First, we observe that the current valuation dispersion in these factors (e.g., the discount of value stocks relative to their expensive peers, and the premium attached to high-momentum stocks) is much greater than historical averages, creating a potentially heightened probability for mean reversion.  Next, we find that the current spreads in valuation multiples are not explained by spreads in earnings growth and returns on equity (ROE) expectations - the key drivers of price-to-earnings (P/E) and price-to-book value (P/B) ratios.  Finally, by examining past drawdowns in value and momentum factors, the analysis indicates that when the mean reversion comes, the “snap-back” will likely happen quickly.

Value investing has always been an inherently unpopular and lonely road to travel.  Cheap stocks tend to trade at discounts to peers for a reason, though the academic debate continues as to whether that reason is more risk-based or behavioral in nature.  Risk-based proponents argue that value stocks price in a higher probability of financial distress.  More generally, those stocks may reflect greater perceived uncertainly around future earnings created from cyclical, structural, or competitive forces.  Behavioral explanations focus instead on the practical limitations facing market participants.  Many investors simply do not have the mandate, liquidity, patience, or conviction to maintain these unpopular positions for extended periods of time.  Value stocks can take years to “re-rate” upward, and holding these out-of-favor stocks will likely cause short-term pain.  As with any consistent investing strategy, a value style will underperform the broader market from time to time.  This is simply the price of being contrarian.

So why invest in value at all?  Its track record is perhaps value’s strongest advocate.  Over longer periods, value stocks have outperformed the markets fairly consistently.  As Exhibit 1 illustrates, despite short periods of weakness, value indices have outperformed benchmark indices over the past 25 years.  Additionally, buying inexpensive or “cheap,” unloved stocks makes intuitive sense.  If a stock trades lower but fundamentals remain unchanged, the upside to downside ratio improves and the increasingly asymmetric return profile becomes more attractive.  At Causeway, we like to add to positions on these price declines.  It may be difficult to do at the time, but this discipline is why you hire an active value manager.  Although timing cannot be predicted, eventually value works, the pendulum swings the other way, and cheap stocks ultimately outperform.  Contrarian investors who can “stomach” value drawdowns are rewarded for their patience by the subsequent snap-back and re-rating.


Despite a small bounce in March 2016, however, value stocks have underperformed broader market indices since the summer of 2014.  The questions that follow are, how long can this underperformance persist, and at what point might value resume its historical leadership?   While identifying the precise turning point is not possible, we can search for indications by examining past valuation spreads, disconnects between market multiples and their respective drivers, and the structure of previous drawdowns.  Though our primary style factor focus is value, we also analyze momentum, cyclicality, and volatility*  given recent trends.

*“Value” measures a stock’s relative cheapness, “momentum” measures a stock’s relative price performance, “cyclicality” measures a stock’s sensitivity to economic cycles, and “volatility” measures a stock’s historical variability.

I. Dispersion in the Valuation Ratios of Style Factors is Elevated Relative to History
We begin by measuring how cheaply or expensively each of these four styles is trading relative to the last 25 years.  We use 25 years based on the availability of constituent-level data for the MSCI World Index. At the end of each month, we distribute the constituents of the MSCI World Index into quintiles based on each stock’s relative exposure to a specific style.  For each factor, we use Causeway’s risk model definitions which combine multiple measures of each style exposure.   Next, we observe the median price-to-book value and forward price-to-earnings multiple for those stocks sorted into the extreme quintiles (the top quintile, or “Q1” and the bottom quintile, or “Q5”).  Finally, we track the ratio of the median valuation multiple of Q1 relative to the median valuation multiple of Q5.  Comparing the most recent measure of this dispersion ratio to its range over the past 25 years will give us a sense of how far away the current ratio is from norms over the period.

Exhibit 2 plots the most recent data point for each style factor standardized relative to its 25-year history.  Current points are plotted in terms of standard deviations from the 25-year median (to diminish the impact of historical outliers).  If a particular factor’s dispersion is either well above or well below its 25-year norm, we would argue that it is overvalued or undervalued, respectively. The gray bars reference a +/- 1 standard deviation difference from the median of the respective valuation ratio between extreme quintiles.


We see from Exhibit 2 that the valuation ratio of cheap stocks relative to expensive stocks is roughly a 1 standard deviation discount to 25-year historical norms on a P/E basis and even greater on a P/B basis.  Stated differently, cheap stocks usually do not trade so cheaply, and value as a factor appears oversold.   Momentum exhibits the opposite characteristic.  High momentum stocks are trading at a much larger premium to low momentum stocks relative to their 25-year history.   Additionally, we observe that cyclical stocks are trading at a larger discount to less cyclical (or “defensive”) stocks compared to their 25-year history, and higher-volatility stocks are also trading at a greater discount to lower-volatility stocks.   These findings are consistent with recent relative performance: Value has underperformed, while momentum, defensive, and “low-vol” stocks all have outperformed.

Why do we care about the current valuation dispersion of each style factor relative to history?  All of these factors have a tendency toward mean reversion.  And we believe Causeway value equity strategies are well-positioned for such a turn.  In our international and global value portfolios, we currently have positive active exposure (exposure relative to that of the benchmark index) to value and cyclicality, and we have negative active exposure to momentum.  Mean reversion in these factors should benefit Causeway portfolios.  It is important to emphasize that these active positions do not represent attempts to “time” factors.  Rather, as value investors, we will generally have positive active exposure to undervalued factors and negative active exposure to overvalued factors.

II. Spreads in Fundamental Drivers Do Not Explain Differences in Current Valuation Multiples
Another way to gauge the degree of disconnect in factor valuations is to investigate the drivers behind the specific metrics that we have been using: price-to-earnings and price-to-book value multiples.  Using the Gordon Growth Model framework, we know that a stock price (P) should reflect the growing stream of future earnings per share (Et+1 growing by g), assuming all earnings are paid as dividends, discounted by the cost of equity (re):


If the market is efficient, a stock’s price should adjust to equal future earnings per share (EPS) divided by the difference between the cost of equity and the EPS growth rate.  Rearranging the terms, the P/E multiple should equal the reciprocal of the difference between the cost of equity and the growth rate.  As the equation indicates, investors should be willing to pay a higher P/E multiple for stocks exhibiting higher earnings growth.  Although the cost of equity is admittedly an important determinant of a P/E multiple, we find that the relationship between market beta (a key input in CAPM-derived re) and valuation quintile is not stable historically.  Expected earnings growth (g), on the other hand, has consistently been positively correlated with P/E, and we will therefore focus on growth differentials in this analysis.

First « 1 2 3 » Next