Basantjee,

In an article published in the February 1997 edition of Professional Investor, Jim Slater shares his winning combination formula for selecting growth stocks:

I have always been attracted to growth shares, particularly those that can be purchased at what I perceive to be a discount to their proper value. I use a number of sieves to ensure that value is present as well as growth. In his recent 'Value or Growth?', Michael Lenhoff of Capel-Cure Myers referred to my style of investing as a hybrid of value and growth , or GARP - 'Growth at a reasonable price.' I prefer to describe it as buying growth shares using value filters to provide a margin of safety.

The task of devising, and helping HS Financial Publishing (Formerly Hemmington Scott LTD) to develop our monthly investment service, Company REFS, removed many fuzzy areas from my mind. All calculations had to be carefully defined and some difficult points needed to be resolved very precisely before the computer software could be written.

The first and most important question was to define growth shares and to eliminate cyclicals and recovery stocks from our growth share universe. In my view, the main characteristic of a reliable growth share is that it should be able to produce increasing EPS year after year, almost irrespective of general economic conditions. In the UK, few companies qualify.

Our aim was to catch mature growth companies, and nascent ones, before all the action was over. We also wanted to eliminate major cyclicals. After many months of trial and error, we finally settled on these parameters for defining a growth share:

• There must be broker forecasts.
• There must be four years of consecutive EPS growth, including any forecast periods. In practice, this means that a company can make the grade if it has two years of past EPS growth and two years' future growth forecast by brokers.
• Each of the past five years' results must have been profitable. None may show a loss.
• Where four periods of growth follow a previous setback, it must have achieved, or be expected to achieve, its highest normalised EPS in the latest period out of the last six.
• All property companies and investment trusts are eliminated.
• Companies in the highly cyclical Building & Construction and Building Materials & Merchants sectors, and those in the Vehicle Distribution sub-sector, are required to meet the above growth criteria. However, they are required not to have incurrred a loss, or suffered an EPS reversal, in any of the last five years of reported results. This last criterion eliminates most of the companies in those sectors, but leaves the few that are arguably genuine growth companies.

In the FTSE All-Share about 300 companies qualified, 50 in the FTSE 100, 90 in the Mid-250 and 160 in the SmallCap.

The next task was to ensure that the financial statistics of all companies were synchronised, so that they could properly be compared with each other. The term 'prospective price-earnings ratio' is bandied around too glibly. It usually refers to the current year and that is fine if the year has just begun. However, if the term is used when the year is almost over, it is clearly much less pertinent.

We resolved this problem by compiling all growth statistics on a rolling 12 months ahead basis. To make this calculation on, say, 1 October, for a company with a calendar financial year, it is necessary to take a quarter of the current year's brokers' consensus forecast and three quarters of the consensus forecast for the following financial year. By adopting this approach for all companies, chalk and cheese are converted into chalk and chalk, enabling tables to be prepared and meaningful comparisons to be made.

Once the characteristics of a growth share have been defined, the first value filter I apply to the resultant universe is to establish that the growth company's price-earnings growth factor (PEG) is attractive. The PEG is calculated by dividing the rolling 12 months ahead price-earnings ratio (PER) by the rolling 12 months ahead EPS growth rate. A share with a PER of 15 and an EPS growth rate of 10% per annum would therefore have a PEG of 1.5, which is about the average. The lower the PEG the better and to provide a margin of safety, I usually invest in companies with PEGs of well below 1.0.

To my mind the PER is a one-dimensional and very limited measure, especially when it is used to compare companies with widely differing financial year ends. I'd like to know exactly how much growth I am getting for my multiple and I like to be able to compare any share I am considering with every other share in my universe of growth shares.

The level of the PEG gives an instant fix on the growth you should get for your money in the coming 12 months. For longer periods the trend of EPS does, of course, have to be taken into account.

I eliminate any shares with rolling 12 months ahead PERs of over 20. In a few years' time, those PERs usually regress to a more normal level and the resultant fall in the rating then has to be amortised over the life of the investment.

EPS growth may more than offset the fall in the PER. However, a company with a very high PER must have a very high (and therefore probably unsustainable) growth rate to justify a low PEG. For example, a share with a PER of 30 and an estimated EPS growth rate of 50% would have a PEG of 0.6. A 50% growth rate would be too high to be maintained for many years. A far better proposition would be a company with same 0.6 PEG but based on a prospective PER of 12 and an estimated EPS growth rate of 20%. The margin of safety would be much better as the 20% growth rate might be sustainable for many years and the PER would be far more likely to be upgraded than downgraded.

My next value filter is to ensure that a company's cash flow is in excess of its EPS for the last reported year and for the average of the previous five years. If a company is expanding rapidly it can sometimes justify a one year deficit of cash flow per share in relation to EPS. However, it is noticeable that most great growth companies seem to have strong cash flow and as a consequence usually have net cash in their balance sheets. My next sieve - positive relative strength over the previous 12 months and previous month - cannot be described as a value filter. It is, however, very effective and there is a clear reason why positive relative strength is such an important factor when buying shares with low PEGs. The intention when investing in them is to benefit from their above average EPS growth and from the status change in their PERs when the market re-rates them. Positive relative strength is important because it indicates that the market is already beginning to appreciate the virtues of the company, so the wait for the re-rating should not be unduly long.

Usually, EPS growth might account for 20%-25% of an annual capital gain, but the status change in the multiple compounds this and is often far more significant. Take for example a company which had a share price of 120p, EPS of 10p, and a prospective growth rate of 20%. It therefore had a PER of 12 and a PEG of 0.6. If, after a year, EPS rose to 12p as expected, and the PER was re-rated to 18, the share price would increase to 216p. In that event, the resultant capital gain would be 96p, of which 75% would be due to the re-rating of the PER, and only 25% to the increase in EPS.

I use further sieves, but first see how well the basic method has worked since the inception of Company REFS. We have back-tested 14 six-month periods at monthly intervals ending on 28 October 1996. The average gains are set out below: