Furthermore, although these are important ideas, they aren't Big Ideas. They explain certain phenomena well, but they have limited reach. Those in search of a really big idea had to look further back in the economics literature. They hit gold with ''The Nature of the Firm,'' a 1937 paper written by the Nobel laureate Ronald Coase.

The Coase paper asked a deceptively simple question: If the market is such a great tool for allocating resources, why isn't it used inside the firm or company? Why doesn't one worker on the assembly line negotiate with the worker next to him about the price at which he will supply the partly assembled product?

That sort of negotiation rarely happens. Instead of using markets, companies tend to be organized as hierarchies, using a chain of command and control rather than negotiation, markets and explicit contracts. Paradoxically, the primary unit of capitalism, on close inspection, looks a lot like central planning.

Mr. Coase didn't just ask this question; he also provided a provocative answer: it all hinges on the costs of making transactions. What economists call firms, he said, are essentially groups of activities for which it is more effective and less costly to use command-and-control than markets to have things done.

New-economy advocates found this a compelling idea. One consequence of the Internet has surely been to make it cheaper to communicate. This should, in turn, lower transaction costs and change company boundaries. Their conclusion was that companies would inevitably downsize and outsource, spin off unnecessary functions, and carry out more and more transactions using the Internet instead of internal memos.

Not so fast. The Internet lowers communication costs, that's for sure. But that means it lowers transaction costs within organizations as well as across organizations. The internal memo might disappear, but only because it is replaced by the internal e-mail message.

It just doesn't follow that lower communication costs lead to smaller companies. In fact, Mr. Coase himself said that ''changes like the telephone and telegraphy, which tend to reduce the cost of organizing spatially, will tend to increase the size of the firm.''

There's a lot of other evidence in the economics literature for Mr. Coase's observation. The Harvard business historian Alfred D. Chandler's classic work ''The Visible Hand'' documented how the deployment of the telegraph and railroad led to the creation of the giant corporation.

Maybe the Internet's role is to provide the inexpensive communications that can support megacorporations. This thought is enough to make a new-economy guru shudder.

What do the facts say about company size? Alas, they are inconclusive. From 1962 to 1992, the average size of a company hardly budged, and the small reduction that did occur could be attributed to the increased importance of the service sector, which tends to have smaller companies.

Later on in his article, Mr. Coase acknowledges that inventions like the telephone might reduce a company's size, if they reduce the costs of using markets by more than they reduce internal communication cost. In other words, it could go either way. We need a deeper understanding of transaction costs to resolve the issue.

So what are the important transaction costs? Mr. Coase, citing another economist, mentions three general categories: search and information costs, bargaining and decision costs, and policing and enforcing costs.

The Internet certainly reduces search and information costs, but, as we have seen, this cuts both ways. Bargaining and decisions still require a team of managers and lawyers sitting around a table. What makes contracts easier is codification and standardization, trends that are important, but are not greatly affected by the Internet, at least so far.

Policing and enforcing costs are the most relevant category. The reason the assembly-line worker doesn't negotiate with the person next to him is that it's too easy for him to say, ''Give me a good deal or I'll stop the line.'' Putting all the assembly-line workers under command-and-control reduces this sort of opportunistic behavior, at least as long as it can be easily observed.

Oliver Williamson, a significant contributor to transaction cost economics, argues that the temptation to be opportunistic is a major component of transaction costs, and hence a major determinant of the boundaries of the company. If certain suppliers are critical to your success, you want them inside, under your control, not outside, where their objectives may differ from yours.

The temptation to be opportunistic hasn't been affected much by the Internet. The Internet has made it easier, however, to monitor some sorts of contracts. Nowadays it is easy to tell if your supplier's claim that it shipped the parts to you on time is true; you can get the tracking number and check the shipper's Web page. Within the company, smart cash registers, inventory management systems, vehicle monitoring systems and the like help ensure that workers are performing their tasks, even when not directly observed.

Computers are monitoring more and more contracts, and that may well, in time, make companies more comfortable about letting functions move outside the command-and-control boundaries.

The real issue confronting a company trying to decide if some unit will be inside or spun off is what incentives the spinoff will have. If you spin off something critical to your business, you leave yourself open to extortion down the road. An internal monopoly supplier may be sluggish and inefficient, but its incentives are at least party aligned with the rest of the organization. An external monopoly can be much worse.

Mr. Coase's point about transaction costs' determining how organizations are structured is both profound and subtle. The Internet certainly affects transaction costs, but determining whether that means companies will be bigger or smaller requires careful analysis of competing forces.