The Microsoft Campus Agreement Andr´as Breuer April 16, 2005 Abstract This paper will attempt the analysis and critical review of an expected deal between Microsoft and the Ministry of Informatics, and Communication of the Republic of Hungary regarding special licensing of software to teachers and students. We will (1) introduce the deal, and its relevance, (2) analyze the economic background of such a deal, (3) compare this solution with an alternative free software solution by reviewing relevant empirical findings, and (4) propose a recommendation.

1

1 Introduction On November 30, 2001, Microsoft Hungary, and the Office of the Prime Minister signed a deal worth HUF 1 billion (more than EUR 4 million or SEK 37 million) for each of the following three years (HUF 3 bn in total[1]) for licensing Microsoft software to students, and teachers working in higher education. “According to the agreement between the Office of the Prime Minister, and Microsoft Hungary, all full-time, correspondence, and night-time students, in addition to all teachers, and employees at every state-operated, religious or private higher education institution may use the specified software for educational or research-and-developmental purposes at the institute or — in the case of students, and teachers — also at home. The Microsoft Campus Agreement allows the use of the most recent version — or an older version should the user so desire — of the following pieces of software in any available language:”[2] • Windows XP Professional • Office XP Professional • Front Page 2002 • Visual Studio • BackOffice client access Since then, responsibility for this agreement has been transferred from the Office of the Prime Minister1 to the Ministry of Education,2 and later to the Ministry of Informatics and Communication (MIC)3 . The deal was struck without any use of public procurement methods — such as tendering or auctioning. According to the information available at the www.campus.hu homepage, the persons entitled to use the software specified above are eligible to use it free of charge. However, if they leave the field of higher education, or if the agreement is not renewed after three years, their license is void. On January 16, 2002, the following text was added to the homepage: “According to the special agreement between the Hungarian government, and Microsoft Hungary, regarding PCs acquired before the signing of the agreement (November 30, 2001), the software that is upgraded is considered authentic[. . . ]”[2, 3] In other words, illegal versions of the software were ex post facto legalized, as long as they were used in computers purchased before the given date. The apparent underlying problem with this addition is that instead of outright dealing with the use of illegal software, the agreement only provides an upgrade style license for the operating system (OS). This only provides for confusion yet again regarding what is, and what is not lawful use. In November 2004, the three years of the agreement were over, and negotiations started between the MIC, and Microsoft Hungary about the possible renewal of the deal. For this reason, the Campus licenses were extended until February 28, 2005. We 1 The home page of the Office of the PM in English is available at: http://www.meh.hu/english All Internet URLs cited in this paper were accessed between March 04, 2005, and April 16, 2005. They are unlikely to change in the near future. 2 English home page: http://www.om.hu/main.php?folderID=137 3 The home page of the MIC in English is available at: http://en.ihm.gov.hu/

are now past that date, but there is still no sign of a renewal agreement. According to the MIC, however, during the negotiation period of the 30 days after February 28, the licenses “are to be considered valid”.[4] While it may seem only natural to renew this contract, if only for simplicity’s sake, it should be contemplated more thoroughly — especially if we take into consideration the booming growth of some alternatives in recent years.

2 Economic Analysis An alleged motivation for this deal is to provide a means for legal software use in the field of higher education. While this is an important, and desirable aim, it is vital to study the background, reasons, and potential effects of any government policy, especially if it is in the field of procurement. In addition, as transparency is considered to have a number of beneficial effects on public procurement, available information stating the reasons, and aims of the agreement would be but highly expected. Unfortunately, there is no sign whatsoever that such an analysis has been performed at all in regards to the deal, let alone having been made public. From an economic point of view certain questions arise concerning the Campus Agreement: • Does the Hungarian higher education need a ministry-sponsored software licensing approach? • Which are the desirable methods of software procurement? • What software should be included in such a deal? • What are the benefits, and costs of the Campus Agreement?

2.1

Does the Hungarian higher education need a ministry-sponsored software licensing approach?

Two arguments come to mind which provide possible reasons for ministry-sponsored software licensing. The advantages of such a licensing policy may come from the fact that software is actually a public good, and thus the socially optimal level of production may not be guaranteed without government intervention. In other words, the effects such an action may have on the level of social welfare should be taken into consideration. Another reason may be the need to cut down on the costs of public services — such as education —, which could possibly be achieved by using the bigger bargaining power of the ministry representing all higher education institutions. A possible drawback of the deal may be the loss of efficiency due to the inability to set the level of production at the socially optimal level. In fact, any output farther away from the social optimum than the free market level is undesirable. 2.1.1

Social welfare

Is software a public good? In order to answer that question a definition of public goods has to be provided. There are two typical characteristics of public goods[5]: they are 2

non-rivalrous, and non-excludable. OS software seems to fulfill both requirements in many cases. Non-rivalry is actually overly fulfilled: the more users use a given operating system, the bigger the beneficial network effects are.[6] This means that not only does an additional user not lessen the utility of all other users, it actually raises the utility of others. This can be caused by the fact that a piece of software may be improved upon more, if it is used by many, and this possible future improvement has a value for current users as well. The improvement might come from the original producer as a profitable activity, or from the users — possibly in the now-so-popular wiki style, which is based on mutual contributions. Moreover, a bigger pool of complimentary software is likely to be produced for a bigger user-base. Viewed from a different perspective, however, this argument also means that independent developers of complimentary applications may want to develop cross-platform applications to reach more customers.4 These applications, therefore, do not add to the value of merely one OS, instead, they benefit all OSes they are available for. Furthermore, if many people use the same program, it may behave like a standard. This makes communication about this given piece of software easier, as most can be presumed to know what one is referring to. It also reduces costs of in-house support if only one platform needs to be supported. In addition, users only have to learn one interface, and thus the costs of learning additional interfaces can be avoided. Non-excludability is a different issue, however. Even though in most cases the transfer of a given piece of software is a simple act of copying it from one computer to another, the license agreements of many proprietary systems (including those of Microsoft Windows) legally prohibit such an action. In the case of free software, however, such copying is allowed. Yet exclusion is neither trivial, nor easy, and far from natural. In the case of a typical (OEM) Microsoft Windows license, for example, the obligatory activation of the product is only possible through the phone — since a recent policy change.[7] Throughout most of the 90’s, the entering of a product key sufficed, but with the introduction of Windows XP, Microsoft also introduced the notion of product activation. Until now, this activation was always possible through the Internet — given an Internet connection at the point of installation. The above changes in copy-protection technology show how difficult it can be to prevent unauthorized use of a piece of software. They also show how easy it must have been in most cases to circumvent copy protection to necessitate such a rapid change of copy-protection technology. In summary, software can often be looked upon as a public good — especially free software5 , where everyone has the freedom to redistribute the software[8]. Proprietary software, however, can be considered to be only a club good (non-rivalrous but excludable). Some argue that due to the problems with successful exclusion all kinds of software are typical public goods.[9] If we consider software to be a club good, profitable production should be possible, but the socially optimal level of output is still not guaranteed — and a deadweight loss may be incurred. 4 There are many examples of cross-platform applications, including Microsoft Office, which is available for Windows and Mac OS X, and most programs that are to be mentioned in the section dealing with possible alternatives to the Microsoft solution. 5 The terms free software and open source software are going to be used interchangeably in this paper. Free software often refers to the self-distinction of those, who consider the issues of political values, and freedom the most important; as opposed to its mere economic significance.

3

The question “What is the socially optimal level of output for software?” seems to be unavoidable at this point. According to microeconomic theory, the optimal quantity of production of any good is at the point where the marginal cost curve of production meets the demand curve. (External effects of software is not taken into consideration in this paper.) Software is a very special product. Once it is written, the marginal cost of producing any number of additional units of it is virtually nonexistent. Depending on the amount of the needed investment, a natural monopoly may be the most efficient way of production. Even more so, if network effects are also taken into consideration. The more people use a given piece of software the bigger its utility is to the users. A considerable investment may be needed to produce the first unit, but only the cost of the medium — that of a blank CD, DVD, or available bandwidth, for example — is incurred in producing any additional units. The production of software, therefore, resembles that of intellectual property the most closely. Regarding the demand for software, it can be seen as a durable good.[6] Overall, once a software product is ready, the socially optimal price for that product should be virtually zero, and it should be available to virtually everyone who wants to use it. In order to make the production of software economically feasible, a means to recover the initial investment might be needed. A desirable amount of production may be endangered by possible under-production due to either the free rider effect, or, in the case of club goods, exclusion. Traditionally there are two solutions to this problem: (1) let the market take care of production, or (2) produce it publicly. Both solutions include high possibility of deadweight costs: (1) because of over-exclusion, and (2) because of the deadweight costs of an associated tax raise. It can be argued that the distribution of these costs is more fair if private firms — and/or wealthy individuals — acquire the good for a fee, while others (possibly including public institutions) are subsidized. Another important aspect may be that the associated benefit of network effects — the higher demand for the product —, when a piece of software is used and taught in schools, may well cover the proportional costs of its production. Furthermore, monetary incentives are not always the needed form of investment for the production of software, neither are markets or firms necessarily the most efficient in the field of information technology. As is the case with free/open source software, which is coordinated by an altogether different process. The theory of commons-based peer production[10] for example suggests that an alternative model of production, or in other words, an alternative technology, may be more efficient. While physical capital is in fact needed for recording and communicating in the domain of information, the costs of this capital have reduced drastically over recent decades. This suggests that “the primary remaining scarce resource is human creativity.”[10] In this alternative model of software production, it is possible after all, that there is no need for a big monetary investment for developing new software. To put things in a more dynamic perspective, distinct models of software production — the retail, the contract or in-house, and the open source model — will be examined more thoroughly here. The retail model treats software as any other manufactured good, and it is based on two fundamental premises: (1) that most developer time is paid for by sale value, and that (2) the sale value of software is proportional to its development costs and its use value.[11] This model supports a producer driven development process. In reality, such development (“prepackaged software”) only accounts for less 4

than 25% of all commercial software development in the USA.[12] Since this is the country that, among other widespread packaged software products, exports Microsoft Windows, the operating system used by an overwhelming majority all over the world, the numbers for a country like Hungary, can be suspected to be considerably lower. This paradigm is only useful in the case of software products that have a mass market, and has considerably lower efficiency than the open source model.[13] In-house and contract development may be superior in the case of differentiating technology — something that makes one business more desirable to customers than a competitor. Software that is used by every actor in practically every market, such as the operating system, and office applications clearly falls in the non-differentiating category. In this case, open source production is favorable to both in-house or contract development and retail production. The in-house or contract model and open source model are user-driven development methods. What are the effects of free software on social welfare? — one might ask. While open source projects are more efficient in regards to the aggregation of information, and their access to the entire pool of talent, not only that of a given firm, “due to the free riding problem, some potentially valuable projects will not be developed under an open source system”[14]. The proliferation of alternative operating systems causes inefficiency because of the inherent network effects. The more people use a given system, the better the overall knowledge about that system is. Moreover, the value of any OS grows with the amount of useful complementary software available for that OS. Given more than one competing operating systems, a duplication of efforts, and thus a waste of resources seems likely. It is, however, not inevitable — a number of applications are developed in a multiplatform way, as mentioned before. On the other hand, competition is known to spur innovation, and drive prices down, while a monopoly in a market usually leads to deadweight losses. Multiplatform applications lessen the possibility and the effects of vendor lock-in, a phenomenon often exploited by monopolies. Another danger is if competition leads to deteriorating profits, the decline of investment into innovation, and thus insufficient monetary incentives could curtail major breakthroughs. It is hard to notice any trend of major breakthroughs in the field of software, however, be it free or proprietary. Additionally, free, and proprietary software may not have the same utility to users — one may meet their needs better. It is often argued that free software is not as user friendly as its proprietary counterparts. On the other hand, given free access to the source code, and the freedom of modification, it can be tailored to the specific needs of any one user. This is usually not possible with a proprietary program. 2.1.2

Ministry bargaining

The ministry is able to achive a considerably stronger bargaining position by aggregating the needs of a significant number of consumers, than those, whom it represents, are — the teachers, students, and employees. The Ministry of Education is in a perfect situation to represent the aggregate demand of higher education, has a good point in wanting to cut costs of software licenses in higher education, and the Campus Agreement is a promising way of doing exactly that. If we take into consideration the number

5

of students, each campus license costs only about USD 11 per year per student 6 . In comparison, an Windows XP Home Edition OEM license alone — which allows the use of the software on the computer it is purchased with — costs HUF 22 856 (or USD 124, or EUR 94) per computer. The whole package of software included in the deal, however, is worth more than USD 700, even if we take into consideration any possible discounts.[15] Large-scale bargaining with Microsoft is often successful. One of many examples is when the city of Munich was considering switching from Microsoft products to an open source solution, it was offered a 90% discount.[16] Another valid reason for the deal is the fact that many, if not most software in use before the deal in the field of higher education is likely to have been illegal — it was used without a proper license. Given the financial hardships most education institutions are dealing with in Hungary, they were not likely to have been able to purchase any licenses given their original prices. The ministry/ministries involved with the deal, however, did make mistakes, as there is no sign that they did any form of market research precluding the deal. They purchased a software package that has no proven connection with the needs of the users. Most teachers, and students, for example, are highly unlikely to need the development package Visual Studio. Moreover, many students might have had licensed Microsoft software already. Buying a smaller amount of software licenses may have been cheaper. A lower price may also have been achievable given a threat to move away from Microsoft software — as seen in the Munich case.

2.2

Which are the desirable methods of software procurement?

There are two important aspects that are to be examined: public procurement methods in general, and the special characteristics of the software market. This part is based on Laffont-Tirole’s Competition in Telecommunications[17] Assuming a situation where the government wants to procure some public good that is a natural monopoly, the most important question is the power of the incentive scheme. The reason why performance-based or incentive regulation is needed is the need to increase the efficiency of production. A high-powered incentive scheme is likely to achieve higher efficiency, as in this case the firm bears a high fraction of its cost at the margin. An example of this scheme is a fixed-priced contract, which was in fact used in the Campus Agreement. A high incentive scheme, however is vulnerable to rent extraction — the firms profits are only partly due to its cost abatement measures; they are also due to exogenous variables, or luck. As the government wants to pay as little as possible for the procured good, rent extraction can be seen as something negative. The government is likely to have one or more of three possible reasons for this. First, it may put a lower weight on the firm than on other agents in the economy. Second, it could also be that transfers to the firm must be financed through distortionary taxation, and thus unwarranted transfers create a social loss. Third, should the firm be subject to budget balance, its rent is financed through markups on the firm’s services, which distort the relative consumption of consumers, and are thus undesirable. (This 6 According to Zolt´ an Fauszt, CEO of Rufusz Computer Kft., the company responsible for the technical details of the deal

6

third point is irrelevant in the case of the Campus Agreement, as the procured services are supplied free of charge — or a nominal fee of the medium —, which seems to be socially optimal.) In case there is asymmetric information between the firm, and the government regarding the costs of production, screening the firm through its contractual choice becomes the key issue. Overall, it is optimal to design a menu of contracts in which the firm self-selects: a high-powered incentive scheme when it is efficient, and lower-powered incentive scheme when it is less efficient. If for example firms compete in an auction, they are likely to report their type or real costs because of fear of losing the market. Benchmarking, that is, comparing the firm’s performance with that of other firms in a correlated situation or technological conditions, reduces the asymmetry of information, and also enables the use of higher-powered incentive schemes. There are three factors limiting the power of incentives. First, in high-powered schemes concerns about quality may arise. This is because the easiest way to reduce costs may be to reduce quality. In contrast, low-powered schemes do not discourage the production of high quality goods. Powerful incentive schemes, and close monitoring of quality standards are clearly complementary. Second, if the dynamics of incentive contracts, and regulatory commitment are taken into consideration, the problem of the ratchet effect arises. If the goods are produced in more than one period, and the contract is renegotiated after each period, a formally high-powered incentive scheme may make a firm hide its good type because of fear of punishment in subsequent periods. If there is enough incentive for the firm to show its type in the first period, a bad type firm may adept a “take the money and run” approach by mimicking a good type in the first period, and quitting the relationship in the second period.[18] The ratchet effect is reinforced by the possibility of renegotiation before the end of the contract. Overall, contract renegotiation ex post rewards the firm’s inefficiency by punishing its cost saving efforts. Third, the possibility of regulatory capture is also imminent in the case of high-powered schemes. In this case a second principal-agent problem arises from the regulator’s role as an informational intermediary between the firm, and the general public. High-powered incentives generate a large benefit for the firms of capturing their regulators, as the assessment of the regulated firms’ cost, and demand environment may fall under regulatory discretion. Two policies are to be mentioned that can potentially curb regulatory capture. A reduction of regulatory discretion reduces the the use of the regulators’ private information by creating a more bureaucratic environment. A reduction of the asymmetry of information between regulators, and their principal may also help. This can be interpreted as transparency, which helps reduce the informational asymmetry between the regulators, and the general public. Transparency is also very important in the field of procurement for a number of other reasons. First, it spurs enhanced efficiency, and increased innovation through enhanced competition. Second, transparency may lead to better value-for-money through more effective competition, which would reduce the level of bids, and thus associated public expenditure as well. Third, transparency may help encourage local, and foreign investment, and partnerships. Fourth, transparency helps reduce the level of corruption. Finally, ensuring transparency is a core element of good governance, which is essential for economic development.[19] It is time to explain why the above analysis is fitting for the software market, and what characteristics of this market need special attention. As it has already been men7

tioned, software is a special good that is characterized by network effects. For this reason it can be argued to be a natural monopoly. It has also been briefly mentioned why the government might want to procure this good. Regarding efficiency, the issue of technology is of utmost importance. Two alternative approaches seem to exist regarding the production of software. The proprietary approach, typically represented by firms, and the free/open source approach, typically represented by individuals working in their free time. The cost implications of the two are strikingly different, as the first type needs a hefty investment while the second type typically has considerably lower costs. In this regard the open source method of production is clearly superior technology. The above economic analysis, however, only relates to firms as producers of the public goods. There are a number of firms producing, and distributing open source software, such as Red Hat or Novell, in which case the production model is a hybrid one including both the classical business model, and peer production. This open source method of software production results in better quality, higher reliability, lower costs, and increased choice.[20] All of these factors are important from an efficiency viewpoint. The reasons for these characteristics are not to be described here in detail, but the most important issues seem to be extensive peer review, and participation. Ample explanation can be found in Eric S. Raymond’s essay The Cathedral and the Bazaar.[21] Apart from acquiring a right to use a piece of software, however, there are a number of other factors that also have an impact on the total cost of ownership. The costs of installation, support, training, maintenance, upgrades, and the principles of reliability, security, functionality, ease of use, interoperability, training, and availability also need to be taken into consideration. From this broader picture, software licenses only constitute a small facet of the costs. Most of these factors, however, may favor the open source production model, due to the fact that it implies greater competition in the fields supplementary to software. The present Campus Agreement is a fixed-cost contract, which is a high-powered incentive scheme. It is thus a high target for possible rent extraction. This is especially undesirable in this case, as the weight put on producer surplus (or the firm) is bound to be much lower than the weight put on consumers, for Microsoft is a foreign company, thus its profits don’t count towards the Hungarian social welfare. Auctioning or benchmarking might reduce the informational asymmetry between Microsoft, and the Hungarian government, but these methods were not used, for the reason that the programs included in the deal are “unique”, and don’t have alternatives[15]. The issue of alternatives will be considered later, but the fact that an OpenOffice based alternative was chosen over Microsoft Office when procuring office applications in another deal, indicates otherwise.[22] There are three factors limiting the effectiveness of highpowered incentive schemes, namely concerns over quality, the effects of renegotiation, and possible regulatory capture. The imperfect quality of Microsoft software is often referred to in online publications. 7 As Microsoft’s predominant licensing policy is one time sale of a licence which allows the use of the software for as long as the consumer wishes, it is hard to sell newer versions of the same software, if the previous version 7 “[. . . ] if Microsoft sells goods that are aesthetically unappealing, or that don’t work very well, [. . . ] It is because Microsoft’s excellent management has figured out that they can make more money for their stockholders by releasing stuff with obvious, known imperfections than they can by making it beautiful or bug-free.”[23]

8

was perfect — or had acceptable quality.8 Renegotiation is not likely to have a noticeable effect on the firm’s behavior, as the contract is long-term (3 years), and there were also no tries to renegotiate before the end of the contract. Regarding regulatory capture, considering the size of the deal, a transparent negotiation, and contract could have helped to lower the informational asymmetry between the government, and the public. Furthermore, a lesser amount of regulatory discretion could have been achieved by using standard auctioning practices, which would also decrease the risk of regulatory capture.

2.3

What software should be included in such a deal?

In the field of education the following software is the most widely needed, and used: an operating system, personal productivity applications, software aiding communication, and applications related to research. As these are different programs that can be separately procured, all tasks should be treated separately. Thus an operating system, such as Windows or Linux, an office suite, such as Microsoft Office, or OpenOffice, an email client, such as Outlook (Express) or Thunderbird, perhaps an instant messenger, such as MSN Messenger, or Gaim, or Skype, a web browser, such as Internet Explorer, or Firefox are the most important pieces. The Microsoft versions of these programs are only available on Microsoft Windows9 , but their alternatives mentioned are available on both Windows, and Linux. The agreement covers software for all of the above mentioned needs. However, there is no apparent widespread need for a development environment or a web page authoring tool (Visual Studio and Frontpage). These programs, therefore, should not be included in the deal, as they may result in unnecessarily increased prices or skewed consumption of these programs due to changes in their perceived prices.

2.4

What are the benefits and costs of the Campus Agreement?

The straightforward benefits of the Campus Agreement in its current form include the diverse gains due to the network effects. A monopoly in the software market is desirable to a certain extent thanks to less time needed for users to learn different interfaces; independent vendors do not need to spend more money to get their applications to run on different platforms; a bigger user base gives higher incentives to develop for the most widespread platform, and more available applications increase the value of the platform to customers. The agreement is also beneficial in the sense that it is clearly a lot cheaper compared to the original price of the included products — even if only Windows and Office are considered. The costs or drawbacks of the agreement include the stark amount of money paid. Moreover, there is a high probability of vendor lock-in due to the proprietary standards and formats Microsoft products tend to prefer, which is highly undesirable. It may 8 Another way to secure market for later versions is to stop supporting earlier versions. For example, Microsoft will end mainstream support of its Visual Basic 6 product despite protests, and a petition signed by close to 4 thousand people[24]. This step, and the fact that VB6 is not included in recent versions of Visual Studio might force users, and developers to upgrade, and migrate to a different programming language. 9 Microsoft Office is also available for OS X

9

result in unreasonably inflated prices. This has in fact been found to have happened in the antitrust lawsuit of the USA DoJ and Microsoft.[25] The external costs associated with this deal include habit formation. By learning to use Microsoft products in school, Hungarians will be more likely to demand the same products in their professional life, thus increasing Microsoft’s market power. As a result, the implied long-term costs may be larger because firms (and the government) will have to bear the costs of monopoly prices or that of retraining their employees. Additionally it seems unwise for a government to support a monopolist proven both in the US and the EU to abuse its monopoly power, and allegedly not even complying with a court order.[26] Especially so, given the fact that possible competitors were not even taken into consideration.

3 Is the Microsoft Campus Agreement a good deal? It most certainly seems to be cheap. But is the price paid really favorable for the supplied software? One can be assured that the beneficial network effects make this deal very tempting for Microsoft. It could be argued that giving away Windows and Office even for free to students is a long-term investment that will pay for itself later, when those same students buy these products for work. Given a hefty sum of HUF 1 billion a year must make it more than worthwhile for Microsoft. But is it worthwhile for the people? If the alternative is that the same software is otherwise pirated, than the external benefits of not having the state-owned higher education institutions brake the law may be very high. If the government does not pay for its software, it is bound to have a much harder time convincing people to pay their taxes, than otherwise. If the alternative would be that those same institutions payed considerably more for the same licences, than this deal in fact is a very good one — achieving a more favorable price through the bigger bargaining power. If, however, the alternative is to get the programs for the given purposes for free, than this agreement does not seem that good again. It is important to note that the Campus Agreement is basically a licensing agreement. Taking a closer look at the included Windows licence shows that it in fact is only an upgrade style licence. In order to legally use Windows XP under this licence, one needs to have a previous Windows licence — or a computer part purchased before November, 2001[2]. This generates uncertainty to whether the operating system can in fact be legally used under the agreement, or not — students who do not major in law may have a hard time finding out whether Windows is included in the deal or not. A few of the above mentioned issues need to be considered in more detail: • Are there any alternatives for the same purposes? • Are the possible alternatives economically favorable? • Are there any political issues to be considered?

3.1

Are there any alternatives for the same purposes?

Without going into much technical detail, alternatives for the following purposes are to be considered: an operating system, office applications, an email client, an instant messenger client, and a web browser. 10

The operating system included in the deal is Microsoft Windows XP (upgrade). An alternative to this operating system is Linux.10 It is a modern, stable[27] operating system that can be as easy to install and use as Windows11 , has a plethora of available software,12 is compatible with Windows’ network interfaces, is compatible with most new hardware, and in many cases supports older hardware better than Windows XP. Another important issue is that of security. Both Microsoft and Linux vendors claim their OS to be more secure than the other, and given that security is in fact hard to determine, different architectural choices suggest that Linux may be the one having an advantage over Windows.[28] Microsoft Office and many other Windows-only applications can be run under Linux with the help of the commercial CrossOver Office 13 or the free software Wine.14 Linux can, therefore, be considered to be a viable alternative to Microsoft Windows. The office suite Microsoft Office was included in the agreement. An alternative to the most often used word processor, spreadsheet, and presentation applications can be the OpenOffice.org15 office suite. It is compatible with Microsoft Office’s formats but also supports a number of other (open) formats. It also has direct portable document format (PDF) export capability, which allows an easy and platform independent way of sharing documents with a professional layout. It is a multi-platform office suite currently available for the Linux, Windows, FreeBSD, Mac OS X, and Irix platforms. It has most of the features of Microsoft Office, and some others. A simpler word processing application is AbiWord,16 which also has most of the common features integrated (such as spell checking), and also runs on Linux and Windows. Outlook Express is bundled with Windows XP as the standard email client and personal information manager (PIM). Thunderbird is a full-fledged email client available for Windows and Linux, and is capable of importing Outlook data. A number of other PIM application suites exist for Linux, including Evolution17 , and KDE PIM18 . Windows Messenger is bundled with Windows XP as the default instant messenger client. MSN Messenger19 is available free of charge on the Internet. Gaim20 is an alternative client that works on Linux, BSD, Mac OS X, and Windows. It is a multi-protocol instant messenger client compatible with the AIM, ICQ, MSN Messenger, Yahoo!, IRC, Jabber, Gadu-Gadu, SILC, GroupWise Messenger, and Zephyr networks. For voice over IP conversations, which are supported by MSN Messenger, GnomeMeeting21 is a free software alternative for Linux and Windows, and Skype22 is 10 The word Linux can refer to the kernel as on http://www.kernel.org, or the GNU/Linux operating system. 11 Most distributions have graphical installers needing minimal user interaction. 12 The Debian distribution for example features more than 16000 software packages, all freely available from the Internet. 13 http://www.codeweavers.com/products/cxoffice/ 14 http://www.winehq.com/ 15 http://www.openoffice.org/ 16 http://www.abisource.org/ 17 http://www.novell.com/products/desktop/features/evolution.html 18 http://pim.kde.org/ 19 http://messenger.msn.com/ 20 http://gaim.sourceforge.net/ 21 http://www.gnomemeeting.org/ 22 http://www.skype.com/

11

a free of charge solution available for Linux, Windows, Mac OS X, and the Pocket PC. Finally, Internet Explorer (IE) is the default web browser bundled with Windows. An arguably technologically superior alternative is Firefox23 . In addition to pop-up blocking (this feature has also been added to IE in its XP SP2 release), it features tabbed browsing, smart keywords, live bookmarks, and useful plug-ins, it is highly configurable, and it is more secure than Internet Explorer.[29] Linux, Wine, OpenOffice.org, AbiWord, Evolution, KDE PIM, Gaim, and GnomeMeeting are licenced under the GNU General Public Licence[30], and Firefox is licensed under the Mozilla Public Licence[31] — they are free/open source software. All these programs are available free of charge on the Internet in source and binary forms.

3.2

Are the possible alternatives economically favorable?

In order to answer that question, one needs not only to look at license costs, as they are only a small portion of the total cost of ownership (TCO), but also the costs of upgrade, maintenance, licence management, and attack vulnerability. As free/open source licences allow free redistribution, and all software mentioned above is available free of charge on the Internet, their licences effectively do not cost anything. Even so, one may want to purchase a distribution of such free software looking for the added benefit of getting pre-packaged24 , supported software. Note, however, that there is no support included in the Campus Agreement. As opposed to proprietary software, technical support for free software can be competed, which implies free market prices. The costs of upgrading Microsoft software generally costs half the original purchase, while free software can be freely redistributed (copied), which implies lower costs. As the Campus licences are paid for in yearly segments, this statement is put in another perspective: as if HUF 1 billion worth of upgrades had to be paid for every year. The costs of retraining because of switching also have to be looked into. In general, most graphical Linux environments are not considerably more different from Windows than different versions of Microsoft Windows itself. Maintenance is a more complex issue. It can include installation of new programs, managing down-time and everyday operative tasks. Microsoft Windows has fairly seamless installation of new programs — most Linux distributions, however provide a way that even eliminates the need of clicking “next” every now and then. Linux has a reputation of stability and reliability, while Windows seems behind in this respect, being more prone to virii, trojans, worms, and various other security breaches. In addition, Linux may have fewer bugs on average than Windows.[32] Several Microsoft studies claim that it is harder to find Linux system administrators and they are more expensive than their Windows counterparts. Others claim that Linux needs less administration than Windows thanks to its higher configurability and its ability to automate many tasks that would be done by hand on Windows systems. Managing Microsoft licences can be challenging due to the many different licencing schemes and the fact that it is not always clear what exactly is one allowed to do 23 http://www.mozilla.org/products/firefox/

24 Certain distributions of pre-packaged free software are available free of charge, while others are more costly.

12

and what is forbidden by the licence. An example to this is the controversy surrounding the Windows licence included in the Campus Agreement. Free/open source software (FOSS) licences are usually simpler, give the licensee more freedom, and are easier to grasp. Regarding attack vulnerability, the development model of FOSS enable developers all around the world to view the source code, discover and fix vulnerabilities with great speed. This supports the hypothesis that FOSS is generally more resistant to malicious attack than its proprietary or closed source counterparts. While Microsoft has announced its Shared Source Initiative, it only allows licensees to look at the source code, but does not allow any modifications or fixes.

3.3

Are there any political issues to be considered?

A number of additional issues, some of which have a political or theoretical nature, and thus have only an indirect connection to economic issues are also useful to consider. The risks and disadvantages of single source solutions are considerably lower in the realm of FOSS than it is with Microsoft products. There is often a need to customize certain parts of a piece of software, which is considerably harder to do without access to complete specification or the source code. Microsoft has a history of withholding certain aspects of the specifications of its products. There is a certain amount of risk involved with the abandonment of a piece of software by its producers. Businesses go out of business and individuals lose interest in projects. There is a stark difference between FOSS and proprietary software in this respect, for in the case of FOSS, if there still is some interest in the given software, it is possible for a new set of developers/supporters to pick up the project. This is usually not the case with proprietary software. Last, but not least, vendor lock-in is highly undesirable because of high long-term costs, and should be avoided if possible.

4 Recommendation It is justifiable for the MIC, and the Ministry of Education to support software licencing in the field of higher education. It is highly desirable that the renewal of the Campus licencing scheme takes into consideration changing its procurement methods. Regarding social efficiency, FOSS is economically superior to proprietary software — such as the Microsoft offering. If a high-powered incentive scheme is to be chosen, measures must be taken to minimize the implied rent extraction. The widespread method of auctioning is capable of reducing the information asymmetry between the Hungarian government and the supplier, and thus reduce rent extraction. Concerns over quality show that a low-powered incentive scheme could also be considered. Striking a long-term deal is desirable in the case of a high-powered incentive scheme, however, it raises the risks and costs of a possible regulatory capture. To minimize this possibility, transparent negotiation and contracting is advisable. The method of auctioning further decreases the risks of regulatory capture. The different applications for different purposes need to be looked at separately. The overall costs of the deal — or, in other words, the total

13

cost of ownership — have to be minimized. Finally, the issue of avoiding a possible vendor lock-in situation, is of utmost importance.

References [1] Tam´as Bodoky. K´etes st´atus´u XP-licencek a Campus Programban. index, last modified: January 15, 2002. Index.hu Rt. Accessed: March 5, 2005. . [2] T´aj´ekoztat´o, 2001. Office of the Prime Minister, Ministry of Education. Accessed: March 5, 2005. . [3] Tam´as Bodoky. K¨ul¨onmeg´allapod´as a Campus Szerz˝od´esben. index, last modified: January 16, 2002. Index.hu Rt. Accessed: March 5, 2005. . [4] Egyel˝ore leg´alisak az egyetemist´ak szoftverei. origo, last modified: March 1, 2005. Axelero Rt. Accessed: March 5, 2005. . [5] Public good. From Wikipedia, the free encyclopedia, last modified: February 25, 2005. Accessed: March 11, 2005. . [6] Nicholas Economides. Durable Goods Monopoly with Network Externalities with Applications to the PC Operating Systems Market. Quarterly Journal of Electronic Commerce, 1(3), 2000. Accessed: March 11, 2005. . [7] Mary Jo Foley. Microsoft to Modify Windows XP Product-Activation Policy. Microsoft Watch, last modified: February 23, 2005. Ziff Davis Publishing Holdings Inc. Accessed: March 11, 2005. . [8] GNU Project − Free Software Foundation. The free software definition. Accessed: March 13, 2005. . [9] Randall Holcombe. Public finance. Chapter 5. Accessed: March 11, 2005. . [10] Yochai Benkler. Coase’s Penguin, or, Linux and The Nature of the Firm. Yale Law Journal, 112(3):369–446, December 2002. Accessed: March 13, 2005. . [11] Eric S. Raymond. The Cathedral & the Bazaar, chapter The Magic Cauldron: The Manufacturing Delusion. O’Reilly, January 2001. Revision 1.20. Accessed: March 31, 2005. .

14

[12] U.S. Infotech Industry Statistics. U.S. Computer Software Industry: Size, Firms, Establishments, Employment, Receipts, last modified: April 14, 2003. Office of Technology and Electronic Commerce. Accessed: March 31, 2005. . [13] Bruce Perens. The Emerging Economic Paradigm of Open Source. Accessed: April 1, 2005. , last modified: February 16, 2005. [14] Josh Lerner and Jean Tirole. The Economics of Technology Sharing: Open Source and Beyond. In NBER Working Paper Series, number 10956. National Bureau of Economic Research, December 2004. Accessed: March 16, 2005. . [15] mikes. Febru´ar v´eg´en lej´ar a Campus-szerz˝od´es. index, Last modified: February 13, 2005. Index.hu Rt. Accessed: March 16, 2005. . [16] Egan Orion. Microsoft loses big deal in Munich. The Inquirer, May 27, 2003. Breakthrough Publishing Ltd. Accessed: March 16, 2005. . [17] Jean-Jacques Laffont and Jean Tirole. Competition in Telecommunications, chapter 2. Munich Lectures. The MIT Press, March 2001. [18] Jean-Jacques Laffont and Jean Tirole. Adverse Selection and Renegotiation in Procurement. Review of Economic Studies, 57:597–625, 1990. [19] World Trade Organization. Report (2003) of the Working Group on Transparency in Government Procurement to the General Council, number 03-3835 in WT/WGTGP/7, July 15 2003. Section I, 9. [20] David A. Wheeler. Why Open Source Software / Free Software (OSS/FS, FLOSS, or FOSS)? Look at the Numbers! Last modified: March 6, 2005. Accessed: March 31, 2005. . [21] Eric S. Raymond. The Cathedral & the Bazaar, chapter The Cathedral & the Bazaar. O’Reilly, January 2001. Revision 1.57. Accessed: March 24, 2005. . [22] Index. Nyolcsz´az¨otvenmilli´oe´ rt kapnak Microsoft-term´ekeket az iskol´ak. index, last modified: March 29, 2004. Index.hu Rt. Accessed: March 24, 2005. . [23] Neal Stephenson. In the Beginning. . . Was the Command Line, chapter Class Struggle on the Desktop. Avon Books, 1999.

15

[24] Martin LaMonica. Microsoft walks VB tightrope. news.com, March 16, 2005. CNET Networks, Inc. Accessed: March 24, 2005. . [25] USA vs. Microsoft Corporation: Court’s Findings of Fact, November 1999. Accessed: March 25, 2005. . [26] Paul Hales. EU sleuths think Microsoft sabotaged Windows. The Inquirer, March 2005. Breakthrough Publishing Ltd. Accessed: March 25, 2005. . [27] Li Ge, Linda Scott, and Mark VanderWiele. Putting Linux reliability to the test. Technical report, IBM developerWorks, December 2003. International Business Machines Corporation. Accessed: March 25, 2005. . [28] Scott Granneman. Linux vs. Windows Viruses. tober 2003. Symantec Corporation. Accessed: .

SecurityFocus, OcMarch 25, 2005.

[29] Scott Granneman. Time to Dump Internet Explorer. SecurityFocus, June 2004. Symantec Corporation. Accessed: March 25, 2005. . [30] GNU General Public Licence, June 1991. .

Free Software Foundation, Inc.

[31] Mozilla Public Licence. . [32] Michelle Delio. Linux: Fewer Bugs Than Rivals. Wired News, December 2004. Lycos, Inc. Accessed: March 27, 2005. .

16