The End Of Management?

The

end of management just might look something like this. You show up for work, boot up your computer and log onto your company's Intranet to make a few trades before getting down to work. You see how your stocks did the day before and then execute a few new orders. You think your company should step up production next month, and you trade on that thought. You sell stock for the production of 20,000 units and buy stock that represents an order for 30,000 instead. All around you, as co-workers arrive at their cubicles, they too flick on their computers and trade.

Together, you are buyers and sellers of your company's future. Through your trades, you determine what is going to happen and then decide how your company should respond. With employees in the trading pits betting on the future, who needs the manager in the corner office?

That scenario isn't as farfetched as you might think. It's called a prediction market, based on the notion that a marketplace is a better organizer of insight and predictor of the future than individuals are. Once confined to research universities, the idea of markets working within companies has started to seep out into some of the nation's largest corporations. Companies from Microsoft to Eli Lilly and Hewlett-Packard are bringing the market inside, with workers trading futures contracts on such "commodities" as sales, product success and supplier behavior. The concept: a work force contains vast amounts of untapped, useful information that a market can unlock. "Markets are likely to revolutionize corporate forecasting and decision making," says Robin Hanson, an economist at George Mason University, in Virginia, who has researched and developed markets. "Strategic decisions, such as mergers, product introductions, regional expansions and changing CEOs, could be effectively delegated to people far down the corporate hierarchy, people not selected by or even known to top management."

To understand the hype, take a look at Hewlett-Packard's experience with forecasting monthly sales. A few years back, HP commissioned Charles Plott, an economist from the California Institute of Technology, to set up a software trading platform. A few dozen employees, mostly product and finance managers, were each given about $50 in a trading account to bet on what they thought computer sales would be at the end of the month. If a salesman thought the company would sell between, say, $201 million and $210 million worth, he could buy a security — like a futures contract — for that prediction, signaling to the rest of the market that someone thought that was a probable scenario. If his opinion changed, he could buy again or sell.

When trading stopped, the scenario behind the highest-priced stock was the one the market deemed most likely. The traders got to keep their profits and won an additional dollar for every share of "stock" they owned that turned out to be the right sales range. Result: while HP's official forecast, which was generated by a marketing manager, was off 13%, the stock market was off only 6%. In further trials, the market beat official forecasts 75% of the time.

Intrigued by that success, HP's business-services division ran a pilot last year with 14 managers worldwide, trying to determine the group's monthly sales and profit. The market was so successful (in one case, improving the prediction 50%) that it has since been integrated into the division's regular forecasts. Another division is running a pilot to see if a market would be better at predicting the costs of certain components with volatile prices. And two other HP divisions hope to be using markets to answer similar questions by the end of the year. "You could do zillions of things with this," says Bernardo Huberman, director of the HP group that designs and coordinates the markets. "The idea of being able to forecast something allows you to prepare, plan and make decisions. It's potentially huge savings."