1. INTRODUCTION

Value investing is a strategy that buys cheap stocks. Cheapness is measured by the ratio of the price to the fundamental value of the stock. An example of such, and probably the most commonly used value metrics, is the book-to-market (BM) ratio - the higher the ratio the cheaper the stock. The value strategy has a long tradition and it dates back to the 1930s when Graham and Dodd (1934) published the popular book “Security Analysis”. The study of Basu (1977) [1] marks the beginning of a more recent, extremely rich academic value literature. The seminal work of Fama and French (1992, 1993) [2]  [3]  and the corresponding 3-factor model have established value (and also size) as one of the key risk factors explaining cross-sectional equity returns. However, value investing is not only limited to equity asset classes, as Asness et al. (2013) [4] show. In their study, they find value (and momentum) premia across several asset classes and markets. Moreover, these returns across asset class exhibit strong co-movements for value (and momentum).

Importantly, value investing has also been a very successful investment strategy over the last century in practice. Most prominently, Warren Buffett’s success can, at least partly, be attributed to the value premia as Frazzini, Kabiller and Pedersen (2013) [5] show (low-risk and quality are the other important factors according to the authors). Next to Buffet, many asset managers utilize value either as a standalone strategy, or in combination with other factors, for example, momentum.

Like any other strategy value investing is primarily a selection strategy (stock A vs. stock B) at any point in time and it has to be differentiated from a value oriented market timing (stocks vs. cash)/time-series value approach.

2. PRACTICAL IMPLEMENTATION

Private investors gain most easily access to value by simply buying passive value ETFs, which are cheaply available for most major equity indices. Most ETF providers offer cheap value solutions, mainly tracking the MSCI value indices.

Constructing a value portfolio involves the same logic as the other factors do, so we only give a brief overview of criteria that could be used.

The value portfolio similar to the one suggested by Fama and French (1993) [3] can be constructed by sorting stocks with respect to their book-to-market value. However, practitioners usually use a combination of several value measures. Let us take the MSCI value methodology as an example. MSCI value indices combine three measures, namely, the book-to-price ratio, 12-months forward earnings-to-price ratio and the dividend yield. The overall value measure of a company is then the sum of the three z-scores. Z-scores are obtained by standardizing each of the variables ($\frac{x-\mu}{\sigma}$). Since most fundamental data contain some outliers, the MSCI methodology employs the winsorization of the data. This means that the values, which exceed the 95th (are lower than the 5th) percentile are replaced by the value of the 95th (the 5th) percentile value. The final index is then obtained by filling up the portfolio with the cheapest stocks starting from top, once 50% of the free-float adjusted market capitalization is reached, no more stocks are added to the value index.

More active investors aim to add (literally) more value, by using more than just three dimensions and by selecting a more concentrated value portfolio, by including less stocks and/or apply some different weighting scheme. Another hint to improve classical value approaches can be found in Fama and French (2012)[6], which indicates that the value premium (and also momemetum) is more pronounced for smaller stocks.

Another interesting paper is Asness and Frazzini (2013)[7] --- they show the importance of using unlagged price information when forming book-to-market ratios. Originally, Fama and French (1992) [2] form value portfolios in June of each year, based on B/M information prior December 31st the previous year. The results indicate, that monthly updating of the price information delivers a significant increase in five-factor alphas in contrast to the yearly approach followed by the original study. This, of course, has practical implications, if a monthly updating and also rebalancing is truly more profitable, as the results suggest, one might ask if an ETF based on semi-annual updates can fully capture the value premia.

For non-equity asset classes, where no fundamental data are available, value can be proxied as in Asness et al. (2013) [4] by using the negative of the 5-year return. The ranks of the worst performers, measured over a period of 5 years, show, in the case of equities, a high correlation with ranks based on BM ratios.

Value is an attractive stand-alone strategy. However, a combined factor strategy with quality or momentum delivers significant improvements, in particular, with respect to risk characteristics.

3. RISKS INVOLVED

Value investing can be difficult for short-term oriented investors, as the past has shown pronounced periods of depressed value returns. For example, prior to the bursting of the dot-com bubble in the beginning of the century, or during the financial crises in 2008 value investors had to endure years of underperformance.

If the rational theory is right, value is painful when you already feel pain (see below). Hence, the risk you face is that, during bad times, you are hit particularly bad by being exposed to value risk.

4. THEORETICAL EXPLANATIONS

Ang (2014) [8] summarizes possible explanations of the existence of a positive value premia. Similar to other factors, there are rational and behavioral theoretical approaches.

The rational theory of the value effect boils down to the idea that all value stocks or stocks with value characteristics contain exposure to a systematic value factor, which cannot be diversified away, similar to the systematic equity market risk. A positive risk-premia of the value factor can be justified, when the value factor performs particularly bad during bad times of the average investor. In finance theory these bad times correspond to a state of the economy, where a decline in asset prices is on average more painful than in good times (see consumption example below). The value premia is simply a risk compensation for holding value during these periods. Zhang (2005)[9]  formalizes the theory in a neoclassical framework. During bad times, value firms face the problem that their capital is less productive and adjustments to the capital stocks are difficult. Hence, value firms perform worse during these periods. On the other hand, growth firms are more flexible, but face higher adjustment costs during booms. Therefore, they tend to underperform during booming markets. These characteristics are priced, when investors have a concave utility function. This becomes clear, when we apply the argument of marginal utility in consumption based asset pricing world:  value performs worst when marginal utility is high (low consumption level) and growth performs worst when marginal utility is low (high consumption level). Using a more simplistic wording, value hurts when most already suffer from a negative economic environment.

One behavioral explanation assumes an overextrapolation of recent news. Strong earnings growth is assumed to persist, while deteriorating earnings probably continue to get worse. Hence, prices are formed on biased expectations. If future earnings turn out to disappoint in case of growth stocks and, on the other hand, surprise positively for value stocks, then the price reversal occurs, driving the prices of previously “overpriced” and “underpriced” stocks to fundamental values.

If the value premia indeed boils down to the behavioral defects, Ang (2014) [8] asks the right question, “why don’t more investors buy value stocks and, in doing so, push up their prices and remove the value premium,...”  

5. SUMMARY

Value has historically been a successful investment approach, however, at a price of pronounced periods of depressed returns. Hence, risk reduction plays the key role for value investors. One way to achieve it is the combination of value with other strategies, such as momentum and quality, as it helps to drastically reduce the length of depressed return periods and drawdown risk. Cost-effective (long only) value exposure can easily be gained through ETFs for most important developed markets/regions, long-short value exposure is typically more pricy and mainly offered by specialized active asset managers.

References

1. Investment performance of common stocks in relation to their price-earnings ratios: A test of the efficient market hypothesis,
   Basu, Sanjoy
   , The Journal of Finance, Volume 32, Number 3, p.663–682, (1977)
   
2. The cross-section of expected stock returns,
   Fama, Eugene F., and French Kenneth R.
   , The Journal of Finance, Volume 47, Number 2, p.427–465, (1992)
   
3. Common risk factors in the returns on stocks and bonds,
   Fama, Eugene F., and French Kenneth R.
   , Journal of Financial Economics, Volume 33, Number 1, p.3–56, (1993)
   
4. Value and Momentum Everywhere,
   Asness, Clifford S., Moskowitz Tobias J., and Pedersen Lasse Heje
   , The Journal of Finance, Volume 68, Number 3, p.929–985, (2013)
   
5. Buffett’s Alpha,
   Frazzini, Andrea, Kabiller David, and Pedersen Lasse H.
   , National Bureau of Economic Research, (2013)
   
6. Size, value, and momentum in international stock returns,
   Fama, Eugene F., and French Kenneth R.
   , Journal of Financial Economics, Volume 105, p.457–472, (2012)
   
7. The devil in HML’s details,
   Asness, Cliff, and Frazzini Andrea
   , The Journal of Portfolio Management, Volume 39, p.49-68, (2013)
   
8. Asset Management: A Systematic Approach to Factor Investing,
   Ang, Andrew
   , (2014)
   
9. The value premium,
   Zhang, Lu
   , The Journal of Finance, Volume 60, p.67–103, (2005)