Cash magic

The key argument for looking at cash rather than earnings is that it is far harder for companies to manipulate and massage the numbers. The formula for CROCI we've used in our screen is as follows:

Cash profits (earnings before interest, tax, depreciation and amortisation) / (total assets – current liabilities)

While CROCI gives an insight into the ability of a company to generate cash from its assets, it tells us little about what price a would-be investor is expected to pay for the cash that is being generated. To bring a value element into this screen we've turned to one of our favourite star-investor screens - Joel Greenblatt's magic formula. Mr Greenblatt uses two components in his very simple but highly effective screen, which measure a stock against both an earnings-based valuation and its ability to generate earnings from its tangible assets.

More specifically, Mr Greenblatt's formula ranks companies by their earnings yields (calculated as EBIT as a proportion of enterprise value (EV)) and also their return on tangible assets, which is a variant of ROA. The shares are then ranked by their combined ranking (the value ranking plus the return-on-tangible-assets ranking). Mr Greenblatt then advocates buying the 30 top ranked shares. For our cash-focused Greenblatt screen we've used CROCI in place of return on tangible assets. And instead of an earnings yield, we've used EV (market cap, minus cash, plus debt)/free cash flow (FCF) as a cash-based equivalent. After eliminating all stocks where ratios could not be calculated or were negative, we were left with about half of the All-Share for the screens we've run using this formula.

While cash is often considered to be the acid test of a business's performance, it can vary markedly from year to year. So we felt our proposed screen needed some back testing to see if it was at all credible. Fortunately, given the leg work, it seems that it is. We've tested the screen on current FTSE All-Share constituents over five 12-month periods starting from the end of the first quarter of 2008.

It needs to be pointed out that the limitation of testing only current FTSE All-Share members is that it flatters our results by including those shares that have been promoted into the index while excluding those that have fallen out. But the marked cumulative underperformance of the lowest ranked fifth of stocks over the five years tested compared with the top ranked fifth, and the incremental performance improvements of higher ranked segments (see bar charts), suggests the system holds some promise.

What's more, in each of the five years we tested, the highest ranked fifth of stocks easily outperformed the lowest ranked fifth. That said, in the most recent year tested, only the lowest ranked fifth underperformed the highest ranked fifth and in the previous year the second-to-top ranked fifth outperformed the top 20 per cent. Overall, though, the results look pretty good to us.

Perhaps it is testament to the wisdom of Mr Greenblatt that the best-performing strategy we found (see chart below) was based on a portfolio of the highest ranking 30 shares. That said, we found that the portfolios we tested showed a high level of variance, which is a measure of how far, on average, each stock varies from the overall portfolio performance. The average annual variance for the five 30-stock portfolios we tested was a whopping 27 per cent. We found a good way to limit this, and thereby limit the associated risk, was not to increase the size of the portfolio but to only select stocks with three-month momentum that was superior to that of the market prior to the portfolio selection date. The average variance on a portfolio of the top-ranked fifth of strong momentum stocks (roughly 35 stocks in most years) was 14 per cent. We actually thought the use of momentum may improve the overall performance of the portfolio, but it didn't.

We list the current highest ranked 30 stocks in our table below and below that the highest ranked fifth of stocks that have shown superior momentum to the market over the past three months.

Top 30

Momentum-Adjusted Top Fifth

Source: S&P CapitalIQ