Herding Behaviour Also Applies to Fundamentals

Adding technical rules can still be beneficial


The latest buzzword in index investing is “Smart Beta” and the approach is gaining traction as a prudent way to approach equities. In most cases, “Smart Beta” is factor-based and includes both new and old ways to approach stocks. Judging by the number of newly published articles surrounding the topic, it seems as if interest is picking up. This article discusses if a factor-driven investment approach creates a herd behaviour, potentially exploitable by investors using an approach based on past performance.

Most Smart Beta strategies are based on rules and require active rebalancing and substantially more trading activity than pure passive indices. Hence, factor-based investment strategies occupy the twilight zone between active and passive investment strategies. Labelling factor investment as “smart beta” is probably a more adequate reflection of the underlying activity. It is active investing for investors that perceive themselves as passive investors. For active investors, it is passive strategies that seek to extract a well defined risk premium from the market.

Factors are commonly defined by mapping certain characteristics, such as size, value, beta, dividends or any other attribute that is supported by empirical research onto a given set of securities. By investing in the factors, rather than the underlying indices, investors seek to generate excess returns by owning stocks with positive expected return and avoiding stocks with negative expected return. In hedge funds, especially equity market-neutral strategies, equities with expected negative returns are usually sold short, both seeking returns but also to stay neutral to the underlying equity market.

Value and small caps
In the early ‘90s, Eugene Fama and Kenneth French [1] argued that factors beyond the market risk (beta) were needed to properly explain stock market returns. Their original framework focused on size and value versus growth. Largely simultaneously, other authors introduced additional factors, such as momentum-driven factors [2] or emphasised the performance of low-beta stocks [3]. The long-term return efficiency of the underlying factors is surprisingly strong. Factor-based strategies are today typically available, not only to the more sophisticated institutional investors, but also to average retail investors through mutual funds, UCITS and ETFs.

In this article, instead of proposing a new and innovative factor, we take a different approach. Herein we explore the behaviour of several popular factor-based investment strategies and if we can observe herding behaviour (AKA trend following). Simply stated, does investor behaviour by buying and selling certain equities create an observable momentum effect?

Our basic belief is that investors exhibit herd mentality, based on other people’s success, driven by the strong returns from a specific factor or motivated by deeply ingrained survival instincts. Herd behaviour has been documented in most otherasset classes. If it was not evident in equity factors the implication would be either that equity investors are not investing into the factors to the extent that we believe, or that factor-based investors share a materially different modus operandi to other investors.

For the first part of the article, we use the traditional and well known Fama French factors: size (“Small minus Big”), value (“High minus Low”) and momentum. In the second half of the article, we use our own factors to verify the conclusions.

Small minus Big is a factor that ranks the investable universe based on size and buys shares in companies that have a market value below the market median and sell shares that are above the median. This factor reflects the excess returns from overweighting small cap and underweighted large cap stocks against a market capitalisation-weighted benchmarked portfolio. The portfolio is controlled for value and growth factors. The High minus Low seeks to own companies that have book value equity to market value equity (i.e., value stocks) that are above the 70th percentile of the universe while not owning the lower 30th percentile. This portfolio is thus long value and short growth against an index. The momentum portfolio is constructed in a similar way to the High minus Low but rather than seeking value, it seeks to have exposure to shares that have the strongest performance over the prior 12 months excluding the current month. The two latter portfolios are controlled for size exposure to avoid extracting small cap premium.

Raw price momentum signals
The momentum factor has traditionally been shunned by value or small cap investors as this factor only seeks to invest in shares that have exhibited strong relative returns and divest shares with the weakest relative returns. For a fundamental investor, this seems a bit too simple to work, as it does not include any other information than the historical price of the company. As shown in Fig.1, the average return for momentum-based equity investments has steadily declined after 2000, and for the last five years the factor has produced negative returns. Whether this is a cyclical low, as some authors argue [4] or if the factor has undergone a regime change, is hard to answer. Other Fama-French factors have worked better.

The three different portfolios are cash-neutral, i.e., an investor can finance the long side of the portfolio using the short side (for a benchmarked investor, the over and underweight could be achieved by only buying the long side of the portfolio, thereby automatically being implicit short against the benchmark). While there are some practical hurdles to implement the market-neutral portfolios, they represent the excess returns for long-only investors that can over and underweight stocks in a portfolio. As such, they can be used to explain excess returns from investors that have the benefit of ownership.

Price momentum of factors
To test our assumption about investor momentum, we built a naïve trend-following strategy based on the recent performance of each factor. For simplicity, we use a standardised moving average, ignoring transaction costs and market liquidity. We vary the length of the moving average and calculate the efficiency of the factor if the trend is positive or not. In the real world, there are additional complexities that need considerations such as transaction costs and restrictions in terms of shorting equities. At this stage, we are only interested in finding out if the underlying factors exhibit profitable momentum or not.

We find that, on average, the shorter the length of the moving average, the more we improve the performance of the factor. For longer moving averages, the result is slightly more mixed and the momentum effect declines the longer the time horizon is. However, since we are ignoring the transaction costs, we can probably expect that the slower strategies are more realistic once we
reduce the turnover by extending the length of the moving average.

Somewhat surprisingly, we find that the largest improvement is for the classical value factor, where we can see the largest improvement from the technical rule, across all time horizons. Paradoxically, while value investors typically ignore the so-called “technicals”, it seems as if the subset of stocks that are considered value, trend better.

More importantly when using a technical rule, we need to be less concerned with the actual sign of the underlying risk factor. If we undergo a regime change and the sign of the factor changes from positive to negative, due to changes in investor demand or structural market changes, the technical rules will tend to pick it up. There are other costs and risks involved with using technical trading rules. In general, such rules introduce model risk and usually a higher turnover, thus increasing transaction costs. Equity factors also contain similar risks.

Proprietary factors
Like most other research-driven investment firms, we have created our own factors. We have made other choices, for instance by keeping the factors strictly sector and beta-neutral over time, while ignoring cash constraints. Fama French factors are only generally cash-neutral and while controlling for size, they do not control for sector or beta exposures. While our own factors include additional drivers, they also include Fama French-inspired factors such as momentum or size. In contrast to the original factors that were built on US data, we built our factors across liquid European stocks. The different universes provide an interesting robustness check.

To verify the prior conclusion that factors do trend, not only in US but also in Europe, we implement the same trading rule across our factors. Our results are compatible with prior observations and we find that the factors do trend and that the same phenomena can be observed throughout our investable universe. The time-based time horizon seems to hold, also for European equities. Noteworthy is that while the Fama French factors include data from 1926 to the present, our factors only cover the period from 1999 to today and there can be substantial time variation in the efficiency of both the factors and technical overlay.

The results in Fig.2 indicate that investors can improve factor efficiency based on simple momentum strategies. In addition, the results have positive implications for strategies that seek to remain equity market-neutral. Remember that the European factors are beta and sector-neutral and can thus be implemented as a true equity market-neutral strategy, preserving a low exposure to equity markets. It is encouraging to see that even when going from cash-neutral to beta and sector-neutral, the results are still valid.

While the inventors of the original equity market factors sought to explain the efficient market hypothesis and argued that one cannot improve returns by deploying trading rules based on historical performance, we have managed to find one simple technical rule has improved the excess returns available to investors.

We believe that this indicates that investors in equities are sensitive to price and exhibit herd mentality, like any other group of investors. By adding filter rules to traditional equity market factors, we can improve the returns from market-neutral factors. An alternative explanation is that factor investors are insensitive to price trends (with the exception of momentum-based factor strategies) but that irrational investors are sensitive to price momentum and will infuse additional momentum in equity markets, thereby benefiting more rational factor-based strategies. It is noteworthy that when we combine a typical fundamental factor, value, with technical rules, we see the largest improvement.

Implementing the technical rules above requires computation power, market access, creative investment professionals and the capability to execute with minimal market impact. While we accept that it is not easy to extract excess returns out of relatively efficient equity markets, we certainly believe it can be achieved, provided that an investor has access to the right technology and people. Even fundamental investors can benefit from adding sound technical trading rules to strategies. In our view based on quantitative research, adding technical rules to fundamentally driven equity strategies is beneficial.


  1. ‘The Cross Section of Expected Stock Return,’ Eugene F. Fame, Kenneth R. French. 2, s.l. : The Journal of Finance, 1992, Vol. 47.
  2. ‘Profitability of Momentum Strategies: An Evaluation of Alternative Explanations,’ Narasimhan Jegadeesh, Sheridan Titman. 2, s.l. : The Journal of Finance, 2001, Vol. 56.
  3. Falkenstein, Eric, Mutual Funds, Idiosyncratic Variance, and Asset Returns. (Evanston, Illinois : Northwestern University, 1994).
  4. Larson, Ryan, ‘Hot Potato: Momentum As An Investment Strategy,’ [Online] Research Affiliates, August 2013, [Cited: October 23, 2013.] http://www.researchaffiliates.com/Our%20Ideas/Insights/Fundamentals/Pages/F_2013_08_Momentum_Factor.aspx.