Why? High technology has always been risky business. Because of important network effects, technology is often only profitable after it has been widely disseminated. After all what is the point of being one of the first few people to have access to the internet, or to have a telephone?

For the economic benefits to exist, someone must be willing to take a huge risk and must have an almost foolish commitment to the technology. History suggests that technology first movers are vital and yet almost always go bankrupt.

In nearly every case after massive investments into enormous, and often enormously unprofitable, technology networks, the original builders failed, and their investors bore huge losses. This is what happened to the pioneers who built the first major railway networks (several times), the telegraph lines, the electricity networks, the phone lines and, most recently, broadband connections.

After the bubble euphoria wore off and excited talk about the brave new world being created was replaced by boring calculations of discounted future cashflows, investors fled and bankruptcies surged.

When can this kind of investment possibly happen except under bubble conditions, when investors are willing to give huge amounts of money to risky proposals? But after the bubbles imploded, there were often subsequent benefits that more than offset the cost to society.

Too much capacity was created for the amount of business that could be supported, and in well-functioning bankruptcy regimes, like those of the US and Britain, this expensive capacity was forced into liquidation, and new investors were able to snap up assets at pennies on the dollar - to the great chagrin of the earlier enthusiasts. This may have been bad for the old investors, but it was a real boon for future users of the technology.

Suddenly the usage cost collapsed, and as a consequence whole new and often unexpected industries were created around the now low-cost networks. What would Google's prospects be, for example, if the whole world hadn't been wired by the likes of the much-lamented WorldCom and Global Crossing?

Overcapacity is a problem for old investors but can be a real boon by creating conditions for an explosion of new businesses. The key is whether or not capacity can be forced into liquidation by new investors and so made available at very low prices to as yet unimagined new businesses.

The debate about rising capacity in property and tradable goods on the mainland centres on whether the government's stimulus package is resolving economic problems or exacerbating overcapacity.

If it is the latter, there will be heartache and disappointed investors.

But if Chinese policymakers are hard-hearted about misallocated investment and force liquidation, the excess property projects, steel, chemicals and bandwidth can be made available at market-clearing prices and transform future business opportunities. Bubbles may have short-term costs, but it is often only bad policy that ensures long-term costs too. It is not a coincidence that major stages of the industrial revolution are nearly always associated with financial bubbles.

Michael Pettis is a professor of finance at the Guanghua School of Peking University