Cheap recovery plays

While Mr Piotroski scores big on the method he set out for measuring quality (the aforementioned F-Score), it can be argued that his take on valuation was less inspired. Indeed, when it came to assessing value, his paper simply opted to use an old-school staple: the price/book value ratio (PBV).

The popularity of PBV as a way to assess value can be traced back to Benjamin 'father of value investing' Graham. The ratio gained another blue-blood stamp of approval in the 1990s thanks to its inclusion in the seminal three-factor model set out by Eugene Fama and Kenneth French, which spawned much of today’s enthusiasm for 'factor' and 'smart beta' investing.

But when financial measures become very popular, they often lose their effectiveness. Many argue that this has been the case with PBV in recent decades given the failure of shares that look attractive based on the ratio to demonstrate meaningful outperformance of the wider market. What’s more, times change and there are grounds to argue that the most important investments made by companies in their operations in the modern age are ignored by PBV. This is because when companies invest in things such as intellectual property, technology and brand, the spending is often treated as an expense in a company’s profit and loss account rather than being capitalised to create a book-value-boosting asset. By failing to take full account of such important, albeit intangible, assets, book value may be falling short in representing the earning potential of some of today’s most promising companies. The measure also has a bias towards highlighting capital-intensive businesses as opposed to the capital-light business that are so adored by modern investors.

The purpose of this week’s screen is to take Piotroski’s insightful quality metric (the F-Score) and use it in conjunction with a range of valuation measures. Given that Piotroski’s F-Score principally involves looking for signs of an underlying improvement in fortunes, it is best used when hunting for 'value' situations; the hope being that it can spot unloved stocks where prospects are improving before the wider market cottons on.

This year I am adding one more value metric to the five that the screen has made use of in previous years. The metric in question is one I have named the ZEUS ratio (I’m sure it’s not a very original idea but I’m yet to be made aware of others using it so have given it my own name). I think the ratio fits quite well with the ideas that make the concept of PBV of interest.

The idea behind targeting stocks that look cheap compared with book value is that book value gives a measure of a company’s ability to generate profit. In the case of a value stock, the notion would be that returns on a company’s assets had dropped due to some kind of rectifiable issue – operational or economic. In this case, P/BV gives an indication of recovery potential at a time when returns (and therefore profits) are historically low or negative. Essentially, P/BV attempts to look at the valuation being placed on the source of a company’s earnings to try to assess the upside should things improve.

While some investors believe that all returns and valuations eventually revert to a common mean average over time, the ZEUS ratio takes the view that different companies and industries retain different characteristics over lengthy periods. This means different companies and industries are likely to have different average valuations over the course of an economic cycle. For example, over a cycle the valuation of a capital-light, high-margin, non-cyclical business is likely to be much higher than that of a capital-intensive, low-margin, cyclical business. What’s more, for the capital-light businesses, turnover rather than book value is arguably a better way to assess the source of earnings, especially in light of some of the issues with P/BV in the modern age discussed earlier.

The ZEUS ratio therefore simply assesses where a company’s valuation against its source of earnings (usually sales but alternatively book value depending on the industry) sits compared with its historical range. Ideally, the comparison is made over 12 years, which should be plenty enough to capture a full cycle, but a minimum comparison period of four years is allowed. The comparison is made using a statistical measure known as a Z-Score. While it is probably only of interest to statistics nerds that the Z-Score represents a current value’s standard deviations from the mean, for everyone else the interesting thing is that a negative score means a stock is cheap compared with the historical average. A score at –1 or below suggests it is very cheap (roughly within the bottom third of the cheap zone) and at –2 it is screamingly, and probably suspiciously, cheap against history (about the cheapest 5 per cent).

Like pretty much all valuation measures, the ZEUS ratio (a shamelessly tortured acronym standing for Z-score of Earnings Ulitimate Source) is best regarded as rough and ready. Over the course of a cycle companies and industries can change in character substantially, causing ratings move to higher or lower long-term average levels. Indeed, it is hard to argue there is any stable level for long-term valuation averages.

To date, the F-Score Value Hunt screen, which tends to highlight many companies on each outing (40 this time around), has put in a decent performance overall. The past 12 months saw a 13 per cent total return based on an equal-weighted average of the performance of each F-Score portfolio (All-Share, All Small and Aim). The average total return from corresponding indices was 8 per cent.

2017 PERFORMANCE

Source: Thomson Datastream

Since I began running this screen four years ago, the cumulative total return stands at 68 per cent, or 60 per cent if I factor in annual costs of 1.5 per cent to represent the notional expense of reshuffling the portfolios (the assumption of this column is that screens are primarily useful to readers for idea generation rather than as off-the-shelf portfolios). The average total return of the indices over the same period is 44 per cent.

The Piotroski F-Score itself attempts to judge whether a company has been improving profitability and cash generation through operational improvement. It does this by assessing improvement of the profit-and-loss and cash-flow accounts counterbalanced with an assessment of the balance sheet that makes sure companies have not had outside financing (share issues or borrowing). Mr Piotroski did come off with a rather wonderful interplay of factors, which are as follows:

■ Positive profit after tax, excluding exceptional items.

■ Positive cash from operations.

■ Profits after tax excluding exceptional items up on the previous year, which Professor Piotroski highlights as being of particular importance as a signal that a company may be in recovery mode and in the process of re-rating.

■ Cash from operations higher than profit after tax, excluding exceptional items, which indicates an ability to convert accounting profit into actual cash.

■ Gearing (net debt as a percentage of net assets) down on the preceding year, which suggests that the company has not had to look for external sources of finance.

■ The current ratio (current assets divided by current liabilities) up on the preceding year, which suggests that the company's ability to service upcoming financial obligations is improving.

■ No new shares issued over the past year, which again suggests that the company has not had to look for external sources of finance.

■ Gross margins have risen in the past year.

■ Improving capital turn (turnover as a proportion of net assets), which suggest greater productivity.

Companies score one point for each criteria met and score of eight and more out of nine are considered high and of interest. As well as having a high F-Score, to pass the screen the stocks must be among the cheapest fifth in their index based on one of more of six valuation measures: P/BV, dividend yield (DY), enterprise-value-to-sales (EV/S), enterprise-value-to-free-cash-flow (EV/FCF), the aforementioned ZEUS ratio and a dividend- and cash-adjusted price/earnings growth ratio (GV, or genuine value).

The table shows the 40 companies that qualified this year ordered by market capitalisation. The valuation metric(s) against which the stocks look cheap is also given in the table and companies are ordered from highest to lowest market capitalisation. A downloadable version of the table is also available through the link below.

F-SCORE VALUE STOCKS

Source: S&P Capital IQ