Arm too far ahead

What most influences that conclusion is the relationship between value-added per employee and the market value of Arm's equity. The underlying logic is that in the knowledge economy - and what type of company could epitomise the knowledge economy better than one that designs templates for microchips? - value-added per employee should have a big influence on a company's performance and hence its share price. In practice, that correlation is hardly clear cut. Even so, the link is sufficient to question why Arm's share price has motored so strongly. At 204p, it is more than double its 2009 low of 81p - and, for those who pay attention to such things, the shares are rated on a multiple of 28 times 2010's forecast earnings.

The reason for this price surge is that, trading-wise, 2009 was a lousy year - globally, shipments of semiconductors were down about 20 per cent on the year - but 2010 will be better. That's obviously good for Arm because it receives a small royalty every time a microprocessor based on its chip templates is sold. So more shipments mean more revenues. And since Arm receives its royalties in the calendar quarter after the product was shipped, its bosses know that the year will start off on a growth tack then, hopefully, improve as it progresses.

On top of this, Arm has gone into the new year with a lower cost base and a bigger share of its markets. Indeed, its management assumes that, as usual, the pace of Arm's revenue growth will exceed its markets'. If semiconductor shipments grow by 15 per cent or so, as analysts expect, then Arm's revenue growth may be a couple of percentage points better - enough to feed through to earnings growth of around 50 per cent to around 7p a share, or so analysts reckon.

Investors also love Arm's shares for the assured stream of revenues that will flow from the company's product development cycle. Arm's aim is to spend big sums developing the right designs, sell licences to use the designs to manufacturers of electronic components who then spend three or four years developing products based on Arm's template. Then, assuming the finished products are right, their sales will generate royalties for Arm for perhaps the following 20 years. Marvellous. Not just marvellous, but the business model works - proof of which is the fact that around 4bn Arm-based microchips are shipped every year and sometime this year microprocessor number 20bn should come out of the foundry.

So, it's a great story but, as I implied earlier, the share rating is also great - arguably too great. The problem is that Arm, while hardly mature, is not the exceptional young thing that it was eight or nine years ago. As the table shows, peak value-added per employee was reached in 2001-02. Since then, value added, which is basically operating profits with depreciation and amortisation added back, has (a) fallen sharply and (b) got itself into a rut of between £40,000 and £50,000 per employee.

Late last year (Bearbull, 27 November 2009) I introduced readers to the notion that there should be a link between value-added per employee and a company's market value. This was based on Modelling the growth of corporations (Palgrave) by a trio of academics and business people. The challenge, however, is to establish the predictive merit of value-added per employee. That's tricky because there is limited data available. Companies report their full-year profits and employee numbers only once a year yet a regression analysis, where we plot value added against stock market value, will need a 10-year time horizon just to be minimally useful.

In Arm's case, the data points for the years 1999-2002 are all over the place. Since then, they have been clustered together neatly. This implies that back in its early days investors really did not have a clue how to value Arm's shares but have got better at it as familiarity grew and Arm's pace of growth moderated. But still not that good, perhaps. After all, Arm is on course in 2010 to produce its best level of value-added per share since 2002, but that's not necessarily a bullish sign. As value-added per employee rose to its high point during 1999-2002, Arm's share price was on a downwards helter skelter. In other words, the share price may be rather good at anticipating the forthcoming direction of value-added per employee. Coupled with that, my other models struggle to find much more than 120p per share of value in the company's equity. Meaning that, at 204p, the price may have been carried away.