Value investing – does a rules-based approach work?

  • Written by 

    Written by Christian Theis, CFA - June 2014

High-profile value investors like Warren Buffett have consistently demonstrated that, when performed with virtuosity and luck, fundamental analysis can be a means to outperforming the market. But can such an approach be turned into rules that benefit the average investor?

Tactical asset allocation review: why sell in May? 4 of 4 The bond-yield conundrum 2 of 4

Residents of the United States, please read this important information before proceeding

Value and growth investment styles

The concepts of value and growth investment styles can be traced back to the teachings of Graham and Dodd at Columbia Business School prior to the publication of their hugely influential 1934 college textbook “Security Analysis”. The growth investment style directly corresponds to the idea of identifying those “growth” stocks with earnings growth rapid enough to allow them to beat the market. Benjamin Graham’s investment philosophy on the other hand, asserts that it is difficult for investors to beat the market, and consequently just aims to mitigate the risk of underperformance due to misjudgement, i.e. overvaluing or undervaluing a stock due to the “tides of pessimism and euphoria which sweep the market”. To reduce this risk, he suggests that investors determine the “intrinsic” equity “value” that is justified by a firm’s assets, earnings, dividends and financial strength, and buy those “value” stocks that appear underpriced.

Graham never fully explained how to determine “intrinsic” value, but considered a firm’s assets a particularly important component. Other factors included earnings, dividends, financial strength and stability. Based on this, value stocks are associated with low price-to-book and price-to-earnings ratios as well as high dividend yields. Similarly there is no single definition of a growth stock, but indicators include high earnings-per-share growth rates or implied internal growth rates (return on equity multiplied by retention rate). Valuations are not taken into account at all, even if they are already relatively high and might seem to be pricing much of that growth in.

The long-term case for value investing

It is now common practice for index providers to divide their investment universes into value and growth stocks, constructing corresponding style indices that, when combined, form the overall market. Financial ratios, such as trailing price to book, have long track records of identifying value stocks, while forward-looking estimates for earnings growth only became commonly used in the 1990s. So historically this started by identifying value stocks and defining growth stocks as the remainder. Methodologies have been refined over time, but still aim for a 50/50 value/growth split and have to deal with the fact that in practice stocks can exhibit both value and growth characteristics or neither. This means that the various approaches have to have room for compromise and approximation. Summing up to the overall capitalisation-weighted market further implies using market capitalisation-based weightings, not style-based weightings such as earnings or dividends. Apart from rebalancing effects, the outperformance of one style implies underperformance of the other.

Figure 1 shows the performance of the MSCI World index alongside its style indices. There are significant regional differences in the style characteristics of stocks (for example US stocks have traditionally quite low dividend yields), so these indices are first constructed on a country or regional basis and then combined to derive global indices. Since December 1974, value has outperformed growth by 2.6% annually, with lower risk. This outperformance on a risk-adjusted basis is the so-called value premium that Eugene Fama and Kenneth French first identified in 1992 and incorporated into their famous three-factor model.

While academic research has shown that value premiums have historically existed in many markets, Figures 2 and 3 demonstrate that the relative performance of value versus growth is quite divergent between countries and regions, and there is no single global ‘style’ factor influencing all regions similarly. Regarding volatility, Figure 5 shows the large span that value versus growth volatility has displayed. One factor contributing to these differences are varying sector weights, as shown in Figure 4.

Over the long term, value has clearly outperformed growth

Value indices currently have the largest overweights in financials, energy and utilities. Whilst the US is roughly neutral for the weights of industrials and consumer staples, these are the largest underweights in Europe. Figures 6 and 7 show that, in the past, sectors have demonstrated extremely volatile style characteristics. For example, consumer discretionary, previously a value sector, has had a higher price-to-book value than the overall market during the last few years. During the financial crisis even consumer staples exhibited higher earnings “growth” expectations than the overall market.

A closer look at the more recent history

A closer look however shows that the overall success of value strategies derives mainly from the 1970s and 1980s. Figure 2 shows that in the US, value has underperformed growth for over 25 years since peaking in July 1988. Globally, value experienced a 30% setback in the late 1990s so that there are now periods with a length of nearly 13 years over which growth has outperformed. For all regions shown, growth has outperformed value since 2007. Over the last 10 years – MSCI methodology was significantly changed in 2003 – the global annual outperformance of value over growth is 10 basis points only and does not cover the higher transaction costs of maintaining style indices: over the last year the MSCI World Value index had a turnover of 19.7% compared to 2.4% for the overall market. Can different index construction methods improve upon this performance? Figure 9 demonstrates, for the S&P 500, the difference between ordinary broad style indices and narrower ‘pure’ style indices that only include stocks with clear style characteristics using style-based weighting. In comparison the broader value and growth indices are very close to the overall market index in absolute performance terms, while the narrower pure value and pure growth indices have both noticeably outperformed the market. That outperformance has, however, come with more risk. Historical volatility of the pure style indices has been 21-22% compared to 16% for the market. As a result, Sharpe ratios for the pure style indices are only marginally higher than the market’s (0.36 vs. 0.41/0.44). From a maximum drawdown perspective, the 69% drop of pure value during the financial crisis exceeded the 51% drop of the overall market.

Figure 8 and 10 show the relative performance and rolling annual relative returns for the style and pure style indices. The charts demonstrate that relative performance has been significantly more volatile for the pure style indices as well; with peaks around 80%, the relative annual performance has been nearly double the amount of the ordinary style indices. Regarding absolute volatility, Figure 11 shows that also for pure style indices there have been both times where value was more volatile than growth and the other way round. A clear outlier is the volatility of pure value during the financial crisis, which peaked at 50% and reflects the large drawdown of this pure style index.

In the last 10 years, relative performance between value and growth indices has been rather muted, offering little attraction to rules-based investors. Pure style indices, which are less investable, come with higher absolute performance, but at the price of higher risk.