Magic Formula has a magic 12 months

The magic formula itself is a super-slick system that gets straight to the nub of what tends to drive stock returns: how effective companies are at producing earnings and what investors are required to pay for those earnings. True, the simplicity of the methodology means it is far from fail-safe and in the six years I've been running the screen a number of companies have been included in the selection that have proved to have been reporting bogus numbers, such as Globo. Other companies simply turn out to be 'cheap for a reason', meaning future earnings and productivity drop along with the share price. Game Digital is a good example of such a stock from last year's screen.

The magic formula system does not need to be exacting in how it identifies stocks. Indeed, it takes the view that on the whole the criteria used by the screen should produce enough genuinely interesting situations to offset the inevitable disappointments. On that basis Mr Greenblatt's recommendation is that the 30 top ranking magic-formula stocks be used to construct a portfolio that is reshuffled once a year. He also suggests investors could use a 20-stock portfolio if 30 stocks seems too onerous. In the six years I've tested the screen, I've also looked at a super-concentrated - and therefore more risky - 10-stock version of the screen. This 10-stock screen has actually done very well, but has tended to suffer most when the strategy has underperformed.

Back on form

The past 12 months have seen the screen produce very impressive returns following two less impressive years. The result is particularly strong as the wider market has been powered by the resurgence of resources stocks from their nadir this time last year. This type of company was not a feature of the 2016 magic-formula selection.

The 2016 performance table below is ordered by total return and the final column provides the position the stocks ranked in last year. The return from Melrose is worth highlighting as it illustrates what I feel is a rather generous interpretation of total return used by Thomson Datastream (the source I use for performance data), which is that all income (dividend or other forms of cash return) from a stock is seamlessly reinvested into shares when it is paid. There was a big cash return from Melrose during the period followed by a strong share price rise, hence the big number.

*Taken over

Source: Thomson Datastream

The long-term performance of the screen is also strong, as can be seen in the chart below. Given that this is a screen designed to be used without recourse to further research; the impact of notional charges on performance seems particularly relevant. The impact of a 1.5 per cent annual charge is given in the returns table below. Over six years this all adds up with, for instance, the return from the best-performing 10-stock portfolio dropping by 15 per cent from 186 per cent without costs to 161 per cent with costs. Importantly, even with costs that's substantially better than the 59 per cent from the FTSE All-Share in the same period.

Greenblatt performance

Source: Thomson Datastream

Combining value and quality

The valuation and returns ratios Mr Greenblatt uses for the magic formula are similar to the price/earnings ratio and return on assets (RoA), which measures profit as a percentage of total assets. However, Mr Greenblatt amends these formulas to take a 'whole company' view of valuation and to focus in on productive assets. The approach of looking at investments as a buyer of the whole company is one endorsed by many investment greats besides Mr Greenblatt himself, including the sage of Omaha, Warren Buffett. The point of putting oneself in the shoes of a company-buyer is based on the recognition that as well as being funded through equity (the total value of which represents a company's market capitalisation), a company has other sources of financing that often rank above its shares. The most notable of these is debt.

To take a 'whole company' view on valuation, Mr Greenblatt uses a PE ratio (in the magic formula he expresses it as an earnings yield but it amounts to the same thing) that is calculated using enterprise value (EV). In its most simple form EV adds debt to a company's market cap and takes away cash. EV can also be calculated to account for debt-like items such as pension deficits and long-term operating lease liabilities (eg long-term commitments to pay rent on a leased property). The earnings figure in Mr Greenblatt's PE is also adjusted to take out interest paid on debt or received from cash held and is given before tax (Ebit).

Mr Greenblatt looks at profits generated from 'tangible assets' as a way of measuring quality. The measure again looks at Ebit to get a clearer picture of the profits being generated from the company's actual operations. The tangible asset measure, which consists of net working capital added to net fixed assets, also has a focus on the assets that are actually being used in the company's operations in order to create its profits. Due to the nature of the balance sheets of financial companies, such stocks are excluded from the screen.

How the screen works

Mr Greenblatt's magic formula uses a ranking system to play off the quality and value aspects of his approach and come up with a list of the most attractive stocks. I'm personally a big fan of the simplicity of this ranking system and have adapted it to use with different metrics for a number of other screens I regularly run. The ranking system works by first ranking stocks individually for both the quality and value facture, then the two rankings are combined to produce a final ranking which the screen results are based on.

I've used Mr Greenblatt's magic formula to screen every non-financial company on the main market and Alternative Investment Market (Aim) with a market capitalisation of more than £50m. These are the results:

2017 magic-formula stocks

Source: S&P CapitalIQ