DEVELOPING THE INTERNET: ENTREPRENEURSHIP AND PUBLIC POLICY IN COMPARATIVE PERSPECTIVE*

Sandra L. Suárez Department of Political Science Temple University Gladfelter Hall Philadelphia, PA 19122 (215) 204-1468 ssuarez@nimbus.temple.edu and Mauro F. Guillén The Wharton School and Department of Sociology University of Pennsylvania 2016 Steinberg Hall-Dietrich Hall Philadelphia, PA 19104-6370 215-573-6267 guillen@wharton.upenn.edu

November 2000 Version

*

This research has been funded by Temple University (Suárez) and the Wharton

eBusiness Initiative, WeBI (Guillén).

2

DEVELOPING THE INTERNET: ENTREPRENEURSHIP AND PUBLIC POLICY IN COMPARATIVE PERSPECTIVE

Abstract The internet has not diffused uniformly throughout the world. Data on 142 countries indicate that, aside from the levels of economic and infrastructural development, cross-national differences have to do with whether the legal system fosters entrepreneurship, and with whether institutional and political conditions are stable, and democracy is the form of government. Competition and privatization policies in the area of telecommunications do not have consistent effects. After examining international patterns of diffusion for the world as a whole, differences between two matched pairs of countries are systematically compared: Ireland and Singapore, and Argentina and Spain. Patterns of entrepreneurship and public policy in each country are shown to have differed systematically, with distinctive consequences for the development of the internet.

Keywords: Worldwide diffusion of the internet; entrepreneurship; privatization.

3 1. Introduction The internet has not diffused throughout the world in a uniform way. The percentage of the population who is a regular user ranges from over fifty percent in Scandinavia to less than 1 percent in many underdeveloped African, Central American, and South Asian countries. The number of hosts, i.e. computers linked to the internet, is also uneven, ranging from more than one for every 10 people to less than one for every 10,000 (ITU 2000). Previous research on the most advanced countries has established that differences across rich countries are accounted for by per capita income, and by the existing infrastructure and competition in the telecommunications sector (Hargittai 1999; Mosaic Group 1998). We believe that, aside from those variables, privatization, appropriate conditions for entrepreneurship, and democracy are likely to affect internet usage and the number of hosts. We seek to understand to what extent countries are limited by their level of economic and infrastructural development when it comes to accessing and using the internet, and to what extent entrepreneurship and public policy can help surmount such limitations. The stereotypical image of the internet is one of entrepreneurs putting together a website to provide services such as internet access, information, e-commerce, or chats. Recent research has shown that internet development is also shaped by public telecommunications policy, including privatization and deregulation (Mosaic Group 1998). Let us deal with public policy and entrepreneurship in turn.

1.1. Public Policy in Telecommunications The growth of the internet is certainly facilitated by the existence of a developed and well-functioning telecommunications sector offering affordable services to a majority

4 of the population. Until 1980, growth rates and prices in the telecommunication industry remained relatively stable and, save for a few highly developed countries, most people did not use telecommunication services frequently. Basically, people could afford telephone services, for example, when their incomes grew as a result of industrialization. During the 1980s and 1990s privatization and competitive deregulation have changed this relatively stagnant situation. Use of telecommunications services has grown rapidly. Previous research has established that privatization and competition tend to enhance the operating and financial performance of public utilities in general (D’Souza and Megginson 1999a; Megginson and Netter 1999). The evidence on the telecommunications sector during the 1980s and 1990s is overwhelming. In a study of 26 privatized telecommunications firms in 21 countries, D’Souza and Megginson (1999b) found that profitability, output, efficiency, capital expenditure, lines in service, and average salary per employee all increased after privatization. (Net employment remained roughly the same.) Furthermore, the reasons for such increases were not higher prices but efficiency gains, which were further enhanced if competition and deregulation accompanied privatization (see also Levy and Spiller 1994, and Guiérrez and Berg 2000). According to the ITU (2000), 19 countries have fully privatized their formerly state-owned telecommunications operators, and 74 others have privatized partially. (Ninety countries still have state-owned operators.) Governments have also deregulated the various telecommunications services, forcing companies to compete with each other. In local telephone services, for example, there are 44 countries with full competition, 19 duopolies, and 120 monopolies. It is important to note, however, that not all governments that privatize also deregulate and introduce competition. Only 33 countries have both privatized (at least partially) and introduced full competition in local services, while 52

5 privatized but kept a monopoly in place, and 8 privatized and allowed a second company to operate (duopoly). Only 10 countries have both fully privatized and deregulated (ITU 2000). We expect these differences in privatization and competition in telecommunications services to affect internet development.

1.2. Entrepreneurship As a new industry, however, internet development has also been shaped by the activities of private entrepreneurs, although their role and stature varies from country to country. Entrepreneurs thrive when: (1) conditions for doing business in a country are stable and predictable; and (2) corporate legislation provides an adequate protection of investor rights, and facilitates obtaining capital. Case studies have indicated that these factors facilitate internet development (Press et al. 1998; Mosaic Group 1998). Countries differ in terms of the extent to which political-institutional and legal conditions enable entrepreneurs to successfully engage in new activities. In certain countries it is relatively easy to pursue entrepreneurial opportunities and to obtain funding; in others, conditions change swiftly, without warning, or tend to favor incumbent firms with established access to capital financing. Henisz (2000) proposes that conditions for investing and doing business are predictable to the extent that political actors and policymakers are constrained by one another because of their preferences and the institutional characteristics of the political system, including veto points, and other similar constraints on policy change. Data for 147 countries during the 1960-1994 period suggest that growth in telecommunications infrastructure is faster for countries where unexpected policy change is less likely because its political actors and policymakers are constrained (Henisz and Zelner 2001).

6 In addition to constraints on policy change, a democratic form of government is also expected to foster internet growth. Democracies tend to impose more constraints on political actors, and they also tend to promote the growth of communication media in which the user has discretion. For example, Buchner (1988) found that in democratic countries the telephone diffused more rapidly and to a greater extent compared to television; the assumption being that the telephone allows for more user autonomy and discretion, while television is a centralized medium. Thus, democracy may have two positive effects on internet development. First, democracy encourages the growth of media in which the user has discretion, the telephone and the internet being two prime examples. Second, democracy generally provides for a stable and predictable institutional framework in which entrepreneurs thrive. In addition to institutional stability and democratic freedoms, scholars have also pointed out that corporate legal provisions may affect entrepreneurial activity by offering different degrees of investor protection. Countries differ systematically in the extent to which their corporate laws foster entrepreneurial activity and protect investors. The most influential categorization of countries regarding corporate legislation has been proposed by La Porta, Lopez-de-Silanes, Shleifer and Vishny (1998). They argue that the degree of investor protection in a country generates distinct patterns of ownership, governance, and financing of business firms. They propose to analyze and classify each country’s corporate legal provisions in order to identify the extent to which the rights of entrepreneurs and equity capital providers (shareholders) are protected. The greater the degree of protection, the more likely new firms will be founded, their ownership will become eventually dispersed over a large number of shareholders, and their financing

7 will be more tilted towards equity than debt. These predictions have been supported empirically (Guillén 2000; La Porta, Lopez-de-Silanes, and Shleifer 1999). Drawing on comparative legal scholarship, La Porta, Lopez-de-Silanes, Shleifer and Vishny (1998, 1999) classify countries into five main corporate legal traditions: English, French, German, Scandinavian, and formerly socialist. English or common law is shaped by the decisions of judges ruling on specific issues. By contrast, French and German law emerged from Roman civil law, which “uses statuses and comprehensive codes as a primary means of ordering legal material” (La Porta, Lopez-de-Silanes, Shleifer and Vishny 1998:1118). The French Commercial Code was issued by Napoleon in 1807, while the German Commercial Code was written in 1897 with Bismarck as Chancellor. Scandinavian legal systems are in part based on civil law, but are not as derivative from it as the French and German are. Lastly, the former socialist countries need to be treated as a separate category because their legal systems, while in many cases influenced by either French or German law, have been in flux since the end of the Cold War in 1989, and have generally failed to provide a sound foundation for entrepreneurial activity. English, French and German corporate law diffused widely throughout the world following patterns of imperial, economic or cultural influence. Thus, former British colonies—including the United States, Canada, Australia, Ireland, Singapore, and many others—adopted English law. French law spread not only to the colonies in the Near East, Northern and sub-Saharan Africa, Indochina, Oceania, and the Caribbean, but also to Portugal, Spain, Italy, and the countries colonized by the latter. The German legal tradition has been most influential in Austria, Switzerland, Greece, Hungary, Yugoslavia, Japan, and Korea, among others.

8 A detailed comparative analysis of the provisions contained in the different corporate legal traditions reveals that the best protection of investors is awarded by English (common) law, followed by German and Scandinavian law. French law provides the worst protection (La Porta, Lopez-de-Silanes, Shleifer and Vishny 1998). Given the potential importance of entrepreneurial activities to internet development, we expect the English, Scandinavian, and German legal traditions to enhance internet growth relative to the French. We also expect internet companies in countries that adopted German or French corporate law to become eventually controlled by large shareholders because in the absence of adequate legal protection of their rights, investors will tend to accumulate large shareholdings as a way to exercise control (La Porta, Lopez-de-Silanes, and Shleifer 1999).

2. Worldwide Diffusion: Data & Results We measure internet development in two ways: the number of (regular) internet users as a percentage of the population, and the number of hosts per 10,000 population (ITU 2000). We measure the levels of economic and infrastructural development with, respectively, GDP per capita (adjusted for purchasing power), and the number of main telephone lines per capita (World Bank 2000). These continuous measures are logged in order to normalize their distribution. We measure the level of competition in telecommunications with two dummy variables. The first equals one for all countries whose incumbent local telephone operator is fully privatized, and zero otherwise. The second equals one for all countries whose incumbent local telephone operator is partially privatized, and zero otherwise (ITU 2000). We also measure the level of competition in local telephone service with a set of two dummies. The first equals one for countries in

9 which a monopoly exists, and zero otherwise. The second equals one for countries in which a duopoly exists, and zero otherwise (ITU 2000). We measure conditions for entrepreneurship in two ways. We use Henisz’s (2000) index of political constraints, which proxies the degree to which a country’s political institutions and the preferences of elected and appointed officials constrain any one political actor from introducing a sudden change in government policy. This measure approximates the credibility of the existing policy regime, thus giving an indication of to what extent entrepreneurs and firms can take business conditions for granted and expect no change. The index is calculated using spatial modeling techniques and ranges between zero (no constraints) and one (maximum constraints). Given that democratic countries tend to impose more constraints, and also to facilitate the development of communication media in which the user has discretion and autonomy, we also use the ten-point democracy score developed by Gurr and Jaggers (2000). Corporate legislation is measured using La Porta, Lopez-de-Silanes, Shleifer and Vishny’s (1998, 1999) classification of countries according to legal tradition. We implement the classification by creating dummy variables for countries influenced by the English, German, Scandinavian, and socialist legal traditions, with French-law countries as the omitted category in the regression analysis. Table 1 shows the variable definitions, sample statistics, and correlation matrix. Table 2 presents the results of regressing internet users and internet hosts on the independent variables. Because Henisz’s index of political constraints and Gurr and Jaggers’s index of democracy are only available for 114 countries in our sample, we first examine results for the full sample of 142 countries (see Model A). Controlling for a country’s level of economic and infrastructural development, full or partial privatization

10 significantly increase internet use but not the number of hosts. Monopolistic local phone markets significantly reduce the number of hosts but not of users. (Excluding telephone lines from the regression analysis does not change this pattern of results.) Entrepreneurship conditions affect the number of internet users. As expected, the Scandinavian, English, and German legal traditions, in that order, are associated with higher proportions of internet use than the French legal tradition (omitted category) or than the formerly socialist countries (whose coefficient is not significantly different from the French). Only the Scandinavian countries have significantly greater numbers of hosts per capita. These effects on internet development are not only significant but also quantitatively sizable. A country that fully privatizes local telephone service has 49 percent more users per 10,000 population than a country with a state-owned telephone company (exp [.40] = 1.49). A country with a monopoly in local telephone service has 58 percent fewer hosts. Countries within the Scadinavian legal tradition have 260 percent more users and 520 percent more hosts than countries within the French legal tradition. Countries within the English legal tradition have 50 percent more users. Table 2 also reports results adding Henisz’s index of political constraints, an indicator of the extent to which conditions for doing business are unlikely to change without warning (see Model B). This variable significantly increases the number of users, but not of hosts. Some of the other variables lose significance because of the smaller sample size. The democracy index significantly increases the numbers of both users and hosts (see Model C). Including both indexes, however, yields insignificant results because of their high correlation (+.85). Overall, the empirical results reported in Table 2 provide evidence that, controlling for per capita income and infrastructure, internet development is enhanced by predictable business conditions, democracy, and corporate

11 legal traditions with good investor protection. Competition and privatization do not seem to exert systematically significant effects.

3. A Comparison of Two Matched Pairs of Countries While a multivariate statistical analysis of internet growth in 142 countries provides a useful overview of the impact of economic and infrastructural development, entrepreneurship, and public policy, a more detailed comparative study of specific national cases helps further assess the roles of entrepreneurship and public policy. The following sections compare the growth of the internet in two matched pairs of countries. We analyze two countries influenced by French law (Argentina and Spain) with two influenced by English law (Ireland and Singapore). Public policy towards the internet has differed in a way orthogonal to entrepreneurial conditions. Argentina and Ireland have pursued full privatization and competition in telecommunications much more assertively than Spain and Singapore. Thus, the four countries differ in the combination of conditions for entrepreneurship and telecommunications policies (see Table 3). Of the four cases, Singapore has achieved the highest level of internet development, even though it has lagged behind in terms of privatization, deregulation, and competition. As of the end of 1999 nearly 30% of the Singaporean population was an internet user, compared to 12% in Ireland, 7% in Spain, and 2% in Argentina. The same rank ordering was true for hosts, online advertising, and the Information Society Index, which is a composite measure. Ireland, however, led in terms of international internet bandwidth and firms listed on the Nasdaq, and Spain in e-commerce spending per user (see Table 4). We first examine internet development in Singapore and Ireland, and then turn to Spain and Argentina.

12

3.1. Singapore: Partial Privatization & Deregulation and Strong Entrepreneurs Perhaps no other country has pursued internet development policies more aggressively than Singapore. A city-state of 3.2 million people, Singapore stands out not only as one of the most developed information economies in the world but also as a country with an interventionist state casting a long shadow on the telecommunications and internet industries. While the state’s policies have been very effective at creating a high-quality infrastructure for e-business, internet access and use by the citizenry is relatively expensive and tightly controlled. The government itself or government-linked corporations (GLCs)—which are less than wholly owned by the state and report to the Primer Minister’s Coordinating Board or the Ministry of Finance—continue to exercise a firm grip over most internet-related activities. Singapore has become a hub for internetrelated entrepreneurial activities. Scores of foreign companies have based their Asian operations in Singapore, but local private entrepreneurs have found it difficult to succeed. The first experiments with the internet began in 1990 with NUSNET, affiliated with the National University of Singapore. In 1991 the university offered access services (Technet). It was not until 1994 that Singapore Telecom (SingNet)—the telecommunications monopoly—launched the truly first ISP in the country. Unlike in Argentina and Spain, the government tightly controlled the supply of ISP services by granting only a few licenses. In fact, until early 1996 only 5 ISPs were allowed to offer services: SingNet, Technet, Sunsite Singapore (National University of Singapore), Silkroute Ventures, and Accel Infotech. The latter two operated out of servers located in the United States. Competition among them did drive access prices down, especially after September 1995, when Technet was acquired by Sembawang, a GLC, and turned into

13 Pacific Internet (PacNet). SingNet and PacNet are the two largest ISPs. The third largest, Cyberway (now owned by StarHub) started operations in March 1996. New licenses were given to DataOne, CWNS and UUNET in 1999. Another GLC, Singapore Cable Vision (SCV) is also licensed to offer internet access. Given the strong grip on the residential market by the top three, however, new entrants are targeting business users (Blanning 1999; Mesher and Zajac 1997). Internet traffic between local ISPs had to be routed via the United States until May 1996, when the three existing ISPs established a Singapore internet backbone. While this development facilitated internet development, some people saw in it a way for the government to better control the flow of information. In fact, Singapore is one of the countries in which citizens’ daily use of the internet is most tightly controlled by the government. Politically and socially undesirable content, as defined by the government, is forbidden. The use of proxy servers as a censorship tool preventing access to forbidden sites has led to slower traffic of information (Hogan 1999; Rodan 1998). Citizens must show a national identify card in order to open an account with an ISP. Despite appearances, there is a great deal of resistance and discontent about censorship (Rodan 1996; Hancock 1999). The government has been more interested in creating an infrastructure to place Singapore at the heart of the Asian internet and attract foreign companies than in allowing its citizens to use the internet freely. Thus, in June 1998 SingTel, Singapore Cable Vision, and various government agencies launched Singapore One, a national broadband network offering multiple services as well as connectivity with over 20 Asian cities and the United States (Tan and Subramaniam 2000). Although the government fully liberalized telecommunications in April 2000, the industry is dominated by a few firms with strong ties to the government. Singapore

14 Telecom (SingTel) was created in 1992 when the government separated the operator and regulator roles, which had been monopolized for over a century. It then granted SingTel monopoly status until 2007. Singapore’s keen interest in a successful completion of WTO negotiations, however, compelled the government to anticipate liberalization to 2000. In 1997 the government initiated the process of granting new telephony licenses (Ono and Aoki 1998). A mobile license was granted to MobileOne, and both fixed and mobile licenses to StarHub (also a GLC). Broadband service was launched in 1998 by SingTel and Singapore Cable Vision (SCV, also a GLC). Conditions for entrepreneurship in Singapore have had paradoxical consequences. On the one hand, the country subscribes to the English corporate law tradition, keeps red tape to a bare minimum, and presents itself to the world as an excellent location for doing business. On the other, the government participates heavily in the economy, and has always been a key entrepreneur in telecommunications and the internet. Observers of Singapore have long maintained that local entrepreneurs have a hard time, but that foreign firms thrive there. Factors such as risk aversion, the prohibition of using homes as offices (in a country where most apartments are developed by the government), and limited access to the local stock market have mitigated against local entrepreneurship (Wingrove 1994; Asiamoney 2000). Of the two internet-related Singaporean firms listed on the Nasdaq, one is controlled by GLCs (PacNet), and the other started as a producer of PC audio systems and is now branching out into the internet as a content and access provider (Creative Technology). Moreover, none of the ten most visited websites were founded by entrepreneurs. Six are the local subsidiaries of foreign companies, and four are controlled by GLCs (Nielsen 2000).

15 The government has not only limited the number of ISPs—and hence opportunities for entrepreneurship—but also controlled the top four because they are part of a GLC. SingNet is controlled by Singapore Telecom (which is still 80% governmentowned). PacNet, Cyberway and DataOne are controlled by a GLC (Sembawang, StarHub and Singapore Press Holdings, respectively). Foreign telecoms have a minority presence in Cyberway (BT and NTT), CWNS (Cable & Wireless), and UUNET (MCI World Com). The three major internet players (SingTel, PacNet, and Cyberway/StarHub) are not only access providers; they also dominate international connectivity, data transmission, mobile internet (WAP), and network services. Although a variety of internet services were liberalized between October 1998 and September 1999—including ISP, international internet exchange services, and value-added network services—GLCs seem poised to continue their firm grip over the city-state’s internet. The Singaporean government has pursued internet development vigorously, and the state casts a long shadow. Privatization and deregulation have only been partial, a factor that has given an advantage to Hong Kong as an alternative Asian information hub (Far Eastern Economic Review, February 2, 2000, p. 49). In addition, the state has assumed the role of leading entrepreneur, even in areas of the internet that are populated by start-up firms in most countries. The country’s entrepreneurship conditions have favored the arrival of foreign firms, with local entrepreneurs lagging far behind. Still, Singapore is one of the most advanced information economies in the world, with a high number of users and hosts.

16 3.2. Ireland: Full Privatization & Deregulation and Strong Entrepreneurs By all accounts the Irish telecommunications market is very competitive and foreign firms are not discriminated against. Foreign investment has been welcomed in Ireland since the late 1950s when the government implemented an export-led economic policy based on the attraction of foreign capital. As a result of this policy shift, hundreds of multinationals have established operations on the island. During the 1990s, the country experienced unprecedented growth which has been the envy of other EU members plagued with high unemployment and low growth rates. In the period between 1994 to 1999, GDP grew at an annual average rate of 9 percent and the economy is approaching full employment (EIU 1999). Not surprisingly, all political parties in Ireland support policies to attract foreign investment. This level of economic growth has given the Irish a newfound confidence. The government is seeking to turn the country into the e-commerce hub of Europe, and an international internet business center. With this in mind, the minority government of Fianna Fáil has taken a number of initiatives to promote the internet. The telecommunications sector has been privatized and deregulated. The Information Society Commission, created in 1997, promotes greater personal and business use of the internet. An investment in a state-of-the-art, high speed and high capacity global telecommunications network was announced in 1999. In an effort to make the country a more attractive location for startups, the government fast-tracked an e-commerce bill in early 2000 giving legal recognition to electronic signatures and legal standing to electronic documents (Irish Times, Jone 30, 2000). In addition, the Industrial Development Authority, the agency in charge of promoting foreign investment through

17 some of the most attractive incentive packages in Europe, has added e-commerce to the list of businesses it targets (EIU 1999) As a result of these strategies the number of people using the internet has grown rapidly in recent years. According to Amarach Consulting (2000), a Dublin-based market research firm, as of March 2000 there were 592,000 internet users in Ireland, an increase of 33% from October 1999. In spite of this growth, Ireland’s comparative standing has not changed much. Internet use was about 12 percent of the population at the end of 1999, behind most other European Union countries. There seem to be a number of obstacles that have prevented a more dramatic increase in the use of the internet. Namely, the cost of local calls is still considered high partly because Ericom’s (the former state monopoly) has continued to control the local loop. Other obstacles include: low PC ownership, low credit card penetration, and a sales tax rate of 21 percent, higher than in most other EU countries (Reuters 1999). By contrast, cable TV and mobile phone use are high, suggesting that there are perhaps other avenues for internet growth. Eircom (formerly Telecom Eireann) was established as a separate corporate entity in 1984. The company was the exclusive provider of telecommunications services until the process of liberalization of the Irish telecommunications market began in the early 1990s. Since 1994 competition has existed for international and long distance telephone calls, and since 1997 in mobile communication. In 1997 the government permitted other companies to build their own fiber optic cable networks rather than use the lines leased from Telecom Eireann. The telecommunications market was completely liberalized in 1998, 13 months ahead of schedule because the Minister for Public Enterprise was afraid that Telecom Eireann’s monopoly could delay the development of the industry (Irish Times, December 1, 1998). Today there are about 30 licensed telecom operators in

18 Ireland. Among the global players with an Irish presence are MCI/World Com, BT, Stentor Communications (now based in Northern Ireland), and Cable & Wireless. In 1999 Telecom Eireann was fully privatized in the largest equity offer in the country’s history. The company then changed its name to Eircom. Esat Clear (a division of Esat Telecom) is considered the only alternative to Eircom in the residential market. Esat Telecom was created in 1991 and from the beginning it lobbied the EU and Irish governments to hasten the tempo of deregulation. Until the infrastructure restriction was lifted in 1997, the company had built its network on cables leased from Telecom Eireann (“Esat Telecom” 2000). To fund the construction of its own infrastructure Esat Telecom raised millions of dollars on the Nasdaq and Easdaq stock exchanges. The company owns the country’s first commercial internet provider, Ireland on Line, founded in 1992. In January 2000 Esat was bought by British Telecom; it is estimated that the takeover gives the latter “an approximate 60 percent share of the [internet] market, reversing almost exactly Eircom’s dominance a year ago” (NMAER 2000). Since 1994 the costs of telephone calls have gone down but are still considered high, and an important impediment to greater use of the internet. Basically, the cost of international and national calls has gone down dramatically, but this has not been the case for the local loop, the link between Irish homes or small business and the national telecommunications infrastructure and, ultimately, the internet. Despite Irish telecommunications market being generally competitive, the local loop is still effectively controlled by Eircom. In 1999 the Office of the Director of Telecommunications Regulation (ODTR) took Eircom to court claiming that the company was abusing its dominant position in the local loop. Eircom challenged an order by ODTR to lower the

19 prices the company charges its competitors for access to its infrastructure (Irish Times, December 20, 1999). Finally, last June the government announced its decision to end Eircom’s control of the local loop, a move that is expected to dramatically lower internet access costs (Irish Times, June 30, 2000). In an effort to attract foreign e-commerce investment as well as retain the existing foreign investment in the country, the Irish government announced in 1999 a new submarine fiber-optic super highway called Global Crossing that will increase the country’s international network capacity to the U.S. and 24 European cities. The Irish government then sold the initial capacity it purchased to telecommunication providers at a price slightly above wholesale. An important factor explaining the government’s underwriting of the purchase of capacity on the Global Crossing network was Microsoft’s decision in 1998 not to add to its investments in Ireland “because of the lack of competitively priced broadband connections with the internet” (Irish Times, June 24, 2000). The Irish vowed not to let that happen again. Global Crossing attracted a private investment in another transatlantic project. After the Global Crossing deal was announced, Canada’s Network360 announced its own investment in a transatlantic cable project predicting that the demand for bandwidth would exceed what Global Crossing would provide (Irish Times, July 14, 2000). In addition to deregulation and investments in infrastructure, private efforts to increase internet use have included offers of free internet service. In 1999 Gateway revolutionized the market by offering free access to Gateway.net (through UUNet, which is owned by MCI/WorldCom) for purchasers of Gateway computers. Oceanfree.net (a joint venture between BT and the Electricity Supply Board) also offered free access to all internet users and signed more than 30,000 customers in the first three weeks (NMAER

20 2000). Within months other companies such as Esat Telecom and Eircom also began to offer free access to the internet. The experiment with free internet was first tried in Britain where Dixons’ (an electronics retailer) Freeserve Internet service became the country’s biggest ISP. Eircom, however, does not let its competitors offer their customers a cheaper access number that allows for net call for about one third of the cost of a regular phone call. Therefore, while internet use is charged at a lower rate than the cost of local calls, this is not always the case primarily because of Eircom’s monopoly (Irish Times, June 10, 1999). In spite of this, Yahoo and MSN are the two most visited websites by Irish users, followed by Eircom, Microsoft, Esat, Alta Vista, Lycos, AOL, Ocean (cotrolled by BT), and Indigo, part of Eircom (Nielsen 2000). Moreover, as explained earlier, it is expected that the government’s decision to end Eircom’s monopoly of the local loop will give a boost to the internet. Other impediments to the growth of the internet include the large number of people who do not own a PC. To combat this problem Esat Clear gives out new free PCs along with an internet connection for a fixed monthly fee. Due to their high penetration rates, some speculate that mobile telephony and digital TV will turn out to be the primary providers of internet access in Ireland (NMAER 2000). Mobile-phone penetration per 100 inhabitants (37%) is above the European average. However, only about 2 percent of adults use WAP enabled phones to access the net (Amarack Consulting 2000). Ireland’s cable TV network covers about 80 percent of homes. The growth of e-commerce in Ireland is hampered by the low rate of credit card ownership (about 11%), which is lower than the EU average. And while it has been suggested that e-commerce revenues will mostly come from business-to-business transactions, Irish businesses have a long way to go (NMAER 2000). A recent survey by

21 researchers at the University of Ulster found that, contrary to expectations, only 8 percent of Irish businesses have an e-mail address and only 4 percent have a website. Still, according to the Irish Internet Association, 62 percent of internet users have bought on line in the past year (Irish Times, June 21, 2000). Unlike in Spain and Singapore, Irish internet policy has not privileged local firms. Yet, Ireland has produced more internationally successful internet startups. Its English legal tradition and entrepreneurial freedoms have been important factors. Among the most prominent examples are: IONA technologies, a designer of “interoperating” systems; Riverdeep Group and SmartForce, which offer e-learning solutions; Trintech Group, a developer of secure transaction systems for e-commerce; and e-First, an electronic bank. The first four of these firms are listed on the Nasdaq. The Irish internet has grown quickly thanks to deregulation and a set of institutional structures that foster innovation and entrepreneurship. The country realized soon that its domestic market is not large enough to sustain growth without attracting foreign firms to do business there. While the ultimate success of the policy of turning Ireland into an internet hub remains to be seen, the country seems to have adopted a set of policies that make sense as a long-term strategy.

3.3. Spain: Partial Privatization & Deregulation and Weak Entrepreneurs Internet development in Spain is lower than its per capita income would suggest. Although the government has embraced the goal of universal access for the general population and for small firms, telecommunications policies and regulations have either purposefully or unintentionally benefited the dominant telecommunications operator (Telefónica), which is partially privatized. Entrepreneurs have played a very limited role.

22 Currently, no Spanish internet company is listed on the Nasdaq. Together with Telefónica, the large domestic banking, multimedia, and energy groups—as providers of capital, marketing knowledge or content—have become key internet players. The first full-fledged Spanish connection to the internet was made in 1990 by the research-oriented RedIRIS, although the first commercial ISP did not appear until 1992 (Goya). Internet development languished during the early 1990s due to low PC ownership, expensive telephone rates, and scarcity of Spanish-language content. A couple dozen ISPs were offering connectivity services by the end of 1995 to fewer than 250,000 regular internet users (0.7% of the adult population). Faced by the underdevelopment of an industry identified as critical to the country’s social and economic well being, the government of the then flagging Socialist Party (PSOE) moved in early 1996 to require the state-owned telephone monopoly company to set up Infovía, a network parallel to the internet and based on its same protocols and tools, which enabled users to log on through a local phone call. For users located far from an access node, this system could represent savings of up to 95% (Rodríguez 1997). Infovía also enabled access by information and service providers, with several hundred of them getting connected within a year. Users could also navigate the worldwide web and other parts of the internet once on Infovía by using the services of an ISP. While the overriding goal guiding the creation of Infovía was to facilitate universal access at an affordable price, it had the side effect of solidifying the position of the telephone company. Telefónica had been awarded the telephone monopoly in 1922, and was nationalized in 1946. The company operated under the regulatory and policymaking authority of the Ministry of Public Works. Historically one of the largest companies in the country and, by the early 1990s, the largest operator in Latin America,

23 Telefónica was in the process of being privatized when the internet fever hit the world. Infovía contributed to a proliferation of ISPs. At the end of 1997, some 450 ISPs (compared to 150 in the U.S.) provided connectivity services to no more than half a million Spaniards. No provider accounted for more than 5% of the total market (Escobar Espinar 1999:182-191; Rodríguez 1997; Solá 1988). ISP fragmentation helped Telefónica in at least four ways. First, the small size of the ISPs eliminated one possible source of competition for the future, and also made it difficult for new network or telephone service providers to enter the market via acquisition. Second, it discouraged BT and Global One—the two international connectivity providers—from investing in their own access network, and thus forced them to pay Telefónica for the service. Third, although the accounting cost of calls to Infovía was five times greater than the actual revenues (at local-call rates), the increased traffic was in part channeled through underutilized circuits, thus helping Telefónica improve the overall profitability of its infrastructure. Fourth, users started to associate the names of Infovía and Telefónica with the internet and its fabulous opportunities (Solá 1998). ISP proliferation resulted in price competition (access rates were among the lowest in the world), but poor quality and frequent service interruption. After seeing internet use rise seven-fold in three years of operation (to more than 5% of the adult population), the government of the right-of-center Popular Party (elected in 1996) terminated Infovía in January 1999. At the time, it accounted for 85% of all internet access calls (about one million daily). The new network providers (BT’s Interpista and Retevisión’s Retenet) found themselves at a huge disadvantage relative to Telefónica’s Infovía Plus (the newly branded network service based on the old Infovía),

24 both of whom had obtained a license for basic telephony and were busy buying ISPs in order to grow their presence in the internet market. Telecommunications policymaking during the 1990s was generally “liberal” and based on market principles, but it exhibited a decisively regulatory bent aimed at maintaining certain aspects of the “status quo,” especially regarding Telefónica, which easily maintained its status as the “dominant” operator (Souvirón Morenilla 1999). The transition from “statist” to “market” principles, however, did not come naturally. Its pace was primarily dictated by European liberalization directives. Initially, the government engaged in piecemeal legislative and regulatory changes, mostly in the areas of satellite and cable, and managed to obtain in 1993 a five-year waiting period to begin in 1997, which it later agreed to anticipate to November 1998 in exchange for Telefónica’s participation in the Unisource consortium, which turned out not to materialize (Cuétara Martínez 1999). In 1996 the European Union intensified its deregulation drive, and the newly elected conservative government of the Popular Party (PP) committed itself to privatization, deregulation, and the end of the monopoly regime for the telecommunications industry. A 1997 law allowed for a second telephone operator (Retevisión), and in 1998 a new General Law of Telecommunications was passed, providing for the gradual liberalization of most markets. Liberalization and deregulation, however, have mostly followed the model of “restricted competition,” i.e. initially allowing only a limited set of competitors (licensees) selected on the basis of their technical and economic strengths. Thus, this approach has tended to create oligopolistic market structures, tempted firms to collude in prices, and helped Telefónica reposition itself as the dominant player in most internet-related activities (Cuétara Martínez 1999).

25 The government has proclaimed its “information society” strategy repeatedly during the late 1990s, and translated it into actual legislation and regulation (Souvirón Morenilla 1999). The latest initiative—Info XXI, the Information Society for All—is a government-wide effort to make the internet accessible to all, and to turn it into a tool to increase employment, develop infrastructures for business, promote culture, enhance quality of life, improve public services, and help Spanish firms (especially small ones) become more international (Robles 1999). In June 2000, the government announced it would mandate a monthly flat access fee of 2,750 pesetas but applicable only during offpeak hours (weekends and 6 pm-8 am weekdays). At current local-call rates, a user needs to be logged on for an average of at least 55 minutes daily in order for the flat fee to make sense. Users, however, are not happy because the flat fee is not available during peak hours. Large ISPs (which tend to be controlled by long-distance and mobile operators) are not enthusiastic either because they have to pay Telefónica an interconnection fee (Cinco Días 6/17/00, p. 16). Internet development in Spain bears not only the imprint of government policy, but also the impact of long-standing patterns of business and entrepreneurship. Because Spain adopted French corporate law, investor protection is relatively weak. Thus, there is a tendency towards ownership concentration, especially when banks, utilities, and other large companies become shareholders. This pattern of corporate development has translated into a relatively high degree of ownership and market concentration in the internet (CMT 2000). Besides Telefónica and the other two large operators that entered the telephone market after 1995—Retevisión, Airtel—local banks, power utilities, and multimedia groups have also become key internet players. The two top banks—BSCH and BBVA—

26 are among Europe’s largest, and have set their eyes on the internet. BSCH was the largest shareholder of Airtel—the mobile phone operator and internet services provider—until 2000, when it allowed Vodafone to take over in exchange for a 2.8% stake in the world’s largest mobile phone operator. BSCH acquired patagon.com, Latin America’s largest financial portal. Spain’s second largest bank, BBVA, is one of Telefónica’s “strategic” shareholders, and participates in Uno-e, an electronic banking site in collaboration with Ireland’s e-First Group. Similarly to the banks, the country’s top power utilities (Endesa, Iberdrola, Fenosa) have expanded throughout Latin America in their traditional business, and taken stakes in telephone operators and internet companies (Retevisión and Amena, in collaboration with Telecom Italia). They have also invested in data transmission networks leveraging their access to the high-voltage electricity transportation network. The large Spanish multimedia groups—Prisa, Unidad Editorial, Prensa Española—were pioneers in internet development with the electronic editions of their newspapers. They are now turning these sites into full-fledged portals carrying information, e-commerce, chats, and other services. The portals and websites controlled by the multimedia groups and the telecoms are among the most visited by Spaniards. It should be noted that telecoms, banks, multimedia groups, and power utilities frequently collaborate with each other as founders or shareholders of internet and telecommunications companies. Given the strength of the telecommunications operators, banks, multimedia groups, and power utilities, it should be no surprise that, with only a few exceptions, individual entrepreneurs have not become important actors on the Spanish internet, even when annual venture-capital flows to the industry exceed $3 billion (El País, June 25, 2000, p. 8). Individual entrepreneurs played a leading a role during the heroic early and mid 1990s, when fewer than a 100,000 Spaniards used the internet. ISPs or search

27 engines that managed to stand out for their growth were quickly acquired by Telefónica (Olé), Retevisión (Servicom, RedesTB), or foreign firms (Arrakis by BT, and CTV-Jet by Uni2, a relatively small telephone operator controlled by France Télécom). Foreign firms like Vodafone, BT, Telecom Italia, and France Télécom are building their presence, but they rarely find themselves among the top four providers of internet-related services. Although the government has attempted to make the internet universally available, the principle of “restricted competition” has by and large guided deregulation of telecommunications and the internet, thus allowing for telecom operators, banks, multimedia groups, and power utilities to become the key actors even though individual entrepreneurs had been the first movers into the new economy. Despite the fact that policy has encouraged price competition over service quality, internet use in Spain remains low by European standards. The government is presently considering tax incentives to facilitate PC ownership and hopes that the four new WAP-based mobile telephone licensees will accelerate internet growth in the near future—there are 20 million mobile phone users compared to 4.3 million internet users.

3.4. Argentina: Full Privatization & Deregulation and Weak Entrepreneurs Like in Spain, slow internet growth during the mid 1990s prompted the Argentine government to introduce a series of initiatives to facilitate universal access and use. Unlike in Spain, however, telecommunications privatization and partial deregulation had taken place in the early 1990s. Thus, internet-related government policy was not affected by the presence and strategic reactions of a dominant telecommunications operator. As a result, local entrepreneurs have played a more prominent role in internet development than in Spain, and foreign firms (telecommunications operators and banks) have become

28 the key actors. Internet access and use, however, remain relatively low in Argentina due to high cost of telephone calls and low hardware ownership. The national telephone company (ENTel) was originally created in 1946 as part of President Perón’s nationalization drive. It operated as an increasingly inefficient and ineffective monopoly until President Menem (1989-99) singled it out to be the first major state-owned company to be privatized. In 1990 ENTel was split into two territorial monopolies. Telephone services in the Southern part of the country and half of the Buenos Aires metropolitan area were awarded to a consortium led by Telefónica of Spain, while the Northern part and the other half of the Buenos Aires area went to Telecom of Italy. Each foreign operator took a 60% stake in their respective Argentine company. In late 1991 and early 1992 the government floated in the stock market an additional 30% of each company, and in 1992 the remaining 10% was given away to employees. The two monopolies were granted seven-year licenses to expire in 1997 (later extended to 1999), with rates updated following the evolution of the U.S. consumer price index. In order to keep their license, they would need to step up infrastructure investments (Moyano 1999). In fact, investment in telecommunications reached a massive $16.7 billion between 1990 and 1998, with telephone lines growing from 3.6 to 8.0 million (CEP 1999:152-153; Petrazzini and Guerrero 2000). Telefónica and Telecom were asked to set up joint companies for value-added services (Startel), and international network connectivity and data transmission (Telintar). The first internet connection in Argentina was provided in 1991 by a consortium of academic, research, and government institutions. Although bulletin board services had become available by 1993, the first commercial internet (dialup) service did not become available until 1995, with Startel (owned by the two licensed telephone companies until

29 the late 1990s). Other firms, Compuserve and Datamarkets, followed shortly thereafter. Internet use, however, remained low due to high prices for local calls and leased lines (Bassi 1998; Petrazzini and Guerrero 2000). In June 1997 the government declared the internet a “matter of national interest,” and issued a Presidential decree asking the Secretary of Communications to explore the possibility of turning internet access into a “universal service obligation.” Policies to increase internet use would include both price cuts and infrastructure development. Unlike in Spain, however, both the government and the private sector took the initiative during the 1990s. Still in 1997, the government passed a decree regulating leased line prices and effectively bringing them down by as much as 45%, and mandated the creation of a special dialing number (0610) for internet calls, with savings of up to 58%, although mostly applicable to users in the large metropolitan areas (Secom 1998; Petrazzini and Guerrero 2000:96-97, 108). Several dozen ISPs began to offer internet access at the time, but their services were hampered by the fact that traffic between each other had to be channeled through the United States. The dire need for a local network access point (NAP) enabling the local interconnection of ISPs was finally met in March 1998 by the private-sector Argentine Chamber of Data Bases and Online Services (CABASE), which included 18 ISPs among its 35 members. (Nowadays 25 ISPs are connected to the NAP, and they account for 90% of domestic traffic.) Although these public and private initiatives contributed to the proliferation of ISPs (136 by the end of 1998 and 170 by mid 1999), fragmentation never reached Spanish levels. The top four ISPs controlled about 50% of the market in late 1998 and more than 80% by the end of 1999 (Bassi 1998; Petrazzini and Guerrero 2000:103). Moreover, the ISPs created by the telephone operators acquired

30 other successful providers. For example, Telefónica’s Advance purchased Satlink, Compuserve, and Overnet (Mercado, April 1998, pp. 82-88). CABASE’s ambition is to establish direct links to other Latin American countries, and hence be able to negotiate with backbones in the U.S. and Europe on a better footing (CABASE 1999). Access remained unevenly distributed by socioeconomic and rural-urban strata. For example, 85% of users were located in the Buenos Aires area, which is home to 41% of the population (Bassi 1998), 41% have a university degree, and 86% are men (Secom 1998). In response, the government launched in 1998 a program to open a thousand Technological Community Centers (CTCs) providing basic computing, audiovisual, and internet services in relatively small towns (argentin@internet.todos). A 1999 decree mandated the recently privatized Post Office to provide each citizen with a free email address and to install PCs at 6,500 post offices throughout the country. And in 2000 the government set up a $1 billion-dollar subsidized loan program to enable citizens to buy a PC (CACE 1999; Petrazzini and Guerrero 2000:98-99; Industry Standard, May 15, 2000, p. 152). Prices of internet-related calls have dropped since the mid 1990s as a result of government regulation. Dial-in rates decreased from $32.50 an hour in 1995 to $2.38 in 1998 (Bassi 1998). A 1997 resolution cut international link rates by 45%. Unlike in Spain, a flat rate for internet access has been available in Argentina since March 1996, when ImpSat challenged the telephone operators with its satellite link. Competition in long-distance calling was enhanced in 1998 with the approval of two additional licenses, and local calls will be liberalized in November 2000 (CACE 1999). Alternatives to the basic telephone network have developed rapidly in Argentina, although few use them to access the internet. At the end of 1999 more than one in two

31 Argentine families had access to cable TV (CEP 1999). In mobile telephony licenses were given for the Buenos Aires metropolitan market to Movistar (1988) and Miniphone, created jointly by Telefónica and Telecom in 1992, and split in 1999. In the interior provinces, the government licensed CTI Móvil (GTE, AT&T, and Grupo Clarín) in 1994, and Unifón (Telefónica) and Telecom Personal (Telecom) in 1996 (Moyano 1999). CTI was allowed in 1999 to offer services in the Buenos Aires area. Broadband DSL access will become available in some areas during late 2000 at a cost comparable to that for cable. In spite of sharp reductions in cost and multiple government initiatives, internet use remains low in Argentina (2.3 percent of the adult population). Revenues from ecommerce are limited because of low use but also because of traditional distrust for direct and catalogue sales, and uneasiness with remote transactions involving credit cards. The Argentine internet has, like in Spain, become dominated by verticallyintegrated groups. The main difference is that foreign groups play an important role in Argentina. The most important groups are Stet/France Télécom (which controls Telecom), Telefónica, and Agea/Clarín, a local multimedia conglomerate which publishes the world’s largest Spanish language newspaper. They offer a range of internetrelated services. Stet/France Télécom and Telefónica dominate basic telephony, international calls, data transmission, and ISP, and are key players in mobile telephony. Telefónica is well positioned in cable TV and its Terra subsidiary has a specific portal for Argentina. Together with CEI Citicorp Holdings, Telefónica controls several publishers, radio and TV stations, and directories (Atlántida Comunicaciones, Torneos y Competencias, and Telinver). Stet/France Télécom differs from Telefónica in its absence from the ISP, portal, and content stages, and in its presence in satellite data transmission.

32 The third main group, Agea/Clarín, started by building up its presence in multimedia content, cable and satellite TV, and data transmission. It obtained licenses for mobile telephony in 1994 and for long-distance telephony in 1998, and operates Ciudad Internet, the third largest ISP and the second largest portal. Clarín controls 3 of the ten most visited sites. The country’s second largest newspaper, La Nación, also has a site among the top ten. Exxel Group controls the country’s largest music chain, Musimundo, which operates on the most visited websites. Local entrepreneurs launched one of the top ten sites (ElSitio). Unlike in Spain, there are two foreign-controlled sites among the top ten (Yahoo, Terra), and two government sites (Universidad de Buenos Aires and Biblioteca Nacional). Internet entrepreneurship in Argentina has been somewhat more vibrant than in Spain, but successful entrepreneurial startups have been acquired by larger players, like in Spain. Some estimates indicate that about 350 of the 600 Latin American internet startups are located in Argentina, and many have customers or specific sites in other Latin American countries (The Industry Standard, May 15, 2000, p. 52). Several firms and funds are playing the role of venture capitalists: Cima Investments, Citigroup, Southern Cross, Quantum (controlled by George Soros), Dolphin, and Federal de Invesiones (owned by the Massot family). The most successful startups include: patagon.com and latinstocks.com (financial portal), MercadoLibre and DeRemate.com (auctions), and ElSitio.com (portal and ISP). Foreign firms have acquired several of these startups: the Spanish bank BSCH (patagon, and a minority stake in Musimundo, together with Exxel Group), Terra (gauchonet, donde, DeRemate), AoL (latinstocks, together with Exxel Group), and Cisneros (ElSitio, with IMPSA).

33 Internet development in Argentina compares favorably to other Latin American countries. (Argentina is one of the wealthiest countries in the region.) Early privatization has helped, but deregulation measures were timid until the late 1990s. While entrepreneurs have contributed to internet development, their ventures were taken over by domestic media groups, and foreign telecoms and banks. As in Spain, the French corporate legal tradition has produced ownership concentration in Argentina’s internet.

4. Conclusion Internet development benefits from institutional and legal conditions favoring entrepreneurship, but the effects of deregulation, privatization, and competition are not as systematic across countries. These relationships hold after controlling for the levels of economic and infrastructural development. Still, a more detailed analysis of two pairs of matched countries differing in terms of the state’s telecommunications policies and the conditions for entrepreneurship reveals a number of other important contingencies and peculiarities in the growth of the internet worldwide. Internet development in Singapore has been very fast even though the government has implemented less-than-ideal policies. For example, state intervention in the forms of regulation and content censorship has had detrimental effects on internet use. The country’s English-based corporate law has helped attract foreign multinationals, but the government’s practice of protecting state-linked firms has stifled local entrepreneurs. Government-linked corporations and foreign firms control virtually every stage of the internet value chain, including access, international connectivity, data transmission, network services, and WAP telephony. In Ireland, by contrast, the government has

34 combined free foreign investment with deregulation and privatization. Low PC ownership and credit card use, however, seem to have slowed down internet development to levels that are still far from Singarpore’s. Irish entrepreneurs, however, have thrived both because of the English-based corporate legal tradition and the lack of an interventionist state such as Singapore’s, which tends to privilege state-owned and foreign firms to the detriment of local entrepreneurs. Internet development in Argentina and Spain has followed a very different path, largely because of the prominent role played by telecom firms, and other large domestic firms and business groups, which have accumulated extensive stockholdings under the mantle of the French-based corporate legal system. In Spain, the incumbent telecom operator is a dominant force in most internet-related activities, while banking, media, and utility groups are also important. In Argentina, ownership concentration has also put telecom and media groups at the center of the country’s internet. Internet use and ecommerce are still low in both Argentina and Spain. The statistical and comparative analyses reported in this paper indicate that internet development is a complex phenomenon shaped not only by public policy and conditions for entrepreneurship but also by specific contingencies in each country. Thus, more research is needed to document and interpret how the internet has developed in different countries.

35 References Abeles, Martín, Karina Forcinito, and Martín Schorr. 1998. Regulación del mercado de telecomunicaciones. Buenos Aires: Flacso. Amarach Consulting. 2000. “Almost 600,000 Irish Now Online.” www.nua.ie Asiamoney. 2000. “Hi-Tech Start-Ups in Singapore.” Asiamoney 11(5) (June):45-46. Bassi, Roxana. 1998. Informe de internet en Argentina. [Send blank message to reporte@arda.com.ar] Blanning, Robert W. 1999. “Establishing a Corporate Presence on the Internet in Singapore.” Journal of Organizational Computing and Electronic Commerce 9:83-99. Buchner, Bradley Jay. 1988. “Social Control and the Diffusion of Modern Telecommunications Technologies: A Cross-National Study.” American Sociological Review 53(3) (June):446-453. CABASE. 1999. Conferencia de prensa. September 22. Buenos Aires: Cámara Argentina de Bases de Datos y Servicios en Línea. CACE (Cámara Argentina del Comercio Electrónico). 1999. Informe 1999 de ecommerce en Argentina. http://www.cacenet.com.ar, http://planeta.gaiasur.com.ar CEP. 1999. Infraestructura: Una reseña de los años ’90. Buenos Aires: Centro de Estudios para la Producción de la Secretaría de Industria, Comercio y Minería. CMT. 2000. Informe Anual de la Comisión del Mercado de las Telecomunicaciones. Madrid: Comisión del Mercado de las Telecomunicaciones. Cuétara Martínez, Juan Miguel de la. 1999. “Algunos conflictos—resueltos y pendientes—planteados en las telecomunicaciones españolas.” In Javier Cremades and Pablo Mayor Menéndez, eds., La liberalización de las telecomunicaciones en un mundo global. Madrid: La Ley-Actualidad, pp. 201-227.

36 D’Souza, Juliet and William L. Megginson. 1999a. The Financial and Operating Performance of Privatized Firms during the 1990s. Journal of Finance, 54:1397-1438. D’Souza, Juliet and William L. Megginson. 1999b. Sources of Performance Improvement in Privatized Firms: A clinical study of the global telecommunications industry. http://faculty-staff.ou.edu/M/William.L.Megginson-1/ EIU. 1999. “Country Commerce Ireland.” www.eiu.com “Esat Telecom.” 2000. www.ireland.com Escobar Espinar, Modesto. 1999. El comercio electrónico. Madrid: Escuela de Organización Industrial and Retevisión. Guillén, Mauro. F. 2000. “Corporate Governance and Globalization: Is there Convergence Across Countries?” Advances In International Comparative Management 13:175-204. Gurr, Ted Robert, and Keith Jaggers. 2000. Polity98 Project: Regime Characteristics, 1800-1998. www.bsos.umd.edu/cidcm/polity/ Gutiérrez, Luis H., and Sanford Berg. 2000. “Telecommunications Liberalization and Regulatory Governance: Lessons from Latin America.” Telecommunications Policy 24:***. Hancock, Bill. 1999. “Internet Users Furious over SingTel’s Computer Security Probe.” Computers and Security 18(4):288-289. Hargittai, Eszter. 1999. “Weaving the Western Web: Explaining differences in Internet connectivity among OECD countries.” Telecommunications Policy 23(NovemberDecember):701-718. Henisz, Witold J. 2000. “The Institutional Environment for Economic Growth.” Economics and Politics, 12:1-31.

37 Henisz, Witold J., and Bennet A. Zelner. 2001. “The Institutional Environment for Telecommunications Investment.” Journal of Economics & Management Strategy 10(1), forthcoming. Hogan, Sarah B. 1999. “To Net or not to Net: Singapore’s Regulation of the Internet.” Federal Communications Law Journal 51(2):429-446. ITU (International Telecommunication Union). 2000. “Telecommunications at a Glance.” www.itu.int/ti/ La Porta, Rafael, Lopez-de-Silanes, Florencio, and Shleifer, Andrei. 1999. Corporate Ownership around the World. Journal of Finance, 54, 471-517. La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, and Vishny, Robert W. 1998. Law and Finance. Journal of Political Economy, 106:1113-1155. ____. 1999. “The Quality of Government” Journal of Law, Economics, and Organization 15:222-279. Levy, Brian and Pablo T. Spiller. 1994. “The Institutional Foundations of Regulatory Commitment: A Comparative Analysis of Telecommunications Regulation.” Journal of Law, Economics and Organization, 10(2):201-246. Megginson, William L and Jeffry M. Netter. 1999. From State to Market: A survey of empirical studies on privatization. http://facultystaff.ou.edu/M/William.L.Megginson-1/ Mesher, Gene M., and Edward E. Zajac. 1997. “The Political Economy of Telecommunications in Malaysia and Singapore: A Structure-Conduct-Performance Comparative Analysis.” Information Economics and Policy 9:183-202. MF. 1998. Estudio sobre un régimen tarifario específico para el accceso a internet en España. Madrid: Ministerio de Fomento.

38 Mosaic Group. 1998. The Global Diffusion of the Internet Project: An Initial Inductive Study. www.agsd.com/gdi97/gdi97.html Moyano, Ramón. 1999. “El proceso de apertura y liberalización de las telecomunicaciones en Argentina.” In Javier Cremades and Pablo Mayor Menéndez, eds., La liberalización de las telecomunicaciones en un mundo global. Madrid: La Ley-Actualidad, pp. 825-864. Neder, Angel Enrique. 1999. “Telecomunicaciones: Regulación y competencia en Argentina.” Actualidad Económica 45 (January-June):19-25. Nielsen. 2000. “Monthly Top Ten Properties.” www.nielsen-netratings.com NMAER. 2000. “Clicking the Blarney Stone.” New Media Age European Report, June 15. Ono, R., and K. Aoki. 1998. “Convergence and New Regulatory Frameworks.” Telecommunications Policy 22(10):817-838. Petrazzini, Ben A., and Agustina Guerrero. 2000. “Promoting Internet Development: The Case of Argentina.” Telecommunications Policy 24:89-112. Press, Larry, Grey Burkhart, Will Foster, Sy Goodman, Peter Wolcott, and Jon Woodard. 1998. “An Intenet Diffusion Framework.” Communications of the ACM 41:21-26. Reuters. 1999. “Ireland Must Act Quickly on e-Commerce.” www.nua.ie Robles, Elisa. 1999. “La estrategia del Gobierno para el desarrollo de la sociedad de la información en España: La iniciativa ‘Info XXI’, la Sociedad de la Información para Todos.” Economía Industrial 325:11-19. Rodan, Garry. 1996. “Information Technology and Political Control in Singapore.” Working Paper #26, Japan Policy Research Institute www.jpri.org/jpri/wp26.html

39 ____. 1998. “The Internet and Political Control in Singapore.” Political Science Quarterly 113(1):63-89. Rodríguez, J. I. 1997. “Las nuevas rutas.” Revista de los Ministerios de Fomento y Medio Ambiente 458 (November):46-57. Secom. 1998. Segunda audiencia pública de internet. Buenos Aires: Secretaría de Comunicaciones. Solá, Javier. 1998. “Requiem por Infovía.” www.aui.es, “artículos” section. Souvirón Morenilla, José María. 1999. El proceso de liberalización y la nueva regulación de las telecomunicaciones. Granada: Editorial Comares. Tan, Leo Wee Hin, and R. Subramaniam. 2000. “Wiring Up the Island State.” Science 288 (April 28):621-623. Wingrove, Norman. 1994. “Singapore Assists High Tech Start-Ups.” ResearchTechnology Management 37(2) (March-April):4-5. World Bank. 2000. World Development Indicators 1999. CD-ROM database. Washington, D.C.: The World Bank.

Table 1: Variable Definitions, Descriptive Statistics, and Correlations (N = 114)
Variable 1. Log internet users (per 10,000 population) 2. Log internet hosts (per 10,000 population) 3. Log GDP per capita 4. Log telephone main lines (per 1,000 pop) 5. Local telephone service fully privatized 6. Local telephone service privatized in part 7. Monopoly in local telephone service 8. Duopoly in local telephone service 9. English legal tradition 10. German legal tradition 11. Scandinavian legal tradition 12. Formerly Socialist country 13. Political constraints index 14. Democracy index Mean 4.49 .86 8.42 4.22 .11 .39 .54 .12 .30 .04 .04 .15 .46 5.71 Std Dev 2.35 4.40 1.18 1.84 .31 .49 .50 .33 .46 .21 .21 .36 .34 3.96 1 .85 .93 .90 .27 .26 -.30 -.08 .05 .26 .35 -.10 .65 .63 2 3 4 5 6 7 8 9 10 11 12 13

.82 .82 .28 .19 -.31 -.08 .03 .22 .29 .02 .59 .58

.95 .27 .28 -.30 -.09 -.01 .28 .30 -.06 .60 .57

.24 .24 -.23 -.09 -.06 .24 .26 .08 .57 .53

-.28 -.25 .05 .03 -.07 .07 -.07 .33 .28

-.15 -.08 .02 .27 -.09 -.09 .16 .17

-.40 -.08 -.23 -.14 -.05 -.20 -.28

.17 -.08 -.08 .07 -.06 -.03

-.14 -.14 -.27 .08 .00

-.05 -.09 -.09 .22 .20 -.15 .23 .23 -.18 .85

Sources: ITU (2000); World Bank (2000).

Table 2: Robust Regressions of Worldwide Internet Development Model A Users Hosts Model B Users Hosts Model C Users Hosts

Dependent variable: Independent variables (expected sign): GDP per capita (+) Telephone lines (+) Full privatization (+) Partial privatization (+) Monopoly (-) Duopoly (-) English law (+) German law (+) Scandinavian law (+) Formerly Socialist (?) Political constraints (+) Democracy index (+) Constant

.83*** (4.12) .57*** (4.63) .40* (1.76) .38** (2.16) -.07 (-.42) -.01 (-.03) .40** (2.56) .34 (1.55) 1.27*** (4.66) -.29 (-1.12)

.92 .87*** (1.13) (3.44) 1.12** .48*** (1.94) (3.02) .61 .02 (.92) (.06) .45 .06 (.85) (.29) -.87* -.19 (-1.88) (-1.07) -.76 -.12 (-1.16) (-.40) .68 .32* (1.32) (1.91) .30 .20 (.64) (.96) 1.82*** 1.02*** (3.18) (4.32) .63 -.37 (1.00) (-1.22) .88** (2.42)

.86 (.84) 1.15 (1.60) .44 (.90) -.15 (-.27) -.88* (-1.80) -.80 (-1.11) .65 (1.24) .03 (.08) 1.17** (2.23) .48 (.67) 1.78 (1.66)

.88*** (3.39) .48*** (2.91) .09 (.37) .07 (.34) -.11 (-.56) -.11 (-.39) .39** (2.26) .23 (1.12) 1.00*** (4.36) -.30 (-.94)

.87 (.85) 1.14 (1.61) .57 (1.16) -.14 (-.25) -.70 (-1.51) -.79 (-1.09) .80 (1.40) .08 (.18) 1.10** (2.12) .65 (.88)

-5.15*** -11.67** (-4.07) (-2.42) .87 .66 153.9*** 46.4*** 142 142

-5.27*** (-3.38) .89 170.9*** 114

-11.76* (-1.94) .73 39.9*** 114

.08*** (2.64) -5.49*** (-3.42) .89 167.9*** 114

.17* (1.95) -12.18** (-2.02) .73 42.1*** 114

R-squared Model F Number of observations

Student’s t shown in parentheses beneath regression coefficient. *** p < .01 ** p < .05 * p < .10

42

Table 3: Entrepreneurship and Public Policy in Four Countries

Telecommunications & Internet Policies: Legal Basis of Entrepreneurship: Full privatization & competition Partial privatization & competition

French (codified)

Argentina

Spain

English (common law)

Ireland

Singapore

43

Table 4: Internet Development in Ireland, Singapore, Argentina, and Spain, compared to the United States, 1999 a Total population, mn Per capita income, int’l $ at PPPs a Telephone lines per 1,000 pop a Privatization local phone service Competition in local services Index of political constraints Index of democracy Internet users, thousands Internet users, % population Internet hosts, thousands Internet hosts per 10,000 pop Int’l internet bandwidth, Mbps b per million internet users per million hosts e-commerce spending, $mn e-commerce spending per user, $ Internet advertising, $mn Internet advertising, % total Internet-related firms on Nasdaq Information Society Index rank

Ireland Singapore Argentina Spain USA 3.7 3.1 36.1 39.0 270.3 21,482 24,209 12,013 16,212 29,605 434.66 562.00 202.73 413.72 661.31 Partial Partial Full Partial Full c d c Full Full Monopoly Full Full .76 .03 .79 .79 .85 10 2 8 10 10 444 950 900 2,830 110,000 11.98 29.46 2.46 7.18 39.82 63,913 148,249 142,470 469,587 53,175,956 172.50 459.72 38.95 119.12 1,925.14 239.0 497.3 147.3 618.0 538 523 164 218 3,739 3,354 1,034 1,316 68.0 83.8 40.0 123.5 61,090.0 153 88 44 44 555 7.3 n.a. n.a. 24.7 1.20 n.a. n.a. 0.25 4 2 0 0 642 19th 11th 33rd 24th 2nd

Sources: ITU (2000); World Bank (2000); www.idc.com; www.nasdaq.com; www.telegeography.com. Notes: a 1998. b From the country’s largest internet hub, i.e. Dublin, Singapore, Buenos Aires, and Madrid. c There is a dominant operator in terms of market share. d Territorial monopolies.