Gambits of Big Blue-sky thinking

In 1991 I blew my big chance. The story begins in 1989, when the guy who was to make Deep Blue, the chess supercomputer, for IBM was my graduate school officemate at Carnegie Mellon. His little side project was called “Chiptest” and I would play the chess program all day long, to the dismay of my professors.

Although I was a US-ranked master, Chiptest (and its successor, Deep Thought) would destroy me nearly every game. Then IBM brought the team over and I was left alone.

That is, until the summer of ’91 when I was offered the chance to work on the Deep Blue project. I turned it down. “IBM?” I thought. “Don’t they make those crappy PCs and some mainframes?”

I regretted my decision for years, particularly when in 1997 Deep Blue beat world champion Garry Kasparov.

When I think PC I think IBM PC because my friends and I would play games on our IBMs. I knew two other things about the company: it was into mainframes and had been around for ever. Also, it had something to do with typewriters way back when.

Now IBM has more than $91bn in revenues, $19bn in cash flows and nothing to do with PCs. In 2004 it sold its PC business to China- based Lenovo Group for $1.75bn. Lenovo became a PC powerhouse, its IBM blessing taking it to revenues of more than $12bn, a fourfold increase.

So what does IBM do? Where do the revenues come from and, more importantly, is IBM cheap now because of people on Wall Street have no clue what it does? As a friend of mine who runs a long/short hedge fund said: “It’s all about mainframes now, but it’s slow-moving, low-margin growth.”

Only 24 per cent of revenues come from mainframes. The rest is software and services. IBM’s $48bn services business has undergone dramatic changes, with more emphasis on higher-margin engagements. Its services division has had eight consecutive quarters of year-to-year improvement in gross profit. And those margins drove cash flows: $15.3bn in 2006 compared with $14.4bn for Microsoft and $11.4bn for Hewlett Packard; more than the cash flows for Dell, Oracle and Accenture combined.

That has allowed the company to go on a mega-buying spree since 2004, acquiring 34 software companies, all in high-growth niches. In 2006 alone, IBM invested $4.8bn for 14 acquisitions (11 software, two services). Its investments in systems management software companies, including the Micromuse acquisition, have contributed to the strong performance of Tivoli (revenues up 44 per cent per cent year-over-year in the third quarter of 2006 and 25 per cent in the fourth quarter).

What has it been doing with the rest of the money? Giving it back to shareholders and quietly taking itself private. Since January 2003 the company has returned $32.4bn to shareholders ($27.2bn through share buy-backs and $5.2bn in dividends). In the past year it bought back more than 1.2bn shares. Companies often do this and it doesn’t reduce their share count because they issue options to employees, use stock to make acquisitions, issue convertible debt and take other dilutive measures.

But in IBM’s case share count has gone down every year since 2003, when it had 1.73bn. It now has 1.553bn shares, a reduction of 10 per cent. Again, slowly taking itself private. This has the effect of increasing earnings per share and creating an imbalance in the supply versus demand equation for the shares.

In terms of the low-margin business, it is worth comparing IBM to its closest competitors – in particular Hewlett Packard. In fiscal 2006 IBM had $91.4bn in revenues while HP had slightly more at $91.7bn. Pre-tax income for IBM was $13.3bn, for HP $7.2bn. IBM’s overall net income margin was 13.2 per cent compared with HP’s
6.7 per cent. Not bad.

I also like the fact that IBM is included in the Ocean Tomo 300 Patent Index, an equity index based on the value of corporate intellectual property. It is based on academic research that shows that companies spending money on research and development tend to outperform broader indices. Wall Street analysts count R&D as expenses and it adds to the intangible book value of a company. In other words, the typical analyst looks at R&D as a negative, even though in the long term it pays off.

IBM received 3,621 patents last year, bringing the total to more than 40,000. No other company has received even 3,000 patents in a year. It got more patents last year than Google, Dell, Apple, Microsoft, General Electric and EMC Corporation combined. Another exchange-traded fund that includes IBM is the PowerShares Lux Nanotech Index.

Although IBM is known as a mainframe company, a big portion of its research is focused on building nanotube transistors. Commercialisation is years away, but when it happens people might forget completely that IBM had something to do with typewriters back in the day.