Value Stocks That Have Been Working
It’s no secret that Value has been cold lately. Although many such stocks have moved off their lows, Value ETFs are, on average, still down 3.68 % in the past year. This does not, however, mean it’s now a bad idea to choose stocks whose prices are reasonable relative to fundamentals. It does, however, mean we have to be especially careful about the fundamentals.
The Essence of Value
Value is not, never was and can never be just about low ratios. Anybody who sorts on the basis of P/E, Price/Sales, Price/Cash Flow etc. and simply buys stocks at the low end of the spectrum is asking for trouble. They may not always get smacked – Lady Luck can accommodate bad ideas for a long time periods. But when the going gets tough, Lady Luck tends to get going; away, far away, from our portfolios.
Stock pricing starts with the academic Dividend Discount Model (DDM), which equates the ideal fair Price to Dividend divided by the difference between Required Return and Growth. In other words:
- P = D/(R-G)
(This formulation copes with an unknown holding period but is based on the idea that a stock is worth the present value of the cash flows we expect to receive as a result of ownership; i.e. on the irrefutable logic that no rational person would spend $100 to receive $5.)
This is no secret sauce. Everybody who takes an entry-level investment-finance course learns this very early on. Then, they forget about it (right after taking the exam) because it’s impossible to strictly implement it in the real world. Many stocks don’t pay dividends. G has to be a very low number (to accommodate a theoretically infinite time horizon which is just as well because if G is greater than R, than P will be negative and obviously, that can’t happen).
The key to making this work lies in the words of Carveth Read, to the effect that it’s better to be “vaguely right than precisely wrong” (many erroneously attribute this adage to Keynes). Instead of dismissing the DDM because the only thing we can do with precision is screw it up, we should practice the art of developing vaguely correct approaches to value flavored with an understanding that well conceived vagueness can be more than enough to allow us to achieve investment success.
I start doing this by substituting earnings per share (E) for dividends and doing some algebraic reshuffling and coming up with this:
- P/E = 1/(R-G)
That’s no more usable as a precise formula than is the original DDM. But now, we see a roadmap to successful value strategies. P/E does not necessarily have to be lower-is-better, and low P/E, by itself, isn’t good enough. P/E does, however, have to be reasonable in light of 1/(R-G). While it’s still impossible to plug specific numbers into a spreadsheet, we can recognize the relationships and articulate three things we need to consider in the real world, especially when we recognize that R is based on the risk-free interest rate and risk:
- As G rises, ideal P/E goes up. You’ve seen that before – it’s the PEG ratio. But PEG is incomplete. We also have two other things to think about.
- As Risk goes up (pulling R up with it), P/E goes down and vice versa. That seems counter-intuitive at first glance: Many think of value as a conservative strategy (margin of safety, avoidance of chasing flying stocks, etc.). Not so! All else being equal, you need to pay up (in terms of P/E) for better-quality less-risky stocks.
- As interest rates go up (thus pulling R up), P/E goes down and vice versa. You’ve heard this many times in the financial media. Now, you can see exactly how and why it impacts P/E.
Why Value Has Been Struggling
Lately, we’ve been living with the reality of rising rates (just a teeny weeny bit and for a brief period) and the fear of it down the rad (a huge concern) and at the very least, the absence of the falling rates we’d been enjoying for 35 years. So we need, under a best-case scenario (rates stay where they are) to wean ourselves away from our multi-decade habit of expecting P/E ratios to rise across the board just for the heck of it. That’s done. It’s over. As they say in old Brooklyn, “Fuggedaboutit!”
But if that’s all it was, all aspects of the market would have been comparably cold as all P/E ratios compressed across the board. That would allow Value to still shine, albeit outperforming a weak market is less fun than outperforming a strong market. But that’s not what we’ve seen. Value has underperformed a lackluster market. So there has to be something going on above and beyond the impact of interest rates on P/E.
The missing link – the reason why Value has been even weaker than can be explained by the system-wide narrowing of P/E ratios – is self-inflicted. It’s the tendency among many in the investment community to choose low P/E stocks without regard to the risk component of R and G and label it “Value.” It’s what quants do when they use Fama-French style factor models. It’s what Value ETFs do because, well, because the index creators, inspired by Fama-French type protocols, haven’t figured out that you can’t use the scientific method (empirical factor analysis) without also addressing financial/economic realities (the need for considerations not just of P and e but also R and G).
What If . . .
What if we tried to get our hands on low valuation-ratio stocks that also give us reasonable quality and decent growth prospects?
That’s pretty much what my Cherrypicking the Blue Chips model (available for free on Portfolio123’s Smart Alpha platform and which was explained here on Forbes.com) winds up doing. On a zero to 100 scale, it scores 81 based on our Smart Alpha Value style rank, which means it does a pretty good job picking up stocks with low ratios. Among my models, that is by far the highest Value-style score.
Based on this alone, I should expect it to have had a rotten year. Actually, though, it’s been pretty good. That’s because the model doesn’t rest only low ratios. Specifically:
- I don’t actually calculate R (there are too many unknowable considerations to make it possible to do this an anything other than a classroom setting) nor do I need to. I’m satisfied to simply require that the stock be something better than a corporate dumpster fire. I do this through a requirement that the stock pass a certain threshold under a multifactor ranking system that includes a strong Quality component.
- I absolutely positively don’t calculate a number for G because I can’t deal with infinity and because I can’t actually predict much of anything. So I look for secondary clues that suggest a stock is more likely than not being seen as having better-than-decent growth prospects. I do that through the historic Growth and Momentum components of the aforementioned multi-factor ranking system and I double down with a requirement that the stock rate well under a Sentiment model that works with estimate revision, earnings surprise, and analyst rating. Now you and I know that none of this assures good growth prospects in the future. But, but, but, if you’ve been paying attention to the markets, the financial media, etc., analysts and the Street tend to get overly carried away with the idea that good news in the past will persist forever. Nonsensical as that really is, it’s more than ample to allow me assume the stock will behave for a little while based on investment-community expectations consistent with a favorable G number. Is this a cheap trick to be playing on Mr. Market? Heck yes. But knowing as we do that past performance doesn’t assure future outcomes, we may as well work with what we can and if a dirty trick will do the job, then so be it. It’s not like I’m marrying these stocks. If the Sentiment rank dips below a certain threshold, I’ll boot it so fast . . . .
So I don’t specifically calculate 1/(R-G), which is just as well because it can’t be done. But I can and do have model that identifies low valuation stocks that are more likely than not to have objective characteristics consistent with the ideal of a decent 1/(R-G) formulation.
It’s Been Working
Table 1 shows the returns achieved over the past 12 months by my Cherrypicking portfolio (this is live money, not a backtest); an equally-weighted Large-Cap Value benchmark I use (with only 10 more-or-less equally weighted S&P 500 constituent stocks in this portfolio, this seems sensible); and a portfolio of all large-cap value ETFs.
Table 1
12-Mo. % Return | |
Portfolio | +5.96 |
LargeCap Value Benchmark | -12.23 |
LargeCap Vale ETFs | -5.92 |
Table 2 compares the portfolio holdings and the S&P 500 constituents in terms of standard value ratios.
Table 2 – Valuation Ratios
Averages | Medians | |||
Portfolio | S&P 500 | Portfolio | S&P 500 | |
PE Estimated Cur Yr. EPS | 11.25 | 21.73 | 11.24 | 17.97 |
PE Estimated Next Yr. EPS | 10.44 | 24.02 | 10.29 | 16.31 |
PEG Ratio | 1.27 | 2.63 | 1.43 | 1.74 |
EV/Sales | 1.96 | 4.33 | 1.43 | 2.87 |
Price/Free Cash Flow | 10.92 | 44.38 | 8.80 | 26.96 |
Price/Book | 1.62 | 6.62 | 1.65 | 2.98 |
Yeah, those are low ratios. If the 1/(R-G) bundle is as bad as the ratios imply, this portfolio should really stink.
Tables 3 and 4 give us a feel for how to portfolio fares in terms of 1/(R-G). Table 3 compares the portfolio holdings and the S&P 500 constituents in terms of company Quality and Financial Strength measures, which I’m using as a general proxy for the risk component of R. Table 4 does likewise for items that can be regarded as a sly proxy for G.
As you look at these tables, recall how low the valuation ratios (Table 2) are relative to the S&P 500. Any R-and-G profile that’s better than the S&P 500 could be considered an excellent information arbitrage value-type opportunity. Profiles that are about the same are very good opportunities. Profiles that are less than the S&P 500 are still OK as long as they are not basket cases (remember how low those value metrics are).
Table 3 – Items Relating to R (Quality/Risk)
Averages | Medians | |||
Portfolio | S&P 500 | Portfolio | S&P 500 | |
Return on Equity – 1 Yr | 14.83 | 34.74 | 13.15 | 14.16 |
Return on Equity – 5 Yr | 18.10 | 18.56 | 12.75 | 7.61 |
Operating Margin – 1 Yr | 20.51 | 14.85 | 12.89 | 16.88 |
Operating Margin – 5 Yr | 21.07 | 19.49 | 13.17 | 16.95 |
Asset Turnover – 1 Yr | 0.49 | 0.72 | 0.39 | 055 |
Asset Turnover – 5 Yr | 0.44 | 0.68 | 0.30 | 0.52 |
Debt to Equity | 1.10 | 1.77 | 0.58 | 0.82 |
Interest Coverage | 8.92 | 19.58 | 8.96 | 7.99 |
Table 4 – Items Relating to G (Growth)
Averages | Medians | |||
Portfolio | S&P 500 | Portfolio | S&P 500 | |
Sales Growth 5 Yr. | 4.56 | 6.77 | 3.55 | 5.24 |
Sales Growth 12 mo. | 1.19 | 1.17 | 1.32 | 0.70 |
Sales growth last qtr. | 1.55 | 2.51 | 3.24 | 1.71 |
EPS Growth 5 Yr. | 23.23 | 12.77 | 8.15 | 10.15 |
EPS Growth 12 mo. | 41.27 | -199.22 | 11.05 | 0.95 |
EPS growth last qtr. | 33.76 | -22.00 | 11.46 | 0.00 |
Sales Volatility | 0.11 | 0.14 | 0.08 | 0.10 |
Oper. Inc. Volatility | 0.24 | 0.36 | 0.19 | 0.16 |
EPS Volatility | 0.36 | 0.58 | 0.38 | 0.25 |
Proj. LT EPS Growth | 9.68 | 9.89 | 9.25 | 9.93 |
Last Qtr. EPS Surprise | 10.37 | 4.59 | 6.85 | 3.90 |
Given the very low valuation profile, combined with fundamental profiles that range from tolerably less than the S&P 500 (with none being basket cases) to significantly better, it’s easy to see why these value stocks came through even during a period when Value was widely considered to have been cold.
The Stocks
Here are the stocks in the portfolio as of this writing. It updates every Monday morning and often there are one or two trades (none this past Monday). For an up-to-date look, go to the model presentation on Portfolio123.
Table 5
Ticker | Company |
AFL | AFLAC |
DFS | Discover Financial Services |
ETN | Eaton Corp |
F | Ford Motor |
LRCX | Lam Research |
PVH | PVH |
RF | Regions Financial |
STI | SunTrust Banks |
TSN | Tyson Foods |
UNM | Unum Group |
Disclosure: None.