Assuming that Microsoft has monopoly power both in the market for operating systems and in the market for applications software, would consumers be better off if those divisions of the company were split from one another, as the government's antitrust lawyers reportedly propose? Paul Krugman's column in the New York Times on Wednesday, April 26, addressed this issue. He applied some basic economics.

First, let me say that with regard to Microsoft, I continue to hold to the view that Netscape went out of business because it did a lousy job with writing and supporting its server software. I do not believe that the browser market was critical for Netscape's survival. Therefore, I believe that the best outcome would be for the government's case to be thrown out. I also wonder how a case that began as a complaint about the way that Microsoft tied its browser to the operating system could end up with this particular remedy, which reportedly includes having the browser be given to the operating system company!

However, if we start with the assumption that something ought to be "done" about Microsoft, then I tend to favor the breakup of the company into an operating system company and an applications company. Krugman argues otherwise, and his point was something I had not considered.

To illustrate Krugman's argument, suppose that a single company has a monopoly in both peanut butter and jelly. When it sets the price of jelly, it knows that the more jelly it sells the more peanut butter it will sell. Therefore, at the margin, it will tend to want to set a lower price for jelly than if it were just looking at jelly as a stand-alone product.

If you then break up the PB and J monopoly into two separate companies, the incentives of the two separate monopolies will change. The peanut butter company is not going to worry about the fact that higher peanut butter prices will reduce jelly consumption, and the jelly company is not going to worry about the fact that higher jelly prices will reduce peanut butter consumption. The net result of the breakup is that prices to consumers will rise.

If you could break up the peanut butter and jelly monopoly into a competitive industry, that would be a good thing. However, breaking it up into separate monopolies in peanut butter and jelly is a bad thing.

A general rule of thumb concerning monopolies is that anything which gives the monopolist an incentive to charge a lower price is good for consumers. Krugman's analysis suggests that allowing Microsoft to keep its monopolies in the operating system and in applications will lead to lower prices for consumers than would be the case if you were to break Microsoft into two separate monopolies. Therefore, two monopolists would be worse than one both for consumers and for Microsoft shareholders.

[Incidentally, this same logic should make companies think twice about creating separate "profit centers" for complementary products. For example, when I was with Freddie Mac, it set up a separate profit center for automated underwriting and for purchasing mortgages.

From the mortgage lender's perspective, these are complementary services that go into the total cost of making a loan. They are like peanut butter and jelly. What Freddie Mac was doing was creating incentives for its separate profit centers to try to over-price their products, thereby reducing demand and lowering overall profit.

Of course, Freddie Mac is not a monopolist. It probably is the case that its competitor, Fannie Mae, was able to discipline Freddie Mac's pricing, regardless of the perverse internal incentives.

My guess, however, is that there are many examples of companies that shoot themselves in the foot by creating separate profit centers for complementary products. If I need Paul Krugman to remind me about the economics of this, then my guess is that so do many corporate executives.]

Two monopolies are not better than one, according to the economics of peanut butter and jelly. However, this model fails to capture the intuition of those of us who think that there is a benefit to this sort of break-up of Microsoft.

The reason some of us like the idea of two monopolies is that we think that each monopoly might be able to reduce the other's power. To see our point, we need to enrich the metaphor. Assume that our near-monopolist produces only creamy peanut butter and grape jelly, and that these products inter-operate well with one another but poorly or not at all with competing chunky peanut butter and strawberry jam, which have low market shares.

Now we break our near-monopoly into two companies. The grape jelly company suddenly has a bigger incentive to inter-operate with chunky peanut butter. The creamy peanut butter company now has a bigger incentive to inter-operate with strawberry jam.

The ultimate effect on the prices paid by consumers is difficult to predict. It feels like one of those models where the results are ambiguous until you know some empirical elasticities, economies of scale, and such.

Even thinking in terms of this extended metaphor, I do not believe that the case for breaking Microsoft into an operating system company and an applications company is particularly strong. However, it is not as weak as Krugman suggests.