GEOFFREY COLVIN – Sam Palmisano’s legacy of leadership at IBM

Click here to learn more about Geoffrey Colvin, Speaker, Journalist and Author

The CEO steps down leaving his company in great shape. Will that legacy last?

FORTUNE—Really good CEOs are rare partly because they use all of their brains. Most people don’t. We carry around either massive left brains—logical, analytical, not very touchy-feely—or, more rarely in today’s world, we rely on big right brains—imaginative, feeling, intuitive, uncomfortable with algebra. Running a giant company obviously demands both, and the best business leaders can bat from either side of the plate. That group includes IBM chief Sam Palmisano who’s stepping down at year-end after nearly a decade as CEO, looking like a business Mickey Mantle. The business world’s ultimate endorser, Warren Buffett, announced recently he’d bought $10.7 billion of IBM stock, saying Palmisano has “delivered bigtime.”

In management the brain’s two halves are the financial side and the human side. Most managers, if they’re lucky, are really good with one or the other.

Palmisano was excellent with both. On the financial side, he understood what actually makes a company valuable, which I regret to report many CEOs do not. The clueless ones do dumb things like pay too much for acquisitions in an effort to plump earnings per share. Palmisano understood that it’s all about capital. He took capital out of businesses with poor returns and put it into businesses with high returns. For example, heoffloaded IBM’s disk drive business to Hitachi (HIT) and then made a five-year deal to buy Hitachi drives; realizing that drives were becoming a commodity, he decided he’d rather be buying them than selling them. He also bought about 100 companies, mostly small ones in the high-return-on-capital businesses of software and services. He squeezed working capital, an unglamorous tactic that pays handsomely.

As a result, IBM’s return on capital has increased from 4.7% when Palmisano got the job to 15.1% today, a tremendous achievement for a company of IBM’s size (estimated 2011 revenue: $107 billion). And as the company grew stronger while interest rates declined, IBM’s cost of capital fell, so the all-important spread between capital’s return and cost widened (data from EVA Dimensions). That’s a big reason IBM’s (IBM) stock is 82% higher than it was when Palmisano took over in 2002—higher even than its bubble-market peak.

2011 Businessperson of the Year

The human side of the CEO’s job is making sure those knockout results keep coming by developing tomorrow’s leaders. Palmisano realized that the most valuable learning is delivered not by a teacher but by the world. “We made a decision to walk out of the classroom and make sure most of the experiences were being lived at the operational level,” says Randall MacDonald, IBM’s HR chief through all of Palmisano’s tenure.

So the company increased its commitment to developmental assignments and invented new ones. For example, it created a Corporate Service Corps, which combines diverse young up-and-comers into teams that work with local leaders on local problems around the world. The employees get stretched and developed; IBM gets insight into their abilities.

Lots of companies send managers from developed economies to emerging ones. IBM also does the opposite, so those emerging- market managers get experience in a mature economy, preparing them for the future. Such efforts are a big reason IBM ranks No. 1 on the new global list of Top Companies for Leaders.

Are those practices worth the trouble and expense? No one can say for sure; much of the payoff from leadership development is far down the road. The larger point is that CEOs can’t be fully assessed when they retire because their decisions play out, for better or worse, over many years after they’re gone. Palmisano made major long-term bets on analytics, supercomputing, and the Smarter Planet initiative, which aims to harness all of IBM’s abilities to solve the world’s biggest social as well as business problems. The performance of his successor, Virginia Rometty, will also be a major element of his record.

So this is still a preliminary report. We have yet to see how well Palmisano managed IBM for tomorrow. But he’s going out like an all-star.

This article is from the December 12, 2011 issue of Fortune.