-month-periods ahead. A similar return pattern was observed for less extreme AMSI levels in the 20-40 range versus the 60-80 range. An analysis of median returns yielded similar results.
While the findings at the extreme levels are not statistically significant due to the small sample size, they are anecdotally very interesting. The forward returns posted after these extreme AMSI readings are telling, because they point to periods when investors may have been acting either out of anxiety, fear, complacency or greed. This is consistent with behavioral finance studies that have shown that investors tend to follow near-term trends. The value of this information is that it alerts investors to the need to seek protection during high levels of complacency or greed and remain invested or perhaps become increasingly aggressive when anxiety or fear is the reigning sentiment.
As would be expected, during the majority of the periods observed, AMSI provided a reading near the middle of the bell curve distribution. This finding is not surprising given that the markets are reasonably efficient and that sentiment is usually not at extremes. However, readings in the 20-80 range can also prove to be relevant despite less dramatic S&P 500 results on a forward-looking basis. Most notably, changes in the direction of AMSI, and its moving average, may be important because they provide information on whether the market is trending toward greed or fear. Furthermore, neutral readings (in the area of 50) may not indicate that the values of most components are near their means, but may instead be reflective of mixed internal signals, the analysis of which may provide substantial insight into market sentiment.
Since the AMSI most often finds itself in a neutral range, its effectiveness lies, in part, in offering perspective on both the level and drivers of that sentiment at a specific point in time. For example, high P/E and strong momentum were the principal factors in the high levels of AMSI complacency or readings during the height of the Internet bubble. Conversely, low-volatility and high-momentum readings were the main factors behind the complacency/greed period that preceded the credit crisis. The interaction of the different forces that drove AMSI sentiment readings in these examples (i.e., high P/E versus low volatility) demonstrate the value of a multifactor model over a single-factor model such as VIX or the put-call ratio.
Figure 7 provides a summary of the values of AMSI and its components before and after three extreme periods in recent financial market history—the crash of 1987, the Internet bubble and the credit crisis. Looking specifically at the time period around the crash of 1987 (Figure 8), at the end of August near the market’s peak, AMSI had a reading of about 70 (90th percentile), well ensconced in the zone of complacency indicating a heightened level of risk before the ensuing market crash. Net selling continued throughout most of September, with modest reprieves, before the catastrophic crash on “Black Monday,” Oct. 19, 1987. After the crash at the end of October, AMSI generated a reading of 12.3 (3rd percentile), suggesting that remaining in the market was warranted and that the selling was overdone.
During the late stage of the Internet bubble in the second quarter of 1999, AMSI, bolstered by extremely high levels for the P/E ratio, provided a reading suggestive of too much complacency or greed. As the bubble began to deflate, momentum quickly dissipated and volatility surged, moving the AMSI toward neutral readings and ultimately toward anxiety and fear. As the three-year market decline initially triggered by the bursting of the internet bubble began to bottom in late 2002, AMSI flashed readings below 25 (10th percentile), suggesting that fear was getting priced into the market and better times lay ahead.
Figure 9 provides a graph of AMSI around the time of the credit crisis. At the end of May 2007, the AMSI reading was 63.1 (75th percentile), suggesting the market may have become too complacent. In this time frame, the subprime crisis was in the midst of unraveling. New Century Financial had declared bankruptcy the prior month, and Bear Stearns would follow the next month with a bailout of its mortgage-related investment funds. During the depths of the credit crisis, over the October 2008 to November 2008 time period, AMSI readings were at their lowest ever, indicating that extreme fear had set in and that it might be an appropriate time to increase risk despite rampant market apprehension. AMSI readings continued to be subdued throughout the first half of 2009 as the market searched for a support level near its lows. By the latter part of 2009, fear had largely left the market, and the surge in stock prices pushed AMSI well into the complacency or greed zone.
While, as noted earlier, there are a host of sentiment indicators, for this part of the discussion, we focus on VIX and the put-call ratio, two of the most widely followed indicators. Both, along with AMSI, indicated high levels of anxiety and fear near the depths of financial crises. However, AMSI signaled high levels of complacency or greed prior to the three financial crises as noted in Figure 10, while VIX and the put-call ratio provided mixed results.
A graph of the three sentiment indicators around specific time periods further illustrates how they differ. For illustration, we have chosen the period around the credit crisis, shown in Figure 11.
To better facilitate comparison to AMSI, we computed 1 minus the percentile values of VIX and the put-call ratio. Low levels of each indicator signify anxiety or fear, while high levels signify complacency or greed.
For the sake of comparison, all sentiment indicators are converted to percentile levels. VIX, designated in red, meandered around its long-term average of 20 (corresponding to its middle percentile levels) from the beginning of 2007 through the middle of 2008, providing little indication of the complacency or risk that had made its way into the market. VIX surged only after the credit crisis moved into full swing in the latter portion of 2008 and remained above its average of roughly 20 (or median percentile) throughout much of 2009. The put-call ratio provided generally high readings throughout the credit crisis. This ratio is often a contrarian indicator, with high levels of the indicator usually considered to be a bullish signal. AMSI, however, provided indications of complacency in the first half of 2007 and then consistently remained at anxiety and fear levels until the second half of 2009, when the credit crisis began to slowly fade into a bad memory for investors.
In our view, AMSI’s use of multiple factors provides it with a more robust indication of market sentiment and, therefore, may give fewer false signals and is capable of adding more value than either VIX or the put-call ratio. During this challenging period, AMSI demonstrated more cyclical behavior than either of the more established indices, consistent with the thoughts of Dow and Graham. The put-call did not provide any indication of the enormous risks that lied ahead. Conversely, VIX stayed in the fear zone throughout much of the credit crisis and its recovery, thereby failing to capture the shift in sentiment that had occurred.
In summary, we believe the benefits offered by AMSI are substantial. First, in our view, tracking7 AMSI on a regular basis will provide a more robust measure of market sentiment than the VIX, the put-call ratio or other indicators. As such, it should help investors gain a better perspective on and insight into the current relationship between the levels of risk and potential return in the market. Second, given that better perspective, it can be of help to professional investors, including those advising institutional and ultra high net worth clients, by providing a framework for a more meaningful dialogue about the nature of risk and return with their constituents. Third, during periods of extreme readings, it may offer some insight into the probable direction of the S&P 500 over the six- to 12-month period ahead. Lastly, because the temptation to abandon well-thought-out long-term plans is usually highest at market and sentiment extremes, tracking AMSI also becomes an important guardrail with respect to adherence to investment policy statements and maintaining proper portfolio risk management controls.
References And Endnotes
- Dow, Charles, 1901, “Watching the Tide,” The Wall Street Journal, Jan. 31, p. 1.
- Graham, Benjamin. 1949, “The Intelligent Investor,” 1st edition, New York: Harper & Brothers.
- J.P. Morgan, Guide to the Markets: 2Q:2013, April 1, 2013, p. 63.
- www.Reference.com (for definition of complacency)
- Nicholson, Francis, 1960, “Price-Earnings Ratios,” Financial Analysts Journal, 16(4): 43-45.
- Marks, Howard, 2013, “Howard Marks: We’re Not at Bubble-Type Highs,” Barron’s, Nov. 27. Accessed at http://online.barrons.com/article/SB50001424053111904642604579223961257360926.html#text.print
- AMSI and its component values, along with commentary on the index, are published each month at www.acertuscap.com.