Excess Stock Returns, Technical Analysis & Market Efficiency

Introduction

This paper addresses two basic questions through a brief literature review. The first question is whether the empirical evidence available supports the predictability of stock returns using technical analysis. The second question asks for a critical evaluation of whether return predictability is a good indicator and test of market efficiency. While some investors can generate strong returns even in the most volatile economic climates, there is generally a correlation between overall market performance and investment returns during the same period, based on the totality of the evidence reviewed here.

Literature Review

Asness et al. (2013) approach both questions in a fairly systematic way. They assert that there is a "value" effect whereby the long-term view of an investment is compared to its current value. Concurrently, there is a shorter-term phenomenon called the "momentum" effect, which tends to indicate what will happen in the near term. It is also noted that these two ratios are typically examined separately rather than together. One interesting assertion in the Asness text is that there are only modest links between some macroeconomic variables compared to others. For example, the business cycle, consumption, and default risk are not seen as having strong correlations with returns, whereas negative liquidity and a few other variables do. In other words, assessing potential returns is about identifying the right variables and disregarding others. The common practice of examining short-term and long-term returns in isolation, rather than together, is therefore less than ideal (Asness, Moskowitz & Pedersen, 2013).

Johnson (2002) makes a similar argument, noting that the market sometimes under-reacts relative to a particular firm's fundamentals and the excess returns it has offered or could offer. The core of Johnson's argument is captured in his statement that "the case for rational momentum effects is not hopeless." He continues by noting that "the key to the model is stochastic expected growth rates," and rounds out his position by observing that growth rates and expected returns are positively correlated. This does not mean that strong returns cannot be found when common economic indicators are unfavorable. However, the probability of this occurring is decidedly lower than it would be if the underlying economic conditions were healthier (Johnson, 2002).

When it comes to market efficiency, Kim, Shamsuddin, and Lim (2011) address the Efficient Market Hypothesis (EMH). The hypothesis holds that all relevant information about a firm and its profit-generating ability is publicly available, making it exceedingly difficult — if not impossible — to achieve consistently higher returns, since there is no hidden or missing information to be uncovered or guessed. This idea was advanced roughly two generations before Kim et al. and has since been found by many researchers to be widely non-credible in its strongest form. Analysts and traders routinely compete and pay significant sums for information advantages. Some succeed, some do not, and some simply have access to insights that others lack. Grossman and Stiglitz (1980), as cited by Kim, argue that a perfectly efficient market is not possible. Relative efficiency is a more defensible position: the market generally aligns with the returns generated within it, but there will always be outliers in terms of information completeness and overall performance. Some firms perform well even when the broader market or economy is struggling, and the reverse is equally true. Kim also points to the work of Lo (2004), among others, in supporting the view that market performance and overall returns are generally correlated (Kim, Shamsuddin & Lim, 2011).

The work of Fama (1998) examines how stock and other financial markets react to information as it becomes publicly available. Fama notes that short-term market reactions to news, when they occur at all, are usually brief and limited in scope. He also observes that markets tend to over-react to information, but that over-reactions and under-reactions will roughly cancel each other out over time, so that the returns on any given investment tend to align with overall market conditions. This is not to say that investors cannot exploit a quick price spike or dip. However, doing so consistently is extremely difficult unless one has advance knowledge of what is coming — which is generally impossible, and in many cases illegal.

One problem Fama identifies with many tests of this subject is that relatively few examine the opposite of market efficiency. Market inefficiency is often not treated as an outcome variable in its own right, likely because — as Fama notes — doing so requires one to "specify biases in information processing that cause the same investors to under-react to some types of events and over-react to others." In other words, it is difficult to explain why investors react differently across event types and why their returns diverge from what market conditions would normally predict. This is a considerably more complex undertaking than simply assessing whether returns and market conditions are in sync — which they are, more often than not.

Fama also points to the work of De Bondt and Thaler (1985), who found that prior winners are not necessarily future winners; indeed, the opposite tends to be true over windows as short as two to five years. There appears to be a protracted market reaction to earnings and related metrics, though the timeframe over which those reactions unfold seems to be shortening. Fama notes that this window was approximately one year in 1980, but has since contracted to roughly three to six months. Even accounting for these outlier reactions, Fama broadly supports market efficiency as the general operating principle of financial markets (Fama, 1998).

The final source reviewed is De Bondt and Thaler's (1989) work on mean reversion and Wall Street. The authors observe that economics is distinctive among the social sciences in that it is grounded in the assumption that actors behave rationally given the totality of available facts. They further assert that financial markets are sufficiently efficient that the stated values of firms — as recorded in financial statements and held in the minds of investors — closely match those firms' intrinsic values. A review of Dow Jones stock prices in the late 1950s and early 1960s found a strong correlation among thirty stocks in terms of daily fluctuations. However, the authors of that study also noted that the magnitude of those fluctuations was exceedingly small and likely insignificant, limiting the practical meaning of the finding. Even so, short-term returns do diverge from long-term patterns and, at least in some respects, from what publicly available news about a firm would seem to imply.

Regarding technical analysis and its link to excess stock returns, it would seem that there is no fully reliable method for identifying excess returns through conventional analysis alone. There are many factors not well captured in the literature reviewed that could drive a shift in firm performance over time. A useful illustration is the widely criticized decision by Netflix to raise its prices for DVD and Blu-ray rentals. The price increase — a few dollars per month — prompted a strongly negative public reaction, and the company's stock likely suffered at the time. Investors scrutinizing the decision purely in terms of customer churn and near-term revenue loss might have concluded that Netflix was in trouble. However, those who looked more deeply would have recognized that Netflix was deliberately steering customers away from physical disc delivery toward streaming — a far more efficient and scalable revenue model. Whether this qualifies as "technical analysis" in the strict sense is debatable, but the broader point holds: technical analysis conducted properly will generally correlate with returns, though those returns will rarely be exceptional. The key for investors is to develop an understanding of why a business is making a particular decision. Many observers believed Netflix's price hike was a serious mistake, and they were wrong. Standard technical analysis would likely not have revealed that Netflix was making the right long-term move, but investors willing to think beyond conventional metrics would have been handsomely rewarded (Scott, 2011).

As for whether return predictability is linked to market efficiency, the answer is a fairly confident yes. If relevant information is available, then the returns that follow will tend to be similarly predictable. The one significant caveat is that standard technical analysis is not always adequate for forecasting outcomes, and markets are sometimes far from efficient. One need look no further than the Enron scandal. That company went from apparent strength to complete bankruptcy — with its executives indicted — in roughly a year, and conditions were already deeply problematic well before the truth became public. To be fair, that was a stark anomaly and such cases are rare. However, anyone who asserts that all information necessary for sound investing is always "out there" is overstating the case. Identifying the right analytical lens, as in the Netflix example, is one thing; when the playing field itself is being deliberately distorted, as in the Enron case, a different set of risks entirely comes into play.

It is generally true that returns, market conditions, and the overall performance of a firm move in lockstep. However, there are always exceptions to this rule. Ford, Apple, and Netflix have all traveled paths on which analysts actively criticized their strategic decisions — yet all three have performed strongly. Analysis is never entirely reliable or uniform, and there will always be excess returns and anomalies that defy conventional prediction.

Asness, C., Moskowitz, T., & Pedersen, L. (2013). Value and momentum everywhere. Journal of Finance, 68(3), 929–985.

Conclusion

Scott, J. (2011). Everyone needs to just calm down about the Netflix price hike. ReelSEO. Retrieved March 6, 2015.