Thoughts of the Week

Created 2000-04-04
Modified 2000-04-04
Go to DeLong's Home Page

Teaching | Writing | Career | Politics | Book Reviews | Information Economy | Economists | Multimedia | Students | Fine Print | Other | My Jobs

Brad DeLong's mailing list | Comments on this page | Questions and Answers

Is Big Bad?

J. Bradford DeLong

J. Bradford DeLong is professor of economics at U.C. Berkeley, a research associate at the National Bureau of Economic Research, and co-editor of the Journal of Economic Perspectives. This piece appeared in the Los Angeles Times on April 9, 2000.


Monday, Federal Judge Thomas Penfield Jackson found as a matter of law that Microsoft had violated the 110-year-old Sherman Antitrust Act. He will now begin the process of determining what remedy will be granted to repair the damage done by this illegal restraint of trade.
     It may be that this decision, shocking to the high-tech sector's stock-market valuation as it was, will wind up as a footnote. For, five years ago, Microsoft, with its dominance of desktop operating systems and productivity applications, was at the heart of America's high-tech economy. But today, because of the remarkable rate of change, the heart of the high-tech economy is the network. It is at least arguable that the key is now in the hands of physical-network companies like AT&T, data-delivery companies like Akamai Technologies, database companies like Oracle, Internet-access providers like America Online and the communities of open-source programmers who maintain and develop the Linux operating system and the Apache Web server. So what happens to Microsoft, specifically, is no longer as critical.
     But even if Microsoft the company is (relatively speaking) no longer as crucial in the new, larger, high-tech economy, Microsoft the case may still be key in the development of antitrust law. It will be a big step down the path of applying antitrust principles more than a century old to the new economy. Ever since the Industrial Revolution began, the United States has wrestled with the dilemma of antitrust: How to deal with highly productive monopolies. Can we somehow have our cake and eat it, too: Gain the benefits of efficient bigness and still keep the benefits of intense competition? Each time the technology changes, the form the dilemma of antitrust takes changes as well.
     America, as Thomas Jefferson or Abraham Lincoln imagined it, had no place for monopolies. For both, competition was an essential piece of individual liberty. If you didn't like the deal someone was offering, you could walk down the street and find an alternative trading partner. Monopoly--someone who made you poor, limited your freedom by telling you to "take it or leave it" and made that stick--was not something America could afford. But steam power and industrial machinery changed things. A single factory could produce enough to sell to a state or the nation. Andrew Carnegie's integrated steel mills, George Westinghouse's electrically powered machines and appliances, Gustavus Swift's steer-disassembly line and others promised huge efficiencies by producing at massive scale.
     The sheer magnitude of economies of scale made it seem likely that, left to itself, the market would produce monopoly after monopoly. But were the efficiency gains--lower costs--worth the loss of individual choice? Without competition, there was every reason to fear that lower costs would come with higher prices and a more unequal distribution of wealth.
     The attempt to carve a solution to the late 19th-century dilemmas of antitrust came with Ohio Sen. John Sherman, younger brother of Civil War Gen. William Tecumsah Sherman, and his 1890 Sherman Antitrust Act. When the dust of the first generation of antitrust litigation settled, the United States found itself with a legal system that tolerated oligopoly, in which four or five large firms competed against each other in an individual market, but sought hard to curb monopoly, a single dominant seller with overwhelming market share, through tight constraints on the monopoly's conduct or through court-mandated breakups of monopoly businesses that had used their power to restrain trade unreasonably. The hope was that this set of legal principles would get us most of the efficiency benefits possible from economies of scale--the oligopolistic firms were extremely large--and most of the benefits of competition as well--the oligopolistic firms would compete against each other.
     Back then, antitrust law mattered, for the large companies it affected had enormous scope and reach throughout the economy. Nearly every construction project used U.S. Steel's steel. The court case that broke up Standard Oil and the threat of antitrust action that turned AT&T into the highly regulated Bell system that we knew before the early 1980s shaped America's economy.
     During the Great Depression, another principle was added to the mix. Congress decided the government should put its thumb on the scale on the side of relatively small producers: There was a public purpose to be served in outlawing practices that made it difficult for the small-scale producer to survive alongside oligopolistic giants, even if they were not the lowest-cost producers, even if the giants' practices were not necessarily unreasonable restraints.
     Thus, antitrust became a tangled web of different laws pursuing contradictory purposes. For most of the past 50 years, lawyers, judges and analysts have gradually picked that web apart. The Chicago school of economists shifted from being enthusiastic advocates of aggressive antitrust enforcement to advocating a hands-off position. Future Judge Robert H. Bork suggested that antitrust law be revised to focus solely on consumer welfare.
     Over the years, Bork's view gained strength. People scratched their heads as they watched the virtual repeal of antitrust statutes like the Robinson-Patman Act undertaken by those who in other areas (civil-rights law, say) exalted the original intent of legislatures and decried judge-made law. But it was not clear that small-scale producers deserved a special edge. A looser approach recognizing that close business links could be beneficial seemed likely to make us all better off.
     Now, however, new technology has once again ripped open the seams. In the days of Standard Oil, to be twice the size of your competitors meant your unit costs were perhaps 10% lower. In these days of the information economy, a much larger share of costs are fixed and sunk: The program has to be written and debugged only once, no matter how many copies are sold. Thus, to be twice the size means your per-unit costs are little more than half as much. The complexity of information-age products means there are subtle dependencies across markets: It seems to be easier to get Microsoft FrontPage working well when the Web server it uploads files to is running Microsoft Internet Information Server rather than when it is running open-source Apache. In addition to supply-side economies of scale, there are demand-side economies of scale: If, say, four of your five co-workers are using Microsoft Word, than either you use it, too, or see a fifth of your life vanish into dealing with format glitches.
     Moreover, the benefits to bigness, or at least coordination, seem larger in the information age. Software for minicomputers stagnated in the 1980s because each brand's version of the Unix operating system was incompatible with the others. The World Wide Web has boomed in the 1990s because its inventor, Tim Berners-Lee, made the software protocols available to everyone for free.
     In the early 20th century, the oligopolies that antitrust established were stable: Economies of scale were limited, so that General Motors had a cost advantage but not that much of a cost advantage over Ford. Our new technologies have far larger and stronger economies of scale, meaning that, as economists Hal R. Varian and Carl Shapiro have written, markets will not and cannot look like the competitive markets of ideal economic theory. Given the size of the economies of scale, it is not clear we want them to.
     So what does this mean for antitrust? First, we do not know how much antitrust law can matter. The economy is so much bigger today that even its largest companies play a smaller role than U.S. Steel, Standard Oil or AT&T did a century ago. The Microsoft case cannot have as much economy-shaking impact. Technology also seems to be moving sufficiently rapidly to make whatever antitrust remedy is reached in that case of doubtful relevance: The market and the industry will have changed too much. The smaller impact of any single case and the difficulty of keeping pace with changing technology may end up making judicial antitrust remedies irrelevant.      Perhaps courts and prosecutors will try to maintain the standard pattern: Tolerate oligopoly, break up monopoly. If so, antitrust authorities will have a busy time as they watch economies of scale create a dominant natural monopoly in sector after sector, then move to break up the monopoly and restore competition. Will such a pattern lead to an efficient and productive economy? We are not sure.
     A second possible direction would be to have greater tolerance for monopolies that played fair: to focus on establishing and monitoring a code of conduct for information-age natural monopolies that allows us to reap all the efficiency benefits of bigness and still maintain a degree of virtual, if not real, competition. But can such a code of standard-setting friendliness be specified and enforced? We are not sure.
     There are other directions in which antitrust law could move, directions that will come as a surprise, but that, perhaps, the Microsoft case will foreshadow. However, the dilemma of antitrust will remain: bigness vs. freedom of choice, competition vs. economies of scale.

Sign up for Brad Delong's (general) mailing list

Add a comment on this page...

Thanks for the info. I really enjoy your information. BTW, have you read any of the Leibowitz/Margolis skeptical regard for increasing returns?

Contributed by ( on April 25, 2000.

I appreciated reading your cogent analysis, but I do think you missed 1/3 of the triad when you wrote

However, the dilemma of antitrust will remain: bigness vs. freedom of choice, competition vs. economies of scale.

without adding also: "fair play" vs. unethical business practices

Your statement "A second possible direction would be to have greater tolerance for monopolies that played fair" was a step toward this conclusion, but did not go far enough.

Indeed, I think the major impact of the MS trial will be in the "fair play" area. For this reason, I cannot agree with your statement:

" ... Technology also seems to be moving sufficiently rapidly to make whatever antitrust remedy is reached in that case of doubtful relevance: The market and the industry will have changed too much. "

If MS gets past this trial with only a slap-on-the-wrist, many other companies will be encouraged to adopt extreme marketing practices like those described for MS in the judge's findings. This could cause a widespread deterioration of the high tech marketplace. As there would be less competition going forward, technical progress would slow generally (just as MS has made very little technical innovation for years), and prices would cease to decline (becoming stable just as MS prices have done). I would call that outcome a major impact on the economy.

Conversely, if MS is seen as being called to task over their practices, then other technology companies will be able to take the risk of competing in areas that were previously too close to MS' sphere of influence to be "safe". In this scenario, high tech will undergo yet another surge of creativity and growth.

Thanks again for putting out a thought provoking piece.

Contributed by Peter Eirich ( on April 25, 2000.

I read your article in today's L.A.Times with considerable interest. But I think you have confused two things in your discussion of economies of scale.

There are indeed considerable benefits to everyone using the same software, or doing things the same way. But this benefit stems from standards, not market domination. We've known for a long time that even a bad standard is better than none, in nearly all situations.

You partly recognize this distinction by referring to "the benefits of bigness, or at least coordination." But it really is coordination that's involved, not bigness and predatory policies like Microsoft's.

Your argument is also needlessly weakened by a couple of statements of dubious validity. For example, you say "It seems to be easier to get Microsoft FrontPage working well when the Web server it uploads files to is running Microsoft Internet Information Sever rather than when it is running open-source Apache." I don't think this is true. There's no logical connection between the server-side software and the Web-page "authoring tools" (as they are known these days). What I think you are referring to is the connection between Microsoft's FrontPage and its browser, which is full of proprietary (and non-standard) features. It's possible that Apache might have difficulties serving Web pages that contain some of the more esoteric features (or bugs!) of the Microsoft software; but Apache has a reputation for reliability and robustness, which is why it has taken over the great majority of the Web servers. So I doubt that there's really a problem there. In any case, there's no reason why the Apache server should make FrontPage work less well. Your may be talking about a real interoperability problem, but if so, I think you have mis-identified it.

Another shaky statement is the assertion that "Software for minicomputers stagnated in the 1980s because each brand's version of the Unix operating system was incompatible with the others." I've used many of those different "flavors" of Unix. While the differences are a nightmare from a system administrator's point of view, the differences are really quite minor from the user's perspective. And why? Because the UNIX community quickly realized they needed standards, and agreed on a set (ever hear of "POSIX"?). Instead, the real problem was that AT&T, the copyright owner of UNIX, began restricting access to the source code. Whereas they had formerly given away the source free, or nearly so, to universities -- which led to the great profusion of new and productive features at your own university, so that "Berkeley UNIX" is a well-known phrase in the software comminuty -- AT&T started charging something like $50k or more for a source license. It was these restrictions that led to the invention of Linux; and this open-source community has now breathed fresh life into the UNIX movement. In short, now that source code is freely available, there are thousands of people working on improvements to Linux, while the closed-source Unices languish and are full of bugs and problems (I can attest to having encountered several in the past year; our system administrator has dealt with these matters by replacing the proprietary utilities with open-source versions that work better.)

Here, again, the benefits flow not from bigness and market domination, but from standards and openness, as opposed to proprietary "closed" code and conventions.

An additional point not covered in your article is the notorious lack of innovation displayed by Microsoft. It's well known that the "Windows" model was copied from Apple's user interface -- and it's a second-hand theft at that, as the whole thing arose at Xerox PARC rather than at Apple. Microsoft wasn't interested in the Internet until Netscape began to be successful.

There are many other examples, but you get the idea. The result has been that Microsoft has done rather badly in areas outside of PC operating systems, where they managed to obtain a monopoly early and dominated the market. Look at their difficulties in the hand-held area, where Windows CE is regarded as slow and underpowered, and has only a relatively small fraction of that market. A lot of what Microsoft touts as innovation arose in other companies they later acquired by using their monopoly power. Many people in the software industry have argued that consumers would have much better PC software today if Microsoft had not deprived them of the opportunity to have it. The inroads that Linux is making on the desktop today suggest this is true.

It sounds as if the anti-trust people are re-fighting previous wars, and don't understand much about what's important in the area of software. It's not that times have changed, but that this is a different type of business than the heavy industries that were the monopolists of the past. The efficiencies in software come from standards, not bigness. And standards thrive in an atmosphere of cooperation, not cut-throat competition.

It's likely that the most productive move the courts could make would be to force Microsoft to publish its source code. Open code would quickly get cleaned up and improved; consumers would benefit; and in the long run, Microsoft would have won the battle for domination it has fought for so long. "Breaking up" Microsoft looks like a futile effort, in contrast. This isn't the phone company, or Standard Oil. --

Contributed by Andrew T. Young ( on April 10, 2000.

>Your argument is also needlessly weakened by a couple >of statements of >dubious validity. For example, you say "It seems to be >easier to get >Microsoft FrontPage working well when the Web server it >uploads files to is >running Microsoft Internet Information Sever rather than >when it is running >open-source Apache." I don't think this is true....

I had thought that the server extensions were the key: that they worked *well* on IIS and were very buggy elsewhere...

>Another shaky statement is the assertion that "Software >for minicomputers >stagnated in the 1980s because each brand's version of >the Unix operating >system was incompatible with the others." I've used >many of those different >"flavors" of Unix. While the differences are a nightmare >from a system >administrator's point of view, the differences are really >quite minor from >the user's perspective.

But not from the viewpoint of the software producer... Phil Greenspun, for example, strongly recommends Suns because Oracle develops on Suns, and then tackles other dialects of Unix...

But since I beat up on Charles Ferguson in the _Harvard Business Review_ for ignoring open source, it's only fair that I get some of the same medicine.


Contributed by Bradford DeLong ( on April 10, 2000.

Go to related links...

Professor of Economics J. Bradford DeLong, 601 Evans Hall, #3880
University of California at Berkeley
Berkeley, CA 94720-3880
(510) 643-4027 phone (510) 642-6615 fax

This document:

Search This Website