Screening for Value: An Earnings-Based Screen

Over the past few weeks, I have generated value investing ideas using the Stockopedia screening tools. This week that quest continues with an earnings-based screen.

Most of the mainstream academic finance research has focused on the Price-to-Book metric, which was the focus of my last screen. However, it has been known for some time that using earnings-based metrics may overcome some of the issues that Price-to-Book faces as a measure of value. Most of the published data has come from “quant practitioners,” who use their research to manage money via quantitative strategies. While I am not aiming to generate a purely quantitative screen, understanding the history of this research helps ensure I pick the right screening factors.

Earnings-Based Value Metrics – A Brief History

Perhaps the best-known of the early quant practitioners are David Dremen and James O’Shaughnessy. In 1980, Dremen wrote a book called Contrarian Investment Strategy, which highlighted that a low Price-to-Earnings strategy was the best-documented way to beat the market. In 1998, he followed this up with a book called Contrariarian Investment Strategies: Beating the Market by Going Against the Crowd. He looked at more up-to-date data for Price/Earnings, Price/Cash Flow, Price/Book Value, and Dividend Yield, finding Price/Earnings to still come out on top.

O’Shaughnessy’s book What Works on Wall Street was first published in 1997 and is now in its 4th Edition. O’Shaughnessy conducted a more thorough analysis and included the Sharpe ratio in his findings. The Sharpe ratio takes the excess return that a strategy generates and divides it by the volatility of those returns. The idea is that a higher return doesn’t gain investors anything if it comes at the cost of higher volatility. If the Sharpe ratio of a strategy was the same as the index, investors could simply leverage up the index to get the same result with less effort. Although the idea isn’t without its issues – leveraging the index is not cost-free, and investors don’t mind upside volatility – it makes sense to conclude that a strategy only outperforms another if it has a higher Sharpe ratio.

What Works on Wall Street contains some interesting findings. The first was that in the period studied (1951-2003 in my Third Edition of the book), Price-to-Book-Value and Price-Sales came out with the highest Sharpe ratios. The second was that the outperformance was concentrated in the largest stocks, which goes against the findings of many other studies of value anomalies.

However, I am not ready to write off earnings-based value metrics just yet. The story continues with one of the biggest pivots in investing. Joel Greenblatt made a name for himself as a Special Situations value investor. His book You Can be a Stock Market Genius describes how investors can make great returns by focusing on situations where other investors are selling for reasons unrelated to the value of the business. The classic example of this is so-called “spin-offs”, where a smaller subsidiary of a larger business is separately listed, and existing investors get the spun-off stock. Investors tend to do what we would all do in this situation, sell the small unknown shares as soon as we receive them and treat them like a bonus. Greenblatt made over 50% compound for 19 years by buying the most promising of these small spin-offs after most investors had dumped them.

This type of return is phenomenal, but the problem is that a limited amount of cash can be deployed into this strategy. So Greenblatt searched for more scalable strategies and found one which he called the Magic Formula. He aimed to implement a quantitative translation of Warren Buffett’s observation that “it is better to buy a wonderful business at a fair price than a wonderful company at a fair price”.

The Magic Formula has two components. The first is intended to find wonderful companies. Greenblatt ranks all companies on their Return on Capital. He uses a specific definition which Stockopedia has in the database as Return on Capital (Greenblatt):

Return on Capital (Greenblatt) = EBIT / (Net fixed assets + working capital)

To find companies trading at a “fair price”, Greenblatt ranks companies by EV/EBIT. EV is the Enterprise Value, the sum of the market capitalisation and net debt. EBIT is Earnings Before Interest and Tax, often called Operating Profit. The Magic Formula adds together these ranks and takes the lowest, representing high RoC and low EV/EBIT. Greenblatt claimed that owning the best 30 ranked stocks on the Magic Formula would have returned a 30.8% compound from 1998-2004.

Although the Magic Formula was based on logical concepts and backed up by data, the astute will have noticed some issues. The first is that applying logical theories to investing doesn’t always work in the real world. The second is that Greenblatt’s chosen period to test the strategy (1998-2004) is relatively short. Investing is replete with data mining issues, where market anomalies are a quirk of the data and don’t persist beyond the initial test period.

Enter Wes Gray and Toby Carlisle, who set out to test the Magic Formula empirically. They go through their findings in their book, Quantitative Value. The first thing that they find is that they can’t replicate Greenblatt’s phenomenal results. However, they did find that the Magic Formula performed well. The top 10% of stocks, as ranked by the Magic Formula, delivered a 12.8% compound return from 1964 to 2011. In comparison, the S&P500 Total Return Index delivered a 9.5% compound return during the same period. The most interesting finding, however, was that Greenblatt’s chosen quality metric appeared to detract from the strategy’s overall performance. He may have been better off going for cheapness alone!

What is the best value metric?

While this sojourn into the history of earnings-based metrics is hopefully interesting, what does it have to do with screening for value? Well, first, it explains why I intentionally don’t include any return on capital metrics in this particular screen. Even more usefully is that in Quantitative Value the authors examine the performance of different value metrics. They chose to analyse a pretty comprehensive list: Market Cap/EBIT, EV/EBIT, EV/EBITDA (which also excludes depreciation and amortisation), EV/Free Cash Flow, EV/Gross Profits, and Market Cap/Book Value. The results of this analysis can be seen in the table below:

[Note that Gray & Carlisle use Earnings Yield = Market Cap/EBIT, whereas Stockopedia uses my preferred definition of Earnings Yield = EV/EBIT.]

While all ratios outperform the index, EV/EBIT (Earnings Yield in Stockopedia language) has the highest Sharpe ratio and lowest drawdown in this study. It appears that Joel Greenblatt certainly got this bit of his Magic Formula right. It makes sense to choose this for my earnings-based screen too.

The Screen

I choose to use Trailing Twelve Month data, so I have the most up-to-date figures. And mirroring the way that the quant studies work, I am using an Advanced Rule to screen for the highest 15% of TTM Earnings Yields on the UK market. (I have included a bit of leeway vs the usual top decile of the quant studies since I will be applying other screening criteria). The way to do this in the Stockopedia screening tool is to use the Rank in Market, in this case, greater than 85.

In addition, I add in some criteria that will now be familiar:

Market Cap Range £10m-£600m I keep this the same as the previous screens, as I tend to find that this range gives companies that are investable but not so diverse in their operations as to make them too hard to analyse. This will exclude large cap companies such as BHP (LON:BHP) or £SHEL. These may be good companies to own to gain exposure to the mining cycle or oil prices. However, I am unlikely to be able to get beyond a surface-level analysis for companies with this scale of operations.

Current Ratio > 1.2 For this earnings-based screen, I have relaxed the current ratio required to 1.2, compared to the 1.5 that I used in the previous screens. Since this is a screen for companies with high earnings compared to their market price, these are likely to be more securely financed. While I will still be mindful of balance sheet strength in my initial analysis, it makes sense to be less cautious in the initial screen.

Excluded Industry Segments: Banking Services, Investment Banking & Investment Services, Residential & Commercial REITs, and Collective Investments. Again, I exclude these to focus on trading businesses. It is notable that a number of investment banks would appear on this screen if they weren’t excluded and with very high TTM Earnings Yields. This can often be a function of the regulatory capital they hold and may not represent free cash. Given this, I intend to look at this sector in more detail in a future article.

As of writing, this screen generates 52 results.

Those who have been reading my previous Screening for Value articles will notice that many of these companies have appeared on my previous screens. This is to be expected since unloved stocks can often appear cheap on many different metrics.

Of these stocks, 10 have already made it onto my watchlist: Angling Direct (LON:ANG) Atalaya Mining (LON:ATYM) Capital (LON:CAPD) Kenmare Resources (LON:KMR) Logistics Development (LON:LDG) National World (LON:NWOR) J Smart & Co (Contractors) (LON:SMJ) Sylvania Platinum (LON:SLP) Wentworth Resources (LON:WEN) Wynnstay Properties (LON:WSP)

And, 13 stocks I have already excluded for various reasons: Altyngold (LON:ALTN) Carclo (LON:CAR) Conygar Investment Co (LON:CIC) Enwell Energy (LON:ENW) Frontier IP (LON:FIPP) Gem Diamonds (LON:GEMD) £MPHC OPG Power Ventures (LON:OPG) Petra Diamonds (LON:PDL) Pharos Energy (LON:PHAR) Rockwood Strategic (LON:RKW) Serabi Gold (LON:SRB) Tandem (LON:TND)

This leaves 29 stocks to look into further:

Next week’s article will follow a familiar path. I will start by eliminating any companies with obvious issues and then look into the remaining ideas in some detail.

[Disclosure: Mark owns shares in Capital (LON:CAPD) Goldplat (LON:GDP) Luceco (LON:LUCE) National World (LON:NWOR) Sanderson Design (LON:SDG) £ZYT]