Is value a buy? But
what sort of value?
We outline the case, both for and against, buying the value factor book yield (another name for book-to-market). We then discuss alternative ways to implement value that might have more appeal for investors with less tolerance for value’s inevitable drawdowns.
The value thesis
Value stocks can be thought of as being cheap relative to their intrinsic worth. It is almost an article of faith among quantitative investors that the value factor outperforms over the long term. There is a wealth of empirical and academic evidence supporting this assertion, with some backtests in the US going back as far as 1926 (see Chart 1). Explanations for this outperformance include:
- investors’ over-extrapolating the recent poor performance of value stocks meaning they then miss out on any subsequent mean reversion;
- investors’ tendency to overpay for ‘expensive, glamorous’ growth stocks, which in many ways are the opposite of cheap, financially distressed value stocks; and
- the fact that value stocks are sometimes neglected by analysts and investors leading to pricing inefficiencies.
Nevertheless value can underperform, or move sideways with volatility, for considerable periods of time and we are most certainly experiencing one of those periods at the moment. For example, as shown in Chart 1, the performance of value in the US has been very disappointing since February 2007, shortly before the start of the global financial crisis.
Value can underperform, or move sideways with volatility
The ‘valuation’ of value
It is informative to look at the cheapness of the value factor compared to history. We do this using a valuation spread, which is defined at a point in time, as the average book yield of the 20% of stocks in a global universe with the highest book yields; divided by the average book yield of the 20% of stocks with the lowest book yields.
This ratio is then plotted over time. The higher the ratio, or spread, then the cheaper value stocks are relative to non-value growth stocks.
Conversely when this ratio is low, and spreads narrow, non-value stocks, with their superior growth rates, are cheap relative to financially distressed value stocks. When this occurs, it is often problematic for the future outperformance of value.
Chart 1: Long term performance of value in the USCumulative long/short performance log scale
Source: Kenneth R. French Data Library, FactSet & Aberdeen Standard Investments (as of February 2019)
Chart 2: Value is cheap relative to historySource: FactSet & Aberdeen Standard Investments (as of February 2019)
From Chart 2 we can clearly see that at the peak of the Dotcom boom in early 2000, spreads were wide and value cheap. When the boom turned to ‘bust’ and non-value, internet-related growth stocks plummeted in price, value went on to outperform for several years (see book yield in Chart 3). Furthermore, in 2006, prior to the global financial crisis, spreads were compressed, which foretold a period of value underperformance. And now, once again, value is cheap by historical standards.
Headwinds for future value outperformance
Despite value’s current attractiveness and the strong economic rationale for its outperformance over the long term, there are headwinds against a value renaissance.
First, it might be the wrong time in the economic cycle for value to outperform. Value is a cyclical risk factor that generally performs best during periods of increasing economic growth and risk-seeking behaviour by investors. But as the economy slows, it is non-value growth stocks that tend to outperform due to genuine growth opportunities being in shorter supply. If, indeed, we are late-on in the current economic cycle, then this could be an issue for value.
Furthermore, quant practitioners and academics are increasingly questioning the efficacy of book yield as a measure of value. Book yield is book value divided by market value where:
book value = tangible assets - depreciation - liabilities
But nowadays, company valuations depend more on intangible assets such as intellectual property and trademarks, which are not usually captured by book value. In addition, book yield is intrinsically biased against companies with capital-light business models that can deliver growth without commensurate increases in input costs.
The value continuum
So far, we have only talked about value in terms of book yield. But we believe it is useful to view value as a continuum, as shown below. It ranges from cyclical measures of value on the left-hand side to those value factors that also possess quality attributes on the right-hand side. Crucially, each of these value factors has its own risk-return profile.
Next, we will look at one particular ‘value-quality’ factor, free cash flow yield, which has some very attractive properties.
Free cash flow yield
There is no standard definition of free cash flow yield (FCF yield), but the definition we use is as follows:
FCF yield = (cash flow from operations – capital expenditure– dividends) / market capitalisation
Note that dividends are viewed as ‘fixed’, in the sense companies are reluctant to cut them.
Chart 3 shows the results of a backtest of FCF yield on a global universe. Here, we go long the quintile of stocks with the highest FCF yields, funded by shorting the quintile of stocks with the lowest FCF yields. Rebalancing takes place every quarter. A backtest of book yield is also displayed for comparison purposes.
Examining the results, we see FCF yield has a relatively smooth upward trajectory with an excellent information ratio (IR)1 of 1.0. In contrast, book yield only generates an IR of 0.3.
“Despite value’s current attractiveness, there are headwinds against a value renaissance”
Chart 3: Free cash flow yield outperforms book yieldSource: FactSet & Aberdeen Standard Investments (as of February 2019)
There is a very intuitive rationale behind these FCF yield results. High FCF yield stocks often have stable revenues, earnings and dividends and hence perform relatively well during bear markets. Furthermore, the FCF yield still has price in the denominator. This means the factor is biased towards cheaper stocks which tend to do well when economic fundamentals are improving.
In addition, further analysis shows FCF yield outperforming during Federal Reserve interest-rate hikes and it appears the factor is not overly sensitive to changes in volatility or macroeconomic shocks.
No single factor outperforms all the time, in all market conditions, but FCF yield has generated superior risk-adjusted returns over the past 25 plus years on a global universe. Furthermore, the factor has a convincing investment rationale which insures against data mining.
Boosting the outperformance of value by excluding value traps
Any basket of value stocks will inevitably contain a number of so-called value traps. These are cheap stocks (according to key valuation metrics), that continue to underperform due to structural problems in the company’s business. An example would be photographic film manufacturer Eastman Kodak in the decade after 2000 when consumers switched en masse to digital cameras. The stock was cheap, but the share price continued to decline due to the company’s increasingly obsolete products.
This observation then begs the question - is there any way to identify and remove value traps from a value universe in a rules-based, systematic way, so as to increase the outperformance of value? We believe that this is indeed possible using quantitative techniques.
A seminal paper published in 2000 by the accounting professor Joseph Piotroski, then at the University of Chicago, used historical financial statement data to identify value traps based on deteriorating fundamentals such as decreasing liquidity, decreasing profitability and increasing leverage.
The original paper was based on an analysis of US stocks, but the underlying methodology can be readily applied to other universes. Backtest results on a global universe are shown in Chart 4.
Value traps can be thought of as ‘low quality’ stocks and so the removal of value traps from a value universe is another way of shifting value to the quality end of the value continuum described earlier.
“Although Value is currently cheap, it may be the wrong time in the economic cycle for value to outperform"
Although Value is currently cheap relative to history, it may be the wrong time in the economic cycle for value to start to outperform. Furthermore, there are increasing questions in the quant community over the efficacy of book yield as a measure of value.
The previous comments notwithstanding, we still believe in value thesis as applied to book yield, but investors should be willing to take a multi-year view and be prepared to withstand periodic drawdowns in performance. As a consequence, investors might want to consider alternative ways of gaining value exposure. This could be achieved by moving across the value continuum from cyclical value towards ‘value-quality’ through targeting free cash flow yield or the elimination of value traps from value universes.
Chart 4: Excluding value traps boosts the performance valueCumulative strategy performance, rebased to 100, as of December 1989:
Source: FactSet & Aberdeen Standard Investments (as of February 2019)
1The Information Ratio (IR) is a measure of the risk adjusted outperformance. It is defined as the average outperformance per period (months in this case) of the long portfolio relative to the short portfolio, divided by the standard deviation of these returns, all annualised (which here means multiplying by √12). A rule of thumb is that an IR of over 0.5 is “good.”
Value can underperform, or move sideways with volatility