trading significantly below intrinsic value occur infrequently, perhaps once
every 10-20 years per asset class. We need to accept that asset pricing (for
long hold periods) is rational and correct most the time, and that cheap asset
classes are usually cheap for a reason.
I discuss methods
I use to determine when an asset class is irrationally mispriced well below long-term
intrinsic value. I’m trying to distinguish a once-a-decade compelling valuation
versus run-of-the-mill everyday value opportunities. Compelling values set up
lucrative multi-year trades to the long side.
While there’s no
formula for identifying these opportunities, I present three steps I take in
the search. First, I find prices that appear to be an extreme outlier with
respect to history, logic and other asset class valuations. Second, I identify the
behavioral effect causing the asset class to be mispriced compared to long-term
intrinsic value. Finally, I ask myself a few common-sense questions, such as: If
I was designing a portfolio to hold for the next 10 years, would I substantially
overweight this asset class?
A compelling value
is a trading edge that’s long-term in nature, so an asset class can stay that
way for years. Since it’s unknown when (or if) the market will recognize a compelling
value, an asset class trader ultimately requires a timing catalyst to initiate
the trade. Getting the timing basically right – for a move that can take years
to unfold – is an important edge for the asset class trader compared to value
investors and asset allocators.
Value investors typically hold cheap securities based on an analysis of future growth prospects and many valuation measures. They sell assets that rise in price (and hit price targets) and replenish the portfolio with cheaper assets that have recently suffered price drops. They manage the portfolio with a low turnover, typically 25% per year or lower. Many securities purchased may be underpriced due to irrational overreactions to bad news, but most are likely cheap for a reason – they are riskier or have inferior growth rates. This is particularly true for the value screens used by index-based value ETFs.
In this post, I examine the popular stock market valuation tool, the Shiller CAPE. The Shiller CAPE valuation approach, based on 150 years of data, appears to have an uncanny ability to predict future S&P 500 returns.
Unfortunately, the benefits of using this tool for actual investment decisions appear to be limited. The Shiller CAPE, along with all asset class valuation measures, has the following significant weaknesses and issues.
Selection bias has likely overstated the reliability of predicting future expected returns.
Using CAPE to shift between equities and T-bills doesn’t enhance risk-adjusted returns.
Using historical valuation data is susceptible to unpredictable long-term regime shifts that can devastate the effectiveness of such a tool.
When the Shiller CAPE is low, risks are high, and many competing asset classes are also priced cheaply. When the CAPE is high, competing assets also have low expected returns. It appears the S&P 500 is efficiently priced on the time scale used for value investing.
The best use of the Shiller CAPE is simply to set return expectations, specifically if valuation multiples revert to a long-term mean. Such expectations should also be contrasted with no-mean reversion return estimates, such as assuming the S&P 500 is simply priced to deliver a 3-5% premium over the current values of any of the following: T-bill yield, S&P 500 yield, 10-year treasury yield or inflation rate.
The primary reason I write this blog is to reexamine the variety of asset class trading approaches I’ve used for years. I don’t want to waste time implementing tools that appear useful, but ultimately fail to enhance risk-adjusted performance. In this post, I examine Shiller CAPE, which should have implications for using historical valuation charts in general and asset class valuation tools produced by firms such as GMO, Research Affiliates and many others.
The most popular stock market valuation tool right now is the Shiller CAPE ratio. The Shiller cyclically adjusted price-to-earnings (CAPE) ratio divides the S&P 500 price by the average of the past 10 years of earnings.1-3 The averaging helps smooth out earnings volatility associated with recessions. Investors love this model because it’s very intuitive and appears to be quite good at predicting future equity returns.
Over the past five years or so, I’ve been reluctant to overweight U.S. stocks in my discretionary portfolios due to the belief that U.S. stocks were significantly overvalued based on the Shiller CAPE ratio. Unfortunately, that hasn’t worked as U.S. stocks have been the top performing asset class over that period. (more…)
This is the first of a multi-part series examining the use of valuation approaches to identify future outperforming asset classes.
I discuss why value investing is an essential and useful tool for asset class traders.
I briefly discuss the Warren Buffett approach as the purest form of discretionary value investing.
I clarify the distinction between fair value (a concept I’ve used to discuss short-term trading edges) and intrinsic value used by value investors.
I briefly review the vast academic literature on mechanically value-tilted portfolios. These portfolios are typically heavily weighted towards cheap and risky assets. These cheap securities are most often fairly priced to deliver long-term returns that are superior to a market cap weighted basket of similar securities.
Finally, I introduce Asset Class Value Investing as: identifying moments in time when an asset class appears to be irrationally mispriced based on an assessment of long-term intrinsic value.
As asset class traders, we must understand the current thinking built into asset class prices. Prices are generally determined by the actions of a global set of large institutions, pension funds, endowments, sovereign wealth funds, banks, mutual fund managers and hedge fund managers considering future prospects and risks. Finding trading opportunities requires us to get into the heads of these large market players.
Many large firms specialize in one or a few asset classes, while other large institutions focus their attention at the asset class level. Most organizations are too big to actively buy and sell securities; they need to act on longer time frames.
The question of assessing an asset class’s relative value is particularly topical right now. U.S. growth stocks continue to outperform all other asset classes despite being “overvalued” for the past 5 or so years. Meanwhile, international and emerging market stocks continue to sink, despite trading at a fraction of U.S. equity multiples. Furthermore, across the global stock market, the value style has underperformed growth stocks for over a decade, despite strong evidence value stocks outperform over the long term. The current market seems a lot like the late 1990s all over again. (more…)
Large price runups, such as a gain of 100% over two years, are rare.1 In a previous blog post, I presented 10 attributes to distinguish asset class bubbles from large price runups that are justified by improving fundamentals. Those bubble attributes are:
Heavy retail investor involvement
Five or more years of swiftly rising prices
Parabolic rise in price
Shorting is unattractive or impossible
Product providers exploit excessive demand
Leverage fuels more buying
Bubbles are late-cycle phenomena
As asset class traders, we are especially interested in bubbles as a potential huge source of alpha when they collapse. As it turns out, bubbles are a lot tougher to exploit than it might seem. In this blog post, we’ll delve into bubble characteristics in more detail, and then investigate the best ways to trade asset classes that are experiencing a bubble.
We’ll examine bubble characteristics over the short term (plus and minus three years around the peak) and then longer term (a decade or more). When doing this sort of analysis, we need to at least acknowledge that various forms of hindsight bias can creep into such work since we are examining known price runups that ultimately crashed spectacularly.
It’s possible that we should include a few historical parabolic runups that did not ultimately pop.1 I’m hopeful that the lack of bubble attributes associated with these moves provides the justification for eliminating these from consideration, but I’m not completely sure. I may have also declared a few large up and down moves as bubbles (for instance, Chinese equities in 2007), when perhaps this price move had no more bubble attributes than a big move that ultimately didn’t end in a long period of underperformance (for example, the 1987 crash). (more…)
If you’re in the investment biz long enough, you’ll inevitably find yourselves searching for profitable ideas when an asset class is experiencing a bubble. The term “bubble” is a heavily overused term in the financial media and among professional investors. Any large price increase over a short time period, such as a 50% gain over a year, prompts a few writers, analysts or professional investors to describe the runup as a bubble. These bearish folks are typically using the term loosely without a nuanced evaluation to determine if prices have simply reflected new highly positive information.
Additionally, how many times have we heard “bubble” used for assets that have experienced long-term, secular bull markets, such as U.S. or Japanese government bonds, when current prices are not experiencing anything like the bubble phenomenon? Then other folks use the term in a variety of ways to describe investor group think, such as “hedge funds are the next investment bubble,” or “there’s currently a bubble in investor complacency.”
I’m specifically interested in those moments in time, which can last months or even years, when price-insensitive buyers become ever more attracted to rapidly rising prices, thus bidding an asset class price way above fair value. It’s a time when pricing is determined by the one-way thinking associated with the “madness of crowds,” rather than the normal, efficient markets “wisdom of crowds” effect. The massive number of performance chasers overwhelms the financial resources of professional investors and traders attempting to push prices back to fair value by shorting and selling the asset class.
Even though the gap between price and fair value grows ever larger, uncertainty about when the market will top and how much further the price will rise creates a situation where the risk-reward of shorting the asset class becomes very unattractive. The arbs, who are usually pushing prices back to fair value, then step away, or perhaps even join the crowd. Other opportunistic professional investors join the crowd by creating new products and/or new firms to exploit the attractive optionality associated with easy money raising and rapidly rising prices.
Bubbles have occurred about once a generation throughout human history, as new investors enter the market with no experience with how bubbles eventually burst. Examples of the large bubbles include the NASDAQ bubble in the late 1990s and the Japanese stock market bubble in the late 1980s. The current bitcoin and cryptocurrency craze has all the attributes of a bubble.
Many Rapid Price Increases Are Not Bubbles
Just because an asset class has a large and rapid price run doesn’t mean it’s mispriced. One or two U.S. industry groups often have a yearly gain of greater than 50%. Most often, prices have increased to reflect unexpected improved future earnings. Also, most bull market tops don’t have a bubble associated with prices. The topping process is more drawn out as the collective wisdom is formed by market players generally using reasonable judgment of future prospects. Prices ultimately collapse when investors sense an emerging bear market or recession.
Eugene Fama, the Nobel-Prize-winning champion of the efficient markets view, believes there is no such thing as bubbles. Bubbles are only known after the fact, when there’s been a large collapse. Careful analysis of large price increases shows that many, perhaps even half, are never followed by a collapse.1 Simple price formulas for defining a bubble, even the approach used by the famous bubble studiers GMO,2,3 are not enough. We need more information. (more…)
Often in trading we become totally engrossed in searching for short-term opportunities with a hyper-focus on news flow and daily price movements. Occasionally it’s good to drastically alter time frames, especially if your creativity has dried up on short-term ideas. One way to search for new trades is to scan asset classes that have performed the worst over the previous decade. This is especially interesting when there’s been a large divergence of performance in asset class returns over the previous 5 to 10 years.
Table 1 shows a ranking of the worst-performing ETFs by 10-year annualized returns as of October 31, 2017. When tabulating this ranking, I excluded the ProShares daily leveraging funds and commodity exchange traded notes. For comparison, the S&P 500 returned 7.51% per year during this time frame. This list contains many ETFs in the energy space, with a few niche asset classes (clean energy, gold miners, steel and nuclear), country funds (Russia, Italy and Brazil) and two currencies (British pound and Canadian dollar).
Table 1: Worst-performing U.S. ETFs based on 10-year annualized returns, as of October 31, 2017. (Note: S&P 500 return over same period was 7.51%/year). Source is Morningstar.
Often these asset classes were popular many years ago, but as the terrible performance persisted over a decade’s time, more and more traders and portfolio managers shifted their focus to better-performing asset classes. With a niche asset class, such as gold mining or solar energy stocks, the last remaining holders are the enthusiasts (gold bugs), index funds and retail investors owning such a small position that to them it’s easier to ignore rather than take a loss on a sale. (more…)
Does experience trading the markets provide an actual trading edge? Surely an investor who’s seen many bull and bear markets has an advantage compared to a novice just starting out in the field. The beginner is an alpha source for seasoned traders as the former pays their “tuition” associated with learning how to trade. The counterargument, using efficient markets logic, suggests that if a great number of portfolio managers (PMs) have significant experience, then the performance benefits of experience become arbitraged away as prices quickly incorporate the collective wisdom of the pros and experts.
Fund marketers ignore the efficient markets logic and advertise portfolio manager experience because it’s very convincing to most customers. I’ve traded asset classes for about 20 years. I’ve experienced two major and several minor bear markets, and seen a variety of bull markets. When I assess my current trading edges, I admit I’ve started putting “experience” as an edge – although at times with a question mark behind it.
Studies show that manager tenure and experience has little impact on mutual fund performance.1 I’ll speculate that the constraints associated with mutual fund portfolio management greatly inhibit the use of experience as an edge. Prospectus limitations on what securities can be purchased and the requirement to stick with an established investment discipline limit the flexibility to use experience to add value. Career risk can also drastically alter a PM’s personal risk-return profile, inhibiting the use of experience to benefit clients.
Seasoned relationship managers and investment advisors can be highly valuable to their clients since they draw on experience to help a client weigh the pros and cons of making a decision – especially in stressful moments. Of course, their job is not about adding alpha, although many claim they can. Just because a person has 20 to 30 years in the industry doesn’t mean they automatically have an “experience edge” that translates into superior performance. Many PMs are not intentional in how they learn from their experience. They’re lazy.
As an asset class trader, free to shift to any asset class and any investment style at any time, experience can become an impactful trading edge. Intentionally developing an “experience trading edge” requires a carefully planned personal mastery process. Much like a training program used by elite athletes, this mastery process is used to successfully play the asset class trading “meta-game” to better assess what approach works best at a particular moment in time. The mastery process is never-ending and consists of learning, practicing, recording results, reflecting and incorporating feedback to get better. To ignore the process is a huge missed opportunity to get better at this game, and perhaps develop an “experience” trading edge. (more…)
In the Market Tells – Part 1 blog, I introduced the concept of market tells as an enduring trading edge that can be used to enhance returns over the long term. Developing the ability to identify market tells and act on these signals takes practice. It’s important to keep track of trading results to get better and gain confidence to jump on these opportunities in size when they occur.
In this blog, I first clarify thoughts on a few instances when the market tell guidelines are not met. The most important guideline for identifying a market tell is the 80% rule, which states that to consider market action as unusual (and thus providing a market tell), the usual behavior must occur at least 80% of the time. Otherwise, it’s just too difficult to associate the unusual behavior with a market tell rather than normal market gyrations. Next, I’ll review three more market tell techniques that can be used to trade asset classes. I’ll close with some thoughts on crowded trades and implementation hints.
On some days, one sector may be up while the rest of the market is down. If the behavior is explained by news, then this price action is totally normal. Even without news, this sort of divergence occurs often enough to consider it normal activity. If a security is illiquid, such as a closed-end fund, then this action might be correcting a stale price from the previous day, which is not unusual at all.
To contrast, if all risky assets are aggressively moving higher (such as when the S&P 500 is up 2%+ for the day), and one sector is flat, then this behavior is much more unusual and may be the basis of a market tell, especially if no news explains the behavior. However, we need to remember that the time scale of this edge is on the order of a day or two.
Watching how a stock or asset class reacts to news associated with a known announcement date (such as an earnings call or a central bank news release) is very difficult to use as a market tell. There is usually an enormous amount of trader attention to these known release dates, which probably means there’s no trading edge. Much of the news and all the probabilistic outcomes have been priced in before the announcement, which is another way of saying the pre-announcement price is set such that there is no edge in buying or shorting the security ahead of the announcement, or in fading or chasing the post-news reaction.
Finally, watch out for interpreting short-term market action when there’s a lot of unwinding of positions. For instance, at the first of the year, it’s best to wait a week before interpreting market action as hedge funds unwind their end-of-year trades. This sort of behavior can occur during bear markets also. This is a very chaotic time when a lot of unwinding is occurring – it’s a very difficult time to interpret market action versus any sort of playbook. Wait until the dust settles and stocks start moving in unison before interpreting market action.
Divergences and Market Tells
In discussing market tells, I’ve used the term “divergence” to describe local price action where an asset class’s price or relative strength line is diverging from normal expectations. Searching for divergences is a staple of technical analysis.1,2 In the past I have falsely interpreted a divergence as a market tell. These misinterpretations generally occurred during relatively quiet times in the market, usually during bull markets, when the 80% rule did not apply.
Ask a market strategist about divergences, and the first thing that comes to their mind is the case where the stock market is making new highs while some other risky asset class is moving lower. For instance, from mid-2014 to mid-2015, the S&P 500 was moving higher, while junk bond markets where moving lower. This isn’t supposed to happen, and such a divergence is often interpreted as a warning of future weakness in the S&P 500, leading to a convergence of the two markets. (more…)
In any competitive field, the awareness of tells can provide a significant and enduring edge. Most people think of poker when they hear about tells.1 Is a player acting strong to encourage other players to fold? Does a player seem nervous when they throw their chips in the pot? Can any useful information about an opponent’s poker hand be gleaned from these actions?
Tells occur in many aspects of life and competition. For most sporting events, searching for tendencies in an opponent’s play is an integral part of game preparation. In the home arena, parents look for facial clues when interrogating a fidgety teenager as the youngster explains what she’s doing on a Saturday night.
Tells are an important source of feedback when trading the financial markets. I call these “market tells” to distinguish between the variety of tells that occur in other forms of competitive environments (more on this distinction later). A market tell is a powerful approach to sensing moments in time when market participants are not positioned correctly.
A simple example of a market tell is a stock that’s acting strong when it should be weak. Perhaps the stock is in an uptrend when extremely bearish news is released about the prospects of the company (such as a product recall). Unexpectedly, after a momentary dip on the news, the stock price continues to go higher. That’s a positive tell for future outperformance.
Properly identified, tells can get you out of a trade much sooner than waiting for a technical trend-following sell signal. Tells can help you identify future outperforming asset classes. They can provide positive feedback that a current trade is working. No matter what your trading discipline or time scale, searching for market tells is a great trading edge.
The more you clarify and develop your thoughts around market tells, the more confidence you’ll have to quickly jump on trades and to trade with high conviction and size. I began noticing market tells soon after I started trading in the late 1990s. In 2005, I decided to be more disciplined about it by keeping track of results. When I observed a tell, I printed out a chart and documented what I expected to happen, and then slipped the paper into my “market tells” folder. I also traded on this information, and over the next four years, avoided looking at the results.
In mid-2009, I checked to see how these predictions worked. The success rates for these trades at both the short and intermediate time scales were excellent. These results piqued my interest to further investigate and refine the use of market tells in my trading. In 2014, I performed yet another trade analysis, and in 2017, I find myself slightly reshaping my views while writing this blog.
Classification of Tells
To provide a framework for thinking about market tells, I’ll review various forms of tells. (more…)
Does momentum really work? Can you outperform benchmarks by simply holding top-performing securities? I have my doubts, but I’ve decided to be open-minded and revisit this idea.
A decade or two ago, we used momentum extensively to select equity asset classes and avoid the poor performers. I lost faith in momentum somewhere in the 2009 to 2010 time frame after the premium suffered a historically large drawdown in 2009.
We turn off momentum at the beginning of new bull markets, so we didn’t suffer from this drawdown, but hedge fund assets had grown exponentially to over $2 trillion at the time, and momentum mutual funds were being introduced. The strategy appeared to be very crowded.
Momentum was a great trading edge for many decades. Now it’s just too easy to do and too popular with an enormous amount of assets implementing momentum in various forms.
Let’s define the momentum effect. The idea is that top performing assets over a 3 to 12 month time frame tend to outperform over a similar time frame. Poor performing assets tend to underperform in the near future. Momentum is sometimes confused with trend following.1 Momentum ranks recent performance among peers, such as ranking U.S. stocks among each other, or U.S. sectors among each other, or individual country stock market indices among each other. Trend following looks at absolute prices and asks if prices are in an uptrend or downtrend, and shifts to cash or shorts an asset when prices are in a downtrend. Momentum indices remain exposed to falling prices, and thus can suffer large losses during bear markets. Long-short momentum portfolios can also suffer enormous equity-like drawdowns.
A typical momentum measure is 12-month performance, although the sweet spot look-back period can vary from 6 to 15 months. For really short time periods, one month or less, there’s a short-term reversal effect where top performers typically underperform during the next month. 2-4 As the look-back period extends beyond a few years, recent top performers begin to lag the averages in the future and the bad performers tend to do better as the latter are now underowned and better valued with higher expected returns. (more…)