Microsoft is one of the largest and most valuable companies in the world. The document discusses key lessons that can be learned from Microsoft's behavior, as its actions have huge effects on the computer industry. Specifically, Microsoft's behavior shows how a dominant firm can use tying of complementary products and exclusive dealing to remain dominant over time, even as products change dramatically. This strategic behavior in a dynamic network industry is the central economic lesson. The document also notes that antitrust issues in such cases can be complex, and remedies are difficult to fashion without unintended consequences.
Microsoft is one of the largest and most valuable companies in the world. The document discusses key lessons that can be learned from Microsoft's behavior, as its actions have huge effects on the computer industry. Specifically, Microsoft's behavior shows how a dominant firm can use tying of complementary products and exclusive dealing to remain dominant over time, even as products change dramatically. This strategic behavior in a dynamic network industry is the central economic lesson. The document also notes that antitrust issues in such cases can be complex, and remedies are difficult to fashion without unintended consequences.
Microsoft is one of the largest and most valuable companies in the world. The document discusses key lessons that can be learned from Microsoft's behavior, as its actions have huge effects on the computer industry. Specifically, Microsoft's behavior shows how a dominant firm can use tying of complementary products and exclusive dealing to remain dominant over time, even as products change dramatically. This strategic behavior in a dynamic network industry is the central economic lesson. The document also notes that antitrust issues in such cases can be complex, and remedies are difficult to fashion without unintended consequences.
Microsoft is one of the largest and most valuable companies in the world. The document discusses key lessons that can be learned from Microsoft's behavior, as its actions have huge effects on the computer industry. Specifically, Microsoft's behavior shows how a dominant firm can use tying of complementary products and exclusive dealing to remain dominant over time, even as products change dramatically. This strategic behavior in a dynamic network industry is the central economic lesson. The document also notes that antitrust issues in such cases can be complex, and remedies are difficult to fashion without unintended consequences.
Business Economics J anuary 2000 47 TheLessons fromMicrosoft
he Microsoft antitrust case has generated lots
of attention, and for good reason. 1 Microsoft is one of the largest companies in the world. Ranked by value of equity, it was the largest as of March 2000, but as of September 2000 was only the fourth largest. Its equity at that time was worth over 350 billion dollars, down from about 550 billion dollars in March. Its growth has been phenomenal. Its equity was not even publicly traded twen- ty years ago, and just ten years ago its stock was worth less than five billion dollars. The actions of Microsoft have huge effects on other firms in the computer industryan industry that, like Microsoft, has exploded in size. It is also an industry responsible for innovation and the recent booming productivity growth of the U.S. economy. With the Internet exploding and reshaping U.S. industry, Microsofts ability to influence the direction of develop- ments in the computer industry raises enormously impor- tant issues of public policy. Though important on its own, the Microsoft case should command our attention because of some very gen- eral economic lessons it provides and because of some hard policy questions that it raises about the appropriate role of antitrust law and enforcement. I want to describe the key economics lessons that we should learn from Microsofts behavior, because its behavior teaches econo- mists a lot about strategic behavior in a rapidly changing environment. Until the Microsoft case, the importance of this type of behavior was neither understood nor appreci- ated, despite the existence of other instances where it seems to have occurred, as for example in some of the IBM cases of about twenty years ago. However, just because we The Lessons from Microsoft MICROSOFTS BEHAVIOR TEACHES ECONOMISTS A LOT ABOUT STRATEGIC BEHAVIOR IN A RAPIDLY CHANGING ENVIRONMENT. By Dennis W. Carlton Dennis W. Carlton is a professor of economics in theGraduateSchool of Business, TheUniversity of Chicago, wherehis teaching and research center on microeconom- ics, industrial organization, and antitrust. Heis also president of Lexecon, Inc., an economic con- sulting firmspecializing in the application of economics to litigation. Heearned both his Ph.D. in economics and his MS in Operations Research fromMIT and earned his AB in applied math- ematics and economics fromHarvard College. This paper is based in part on my presentation to theNational Association for Business Economics, September 2000, and on my Alexander Brody Distinguished ServiceLecturedelivered at Yeshiva University, March 2000. This paper examines thelessons economists can draw fromtheMicrosoft litigation. Thepaper explains that themost interesting economic lesson is how strategic behavior involving tie-in sales and exclusionary acts can beused in a dynamic network industry to allow an ini- tial monopolist to remain themonopolist even in theface of rapid technological change. Thepaper goes on to explorethecomplications for antitrust policy that Microsofts conduct and theproposed remedy in the antitrust caserepresent. T 1 United States v. Microsoft, Civil Action No. 98-1232. I have served as a consultant for Sun Microsystems and others in matters adverse to Microsoft. 48 Business Economics J anuary 2001 TheLessons fromMicrosoft now understand certain strategic behavior does not neces- sarily mean that this behavior should always be attacked as an antitrust violation. I will use this new understanding to briefly discuss some of the antitrust issues that this case raises. The major insights fromthe Microsoft case that I dis- cuss are as follows: Microsofts behavior shows how tying of complemen- tary products and exclusive dealing can be used in a dynamic network industry so that a dominant firmcan remain dominant over time, even though the products change dramatically over time. This is the central economics lesson fromthe Microsoft case. Some antitrust issues raised by the Microsoft case are simple; others are more difficult. The case confirms why it is appropriate for the antitrust laws to use a Rule of Reason (i.e., weighing of benefits vs. costs) to assess whether exclusionary conduct that prevents (or impairs) rivals fromobtaining customers can be an antitrust violation. 2 Tie-in claims of bundling Internet Explorer, Microsofts browser, with Windows, Microsofts operating system, raise complicated issues; and how those claims are ultimately decided could have a huge effect on competition in high tech industries. Antitrust remedies are hard to fashion in a case like Microsoft. The goal is to deter future violations in this and other industries. One remedy is to impose a large financial penalty, a course of action that I understand the government is not authorized to pursue in its civil action against Microsoft. Another approach, the one taken in this case, is to pursue structural relief. The trick is to reduce market power at the operating sys- temlevel, the target of the government complaint, without creating huge losses by breaking up a suc- cessful firm. Basi c Fact s Although everyone is probably familiar with several of the facts in the Microsoft case, let me present some basic facts in order to illustrate the main economic insights. Because I amless interested in the exact details of the Microsoft case than in the general lessons that economists should draw fromthat case, I will consciously oversimpli- fy in order to focus on my main points. A reader interest- ed in only the Microsoft case can find plenty of details in the public filings, the Department of J ustice Websites, and the numerous conferences that have occurred. To function, a computer needs an operating systemthat handles the internal management of tasks such as file shar- ing and printing. Windows is Microsofts operating system. Application programs such as Word are written to work on a specific operating systemlike Windows. They do not generally work on other operating systems like IBMs 0S2 or Linux or Apples operating systemunless developers incur the costs of adapting the applications to these other operating systems. Figure 1 illustrates these facts by show- ing that, unless the application programhas the correct grooves, it will not fit into an operating system. The demand for an operating systemdepends on the number and quality of applications written for it. The more high-quality applications there are, the greater is the demand for the operating system. There will be more applications written for an operating systemwith a large installed base of existing users than for an operating sys- temwith a small base, because switching operating sys- tems in order to use a given application is costly. Hence, a new application will penetrate the user population faster if it is written to work on the operating systemwith the largest existing installed base. This linkage makes the computer industry a network industry and explains why the demand for an operating systemwith many applica- tions exceeds that for a rival operating systemwith few applications. The notion that network industries are somehow new and require a new understanding of economics is mis- leading. Network industries have long existed and been 2 Nothing is as simple as it sounds. Microsoft often did not demand exclusivity by the terms of its contract, but instead used various terms to attain such exclusivity. This in itself raises complicated issues that I have discussed elsewhere. See Carlton (2000) and footnote 4 below, as well as Paragraph 241 of Judge Jacksons Findings of Fact, U.S. v. Microsoft, Civil Action No. 98-1232, 11/5/99. Wi ndows-specfi c Appl i cat i on Wi ndows Wi ndows-specfi c Appl i cat i on Compet i ng Operat i ng Syst em W I N D OW S - S P E C I F I C A P P L I C A T I O N W O N T W O R K W I T H A N O T H E R O P E R A T I N G S Y S T E M F I G U R E 1 TheLessons fromMicrosoft Business Economics J anuary 2001 49 important. The railroad industry, which began almost 200 years ago, provides a good example; but there are many others. Indeed, the features of the Microsoft case that are of particular economic interest include the complementarity between two products, operating sys- tems and applications, and the economies of scale that exist in these products. This particular set of circum- stances is certainly not a new phenomenon. 3 The prin- ciples of microeconomics do not have to be repealed to understand these circumstances, just applied appro- priately. The functions of an operating system have increased over time. For example, the development of the graphical user interface (in which a user can picto- rially see and choose functions) illustrates an improve- ment in the usefulness of an operating system. There is no reason to believe that the operating systemwill not continue to evolve. A browser allows one to view Web pages. In the mid-1990s, the most popular browser was Netscape. The Netscape browser was a formof middleware. It sat on top of an operating systemand displayed Web pages that were downloaded off the Internet. The Netscape browser ran on all of the leading operating sys- tems because Netscape had incurred the costs of adapting its browser to each of these operating systems. The Netscape browser had its own application pro- gramming interfaces (APIs)the building blocks that software developers use to write application programs. Because Netscape had its own APIs (and the potential to expand those APIs) and because Netscape ran on all of the leading operating systems, Netscapes browser had the potential to become a substitute for Windows as a platform for application programs. Software developers might have written programs based on Netscape APIs instead of bas- ing their programs on the Windows APIs (the grooves in the interface between Windows and applications in Figure 1). If Netscape (and application programs) had developed in this manner, then the unique link between application programs and the operating systemwould have been bro- ken because application programs written to the Netscape browser would have worked on any operating system. Therefore, a large installed base for a particular operating systemwould no longer have provided an advantage for that operating systemover other operating systems when new applications are written. Figure 2 illustrates how a browser could theoretically fit into any operating system. In this schematic, the grooves on its bottomside enable it to work with each operating system, while the common grooves on the top provide that applications written to the browser will work on any operating systemon which the browser sits. J ava was (and, to some extent, still is) another substi- tute for Windows. J ava is part programming language and part middleware. The middleware portion of J ava is the J ava Virtual Machine that sits on top of an operating sys- tem. There is a J VM for each of the leading operating sys- tems. The J VM enables an operating systemto process application programs written in J ava and that use J ava APIs. Thus, like Netscape, J ava also had the potential to become a substitute for Windows as a platformfor appli- cations programs. Netscape facilitated the distribution of J VMs and the use of J ava. Every time a user put Netscape on his machine, a J VM was installed. Thus, Figure 2 also illustrates how J ava breaks the unique link between the operating systemand applica- tions for any middleware products (such as a browser) that fits on top of every operating system, yet provides a common top layer for applications. Rapid technological change and diffusion are familiar stories. Using one common measure of processing speed, chip speed has over the last twelve years increased by a factor of about sixty. The number of PCs shipped annual- ly has risen from19 million in 1990 to about 125 million in 2000. And, of course, the rise of the Internet has been phenomenal. Over the last decade, Microsofts share of operating systems on new personal computers based on 3 Whether scale directly alters the product characteristic (as in mak- ing it easier to interchange files with users of the same application programand operating system) or affects the price of a complementa- ry product, the economic effects in a dynamic model are similar. See Carlton and Waldman (2000). Browser/ Mi ddl eware Appl i cat i on Appl i cat i on Browser/ Mi ddl eware Wi ndows Compet i ng Operat i ng Syst em C O M P E T I N G O S A T N O D I S A D V A N T A G E : B R OW S E R / M I D D L E W A R E P L A T F O R M C A N W O R K W I T H A N O T H E R O S F I G U R E 2 50 Business Economics J anuary 2001 TheLessons fromMicrosoft Intel-type designed chips has been over ninety percent and is growing. It is in excess of eighty percent (Apple is about five to ten percent) when one considers all PCs. What Mi crosoft Di d: Lessons for Busi ness St rat egy Microsofts Conduct I now want to use these simplified facts to focus on the economic lessons fromMicrosofts behavior, focusing on two aspects. I emphasize that I amnot discussing the legality of Microsofts actions. The first aspect of their behavior relates to exclusive dealing. Exclusive dealing occurs when a firmsays to its customers or distributors that they are not allowed to deal with a rivals product. So, for example, if Microsoft said to a distributor (e.g., an original equipment manufacturer [OEM] that makes computers and sells themwith an oper- ating systeminstalled) that it could distribute computers only with Windows, it would be engaging in exclusive dealing. As I understand the facts, Microsoft did not typi- cally engage in such explicit exclusive dealing but did offer pricing structures and other terms to encourage exclusivity. For example, in conduct attacked by the Department of J ustice in 1995, Microsoft would charge OEMs a fee for Windows based on all computers shipped whether or not Windows was used in those computers. This means that the marginal cost of Windows was zero, creating an incentive for an OEM to exclusively use Windows. 4 As another example, Microsoft arranged an extra dis- count that Compaq (an OEM of computers) would receive provided Compaq shipped its computer with Microsofts Internet Explorer browser and promoted only IE and not Netscape. Other OEMs also received discounts on Windows provided they favored IE, not Netscape. There can be good pro-competitive reasons to engage in exclusive-dealing type arrangements. It is well known that they can protect property rights and thereby encour- age promotional effort. (See, e.g., Carlton and Perloff [1999], Chapter 12.) So, for example, a manufacturer of cameras might demand exclusivity of camera shops and prevent shops selling its camera fromcarrying competing brands so that, when the manufacturer engages in promo- tion or in dealer training, or provides customer leads to stores, other rival manufacturers of cameras will not be able to free ride on those efforts. But, it is also known that exclusive dealing can harmcompetition by making it more difficult for rivals to obtain distribution, especially when distribution is handled by only a handful of distrib- utors. (If there are many distributors and there are no scale economies in distribution, exclusive dealing cant usually harmcompetition.) Here, Microsofts conduct made it more difficult for a competing operating systemto Windows to obtain distribution, for Netscape (which might have been a rival to the Windows operating system) to obtain distribution, and for J VMs to be installed on com- puters (since Netscape was an important distribution vehicle for J VMs). This use of exclusive-type arrange- ments in these ways can be viewed as a pretty standard application of foreclosure in which the foreclosure pre- vents rivals fromcompeting effectively against Microsofts operating system. The second aspect of Microsofts behavior that I want to focus on is its so-called embrace and extend strategy. This is really the most interesting part of their economic behavior. In this strategy, Microsoft recognizes the value of the browser but is concerned about the spread of a browser such as Netscape that might become a platform for applications written for any operating system. Microsoft therefore created IE, which it either physically bundles with Windows or includes it at no extra charge and which does not obviate the need for Windows-specif- ic APIs for applications programs written for Windows. Much of the governments case at the antitrust trial was to establish that there were no efficiency reasons to bundle IE and Windows. Therefore, the consequence of Microsofts strategy was to reduce the demand for Netscapes browser even though the IE browser was not superior. In this way, Microsoft hindered Netscape from developing into an alternative platformto Windows, and applications programmers writing a programto run on Windows still had to use Windows-specific APIs, preserv- ing the link between the applications program and Windows (see Figure 3). Microsofts strategy with respect to J ava was similar. Microsoft recognized the advantages of J ava as a pro- gramming language and therefore became a J ava licensee. Because Microsoft is a J ava licensee, the IE browser, like Netscapes, comes bundled with a J VM. But, after embracing J ava, Microsoft then extended it. For example, it created tool kits for J ava software developers that made it easy for themto mix Windows APIs with J ava APIs. 5 Figure 3 illustrates this Windows-specific exten- sion of J ava. The consequence of Microsofts actions is that many application programs written with Microsofts 4 Deciding when exclusivity effectively achieved through pricing (or other means) should be treated as the equivalent of explicit exclusivity for legal purposes is a tricky issue. See Carlton (2000) and Thomet al. (2000). I do not want to analyze here when contractual restrictions turn into exclusionary conduct, a topic analyzed in the just cited papers. For purposes of this paper, assume that Microsofts actions can be charac- terized as equivalent to exclusive dealing. 5 Microsofts behavior regarding Java is the subject of a breach of con- tract suit by Sun against Microsoft. TheLessons fromMicrosoft Business Economics J anuary 2001 51 J ava will not run on other operating systems, and hence Microsoft again preserves its application advantage. Consequences of Microsofts Actions Microsofts actions toward Netscape and J ava each accomplish the same goal. They prevent the application advantage of Windows fromslipping away and thereby prevent the source of Microsofts operating systemmonop- oly fromeroding. Importantly, Microsofts actions show how a dominant firmcan use a variety of conducts to pre- serve and maintain its monopoly position in a rapidly changing environment. That is a new lesson from Microsoft, though it is probably one that some of the IBM antitrust cases should have suggested. 6 By preventing new applications frombeing simulta- neously usable on both Windows and other operating sys- tems, Microsoft preserves its advantage. But it also means that, as the operating systems evolve, Microsoft will con- tinue to have an advantage over competing operating sys- tems as long as its operating systems use exisiting appli- cations. Perhaps more importantly, even if operating sys- tems for PCs vanish, Microsoft is well positioned to become the monopolist of any new systemas long as that new systemcan utilize the stock of Windows-based appli- cations. So, for example, if handheld devices replace PCs, then Microsoft could monopolize operating systems for themif such systems can use existing Windows- based applications. In this way, Microsoft can swing its monopoly over time fromproduct to prod- uct by using its application advantage at each point in time to advantage its own products. This behav- ior explains how one firmwith a head start can use tie-in sales and other behavior to prevent generic complementary products from being developed. Moreover, at each stage it can use its special link to the complementary product to become the monopolist of the primary product, even when the primary product is changing rapidly over time. Indeed, it is the combination of some scale economies (or network economies) and rapid tech- nological change that produces the circumstances under which a dominant firmcan remain dominant over time in the face of rapid technological change. 7 Figure 4 illustrates these points. Initially, there is a monopolist of A (operating system), which uti- lizes complementary products B (applications that work only with A) and B (applications that work with A or a future rival A). By preventing B fromsuc- ceeding, the competition between A and a rival A in later periods is avoided, and the initial monopolist of A remains the monopolist in later periods. An important point is that product A in the later period could be a dif- ferent product than A in the initial period, as long as A in the initial period and A in the later period both utilize B as a complementary product. Lessons for Ant i t rust Let me now turn to the economic issues related to the antitrust challenge of Microsofts behavior. First, it is per- fectly legal to be a monopolist. U.S. antitrust laws (and patent laws) recognize that firms that innovate will have market power and that the associated profits provide the incentive to firms to develop new products that consumers want. And our antitrust laws recognize that tough compe- tition helps consumers but harms rivals. J ust because a rival complains does not mean that a firmis doing some- thing wrong. Our antitrust laws do, however, prevent the creation or maintenance of market dominance through certain types of conduct deemed bad acts. Although the Microsoft case raises some hard ques- tions about antitrust law, some of the issues are relatively routine. 8 This is the case for the claims regarding exclu- sionary conduct aimed at preventing OEMs fromdistrib- 6 The IBM cases (e.g., the private cases as well as the government case) often focused on peripherals. Once one realizes that intelligence and processing capability can reside in peripherals, some of the issues in the IBM cases become similar to those in the Microsoft cases. 7 If the technological change is so rapid that the value of the comple- mentary product vanishes, then the advantage of the incumbent is eroded. 8 With the important caveat of footnotes 2 and 4. Many Wi ndows-speci fi c Appl i cat i ons Mi ddl eware Wi ndows Few Appl i cat i ons Compet i ng OS W I N D OW S A D V A N T A G E R E S T O R E D U S I N G E M B R A C E - A N D E X T E N D S T R A T E G Y F I G U R E 3 52 Business Economics J anuary 2001 TheLessons fromMicrosoft uting Netscape and J ava (or competing operating systems fromobtaining distribution). The relevant issue is whether the harmto competition fromexclusion exceeds the effi- ciency benefits presumably associated with the exclusion. This is a standard type of claimin which the dominance of Microsoft would play a key role. In the 1990s, Microsofts share of the operating systemmarket has been over ninety percent and growing. In such a setting, an eco- nomic analysis under the Rule of Reason would require that the efficiency benefits fromexclusion would be very large in order to offset the possible harmto competition. The claims regarding the embrace and extend con- duct aimed at Netscape and J ava are more complicated to evaluate. Let me focus solely on the tie-in claimregarding IE. Tie-in law is muddled froman economic viewpoint. The standard tie-in case is that a monopolist of product A requires that you also buy product B. Why? Why not just charge more for A? The case law provides feeble answers. The most common explanation is that such a tie enables price discrimination to occurwhich has ambiguous con- sequences on social welfare. The bottomline is that there is little economic justification for traditional antitrust hos- tility toward tie-ins. (This insight, due to Aaron Director and Edward Levi, still applies today [with some modifica- tions]. 9 ) But the tie here creates true competitive harm, by dis- advantaging rival operating systems by preserving the Microsoft advantage in applications. So the economic underpinnings of tie-in law can finally be made sensible once one understands the rationale for Microsofts behav- ior. This should be one of the key economic insights of the Microsoft case applied to antitrust. A difficult issue is whether product design should be subject to antitrust attack. Recall that Microsoft was chal- lenged for making the code of Windows and IE insepara- ble (a physical tie-in). There are some plausible benefits of expanding the functions of the operating system. Do we want courts weighing the efficiency benefits vs. the com- petitive harmissue, as a general matter? Do we want the government telling Ford how to design a car in terms of what components have to be removable? The costs of impeding technological change could be enormous. If one goes down this route, one should give significant weight to efficiencies, which are likely to be too difficult for courts to evaluate. 10 One unique feature of the Microsoft case may be the unusual amount of internal e-mails that could allow one to figure out the real reason that IE was tied and what the costs of untying it would be. But it would be foolish to believe that such e-mails will exist in the next antitrust case. So, if the tie-in claimagainst Microsoft stands, I would hope that it would be interpreted as a reflection of the rather unique cir- cumstances of this case and not as a general attack on tie- in sales achieved through product design. Remedy What remedy makes sense in this complicated case? The purpose of a remedy is to deter future bad acts in this and other industries and to restore competition. There are at least three possible remediesa fine, conduct restric- tions, and structural relief. First, a large fine (e.g., many billions) would be appro- priate so that the threat of future large fines should serve as a deterrent to future bad conduct in both this and other industries. The beauty of a fine is that one does not have to get involved in figuring out the costs of structural relief in which a successful firmis torn apart. The fine should be large enough to make this firm(and other potential vio- lators) take notice and must specify the types of undesir- able conduct with enough specificity so that the firm knows what to avoid in the future. The major difficulty is that the government (though not private plaintiffs) has no legal basis to request monetary damages in a civil antitrust suit. This represents a serious impediment for the government to fashion an effective remedy quickly, 9 See Carlton (2000) for a discussion of tie-in doctrine and, in particu- lar, the discussion of Whinston (1990). 10 There is a distinction between a physical tie and a contractual tie. In a physical tie, the product comes bundled so that a requirement that the components of the bundle be sold separately requires analyzing howproduction costs within the firmwill be affected when the physi- cal nature of the product is altered. In contrast, a contractual tie requires that a user consume, say, two separate components fromone firm. A requirement that the components cannot be tied will not entail a change in the physical nature of each of the components produced. Courts have been reluctant to intervene within a firmand tell firms how to operate. In contrast, courts have frequently intervened in market transactions achieved through contract. So, for example, it would be unusual for a court to order GM to allowFord to use its spare produc- tion capacity, but would not be unusual for courts to prevent GM from contractually monopolizing independent dealers (if it had the ability to do so) to prevent distribution of Ford cars. P R E S E R V I N G T H E M A R K E T P OW E R I N D Y N A M I C I N D U S T R Y F I G U R E 4 Now Lat er PRODUCT A B PRODUCT A B B' A' B' If the presence of B' will erode As power later, then attack B' now. Product A in later period could even be a different product than A now. TheLessons fromMicrosoft Business Economics J anuary 2001 53 and this impediment could be removed only through leg- islation. The second possible remedy involves conduct restric- tions. Basically, Microsoft is told: Dont do it again. Dont engage in conduct that places significant restrictions on third parties (e.g., OEMs) fromdealing with your potential rivals. There are two difficulties with conduct remedies. First, they are hard to enforce because the third parties may be reluctant to complain about Microsoft if their suc- cess depends on Microsoft. Because of the constant ongo- ing assistance that Microsoft must provide an OEM, some OEMs could be fearful to complain about Microsofts con- duct to enforcement officials. Second, the conduct that will work tomorrow in harming rivals may be difficult to predict today. Hence, specific conduct remedies may turn out to be obsolete quickly. The third remedy, the structural dismembering of Microsoft, is the most radical. Any structural remedy imposes costs by destroying the internal knowledge that a successful firmhas developed that has allowed its con- stituent parts to work together successfully. The advantage is that, unlike a conduct remedy that requires enforce- ment, a structural remedy relies on the forces of competi- tion to remedy the complained-about behavior. There are a variety of possible structural remedies that could restore competition in operating systems, which was the focus of the governments case. One is to create several operating companies. The J ustice Department specifically chose not to do this, fearing that the costs of doing so were too high with one of the costs being different versions of Windows. Instead, the J ustice Department chose to create one oper- ating systemcompany and one applications company. The idea is that soon, there would be an incentive for the operating company to develop its own applications and for the applications company to develop its own operating systems. That remains to be seen. Concl usi on The Microsoft case illustrates how an initially dominant firmcan use contractu- al and physical ties to remain a dominant firmin a rapidly changing industry. It rais- es some antitrust issues that defy easy solutions. One can perhaps argue that the Microsoft violations were so egregious and the reported e-mail evidence seems overwhelmingthat a dramatic remedy is needed and justified. Maybe so. However, the harmfrom a general interventionist antitrust policy could dwarf any benefits. If the drastic remedy of structural relief is imposed in this case, it should best be regarded as an exceptional remedy for an exceptional case. I A C K N OW L E D G E M E N T S I thank Gregory Pelnar and Robert Stillman for help- ful comments. R E F E R E N C E S Carlton, Dennis W. and J effrey Perloff. 1999. Modern Industrial Organization. Addison, Wesley, Longman, third edition, Ch 12. Carlton, Dennis W. 2000 forthcoming. A General Analysis of Refusal to DealWhy Aspen and Kodak are Misguided. Antitrust LawJ ournal. Carlton, Dennis W. and Michael Waldman. 1998, revised March 2000. The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries. Working Paper No. 145. George J . Stigler Center for the Study of the Economy and the State, University of Chicago. Thom, Willard, David Balto, and Neil Averitt. 2000. Anticompetitive Aspects of Market-Share Discounts and Other Incentives to Exclusive Dealing. Antitrust LawJ ournal. 67, pp. 615- 639. Whinston, Michael. 1990. Tying, Foreclosure and Exclusion. American Economic Review. 80, pp. 837-859. So the economic underpinnings of tie-in law can finally be made sensible once one understands the rationale for Microsofts behavior. This should be one of the key economic insights of the Microsoft case applied to antitrust. 54 Business Economics J anuary 2001 Tomorrows Electric Distribution Companies T he issues surrounding electric industry restructuring, merger and acquisition activi- ties, changing regulations, and ever-acceler- ating technology enhancements are daunting to electric utility distribution companies and poorly understood by their customers. The regulatory safety net providing a guaranteed double-digit earnings streamis not what it once was. Even the apparently safe option of a pure poles and wires strategy is potentially threatened by load (sales) losses fromdistributed genera- tion resources (such as large customers generating their own power), energy conservation measures and services, and distribution bypass (sales direct fromgenerators to ultimate customers). The electricity industry value chain depicted in Figure 1 illustrates the vital position of distribution sys- tems and services in the provision of electricity. While todays customers refer to their utility when they think about electricity, their attitudes about the company are largely determined by their interaction with the distribu- tion part of the company. Metering, billing, customer service, outages, and power quality are all functions that currently reside in distribution services. Tomorrows Electric Distribution Companies CRUCIAL STRATEGIC DECISIONS WILL AFFECT THE ENTIRE U.S. ECONOMY. By Ahmad Faruqui and Ken Seiden Ahmad Faruqui is Area Manager, Retail & Power Markets, at EPRI (theElectric Power Research Institute). Dr. Faruqui has held senior consulting positions with Barakat & Chamberlin, EDS, Hagler Bailly, and Battelle. He has BA and MA degrees in eco- nomics fromtheUniversity of Karachi and a Ph.D. in econom- ics fromtheUniversity of California, Davis. Ken Seiden is VicePresident of Quantec, a con- sulting firmdedicated to quanti- tativeeconomic research for the utility industry. Dr. Seiden has worked for Essential Economics, Barakat & Chamberlin, Pacific Bell Telephone, and BonnevillePower. Heholds a Ph.D. Economics fromthe University of Oregon. Theelectric industry has been dominated by vertically integrated utilities for most of its history. Theseutilities arebeing restructured as theelectric industry is being deregulated and moves toward competition. Someverti- cally integrated utilities havesold off their generation assets and arebecoming distribution companies that deliver electricity. Someobservers contend that such companies will simply providephysical delivery service, that is, they will becomepoles and wires companies that will performa few basic functions, such as line maintenanceand treetrimming. This articleshows that such a de minimusview of thefutureelectric distribution company is myopic. Several alternativefutures await the creativedistribution company of tomorrow. Thekey to success is choosing that future, which best matches the companys corecompetencies. Sinceelectricity supply affects every aspect of theU.S. economy, thesechoices are likely to haveramifications far beyond theutilities and should beunderstood by all businesses for which electric- ity is important in their operations and planning.