View source for Political Determinants of Competition in the Mobile Telecommunication Industry

<!-- you can have any number of categories here -->
[[Category:Mara Faccio]]
[[Category:Luigi Zingales]]
[[Category:Monopoly, Oligopoly, Market Power and AntiTrust]]
[[Category:Telecommunication Industry]]
<!-- 1 URL must be followed by >= 0 Other URL and Old URL and 1 End URL.-->
{{URL | url = https://research.chicagobooth.edu/~/media/3FDE38093FF24815A1D2259A4B25A512.pdf}}
<!-- {{Other URL | url = }} -->
<!-- {{Old URL | url = }} -->
{{End URL}}
{{DES | des = "We show that the way a government designs the rules of the game has an impact on concentration, competition, and prices. Pro-competition regulation reduces prices, but does not hurt quality of services or investments. More democratic governments tend to design more competitive rules, while more politically connected operators are able to distort the rules in their favor, restricting competition. Government intervention has large redistributive effects: U.S. consumers would gain $65bn a year if U.S. mobile service prices were in line with German ones and $44bn if they were in line with Danish ones." | show=}}
<!-- insert wiki page text here -->
<!-- DPL has problems with categories that have a single quote in them.  Use these explicit workarounds. -->
<!-- otherwise, we would use {{Links}} and {{Quotes}}  -->
{{List|title=Political Determinants of Competition in the Mobile Telecommunication Industry|links=true}}
{{Quotations|title=Political Determinants of Competition in the Mobile Telecommunication Industry|quotes=true}}

{{Text |
Abstract
We study how political factors shape competition in the mobile telecommunication sector. We show that the way a government designs the rules of the game has an impact on concentration, competition, and prices. Pro-competition regulation reduces prices, but does not hurt quality of services or investments. More democratic governments tend to design more competitive rules, while more politically connected operators are able to distort the rules in their favor, restricting competition. Government intervention has large redistributive effects: U.S. consumers would gain $65bn a year if U.S. mobile service prices were in line with German ones and $44bn if they were in line with Danish ones.
* We gratefully acknowledge financial support from the Stigler Center at the University of Chicago Booth School of Business. We also thank Sara Bagagli and Andrea Hamaui for their excellent research assistance, as well as Deniz Yavuz, and several participants at seminars at Emory University, Purdue University, the 2016 Global Corporate Governance Colloquia in Stockholm, Tsinghua University, the University of Michigan, the University of Chicago, Wayne State University, and at the 2016 Dawn or Doom Conference at Purdue University for comments. We thank John Fassnacht and Samodya Abeysiriwardane for developing the C# parsing program that was used to identify political connections. We also thank DotEcon for sharing their spectrum auction data. Luigi Zingales was an independent board member of Telecom Italia from April 2007 to April 2014.
In 2014 Mexican telecom tycoon Carlos Slim was the richest person in the world with an estimated wealth of US$79.6 billion. By 2016 Slim had dropped to fourth place in the Forbes rankings, with his wealth down to only US$47.1 billion, at least in part because the “shares of his pan-Latin American mobile phone operator, America Movil, took a beating in 2015 and early 2016 in the wake of new Mexican telecom regulations.”1
In 2012, an OECD study found that in Mexico “insufficient competition has resulted in poor market penetration” and produced a welfare loss of $129.2 billion during the 2005-2009 period. While heavily contested by Hausman and Rios (2013), these conclusions informed a telecommunication reform, which tried to promote competition. Between 2012 and 2016, the average revenue per user (ARPU) in Mexico dropped by 47% and traffic increased by 59%.2 At the same time, Carlos Slim’s wealth dropped by more than $30 bn.
This episode raises several important questions. Is this association between new rules to promote competition and Slim’s wealth just a coincidence? Can government intervention make markets more competitive? If it can, why aren’t all governments promoting competition? Why does the price of the same basket of mobile phone services vary around the world from $10.07 to $47.25? Why does the price of a 1GB mobile-broadband internet plan vary from $11.24 to $100.28?3
These questions are not limited to the mobile phone industry. Yet, in this paper we try to address them in the context of this industry for several reasons. First, the mobile communication industry is one of the most (if not the most) nationally segmented industries. In every country, including tiny ones like Andorra and Lichtenstein, national governments have the power to shape the competitive landscape by defining property rights (number portability), allowing the use of the Voice over Internet Protocol (VoIP, i.e., phone services over the Internet), and restricting foreign entry or ownership. Not only is the mobile communication industry highly regulated at the national level in every country, but it is also heavily dependent upon a scarce resource controlled by the government: the electromagnetic spectrum used to transmit. Thus, if there is a
1 http://www.forbes.com/profile/carlos-slim-helu/?list=billionaires
2 Based on data from © GSMA Intelligence (2015).
3 These prices correspond respectively to the 5th percentile and the 95th percentile of the distribution of purchasing- power adjusted prices for 2015. The data come from International Telecommunication Union (ITU).
 1
sector where the government can affect the degree of competition, the mobile communication industry is one.
The second reason is the technology underlying this industry is fairly similar across the world and less affected than most by access to energy, water, or pre-existing conditions. Nevertheless, we show that the price variability across countries is larger than that normally observed for other goods and services.
Last but not least, the presence of national and international regulators favors the availability of data. The International Telecommunication Union (ITU), a U.N. specialized agency for Information and Communication Technologies, provides cross-country data on prices and quality of telecommunications; Groupe Speciale Mobile Association (GSMA), an association of operators and companies in the broader mobile sector, collects data on margins and concentration.
Therefore, in this paper we study how in the mobile communication industry the government power to set the rules of the game is used around the world and in whose interest it is used. Employing ITU and GSMA data in 148 countries, we show that a government’s pro- competitive policy –from number portability to VoIP– has a significant effect in reducing concentration and prices. For example, on average, number portability reduces the market share of the two largest operators in a given country by 4 percentage points, reduces the price of a mobile-broadband internet plan with a 1GB volume of data by US$10 per month, and reduces the operators’ EBIDTA margin by 4 percentage points. Israel, whose regulation went from a pro- competitive (overall) regulatory score of 17.5 in 2010 to one of 52 in 2012, saw the average revenues per user (ARPU) drop by 33% and the local operators’ EBITDA margin drop by 10 percentage points.4
One hypothesis –often encountered in the lobbying material of the mobile operators– is that excessive competition reduces not only prices but also the capacity of the industry to invest in technology. According to this hypothesis, reduced competition enforcement is a policy variable to obtain more investment and higher quality of service in the mobile sector. In other words, regulations that restrict competition could reduce market failures by ensuring that operators provide the socially desirable quality of service and investment (Pigou, 1932). We test
4 Based on data from © GSMA Intelligence (2015).
 2
this hypothesis and we find no evidence that a higher degree of competition leads to lower quality of service or less investments. If anything, the results go in the opposite direction. A higher degree of competition is also correlated with lower wages, but the effect is not statistically significant at conventional levels.
A second hypothesis is that governments want to restrict competition so as to maximize the revenues they raise in spectrum auctions. We test this hypothesis and we find the opposite to be true: countries with more pro-competition rules raise relatively more money in auctions. The reason is pretty simple: auctions do not work well with limited numbers of bidders. Pro- competition regulation reduces concentration and thus allows auctions to raise more funds.
A variation of the above argument is that governments do not want to promote full competition because they can raise more revenues by taxing monopoly (or oligopoly) profits. If this motivation was important, it would more likely play a role in countries where it is difficult to raise other forms of taxes (because of bad administration or tax evasion) or where profits are taxed more than consumption. We find no evidence in support of any of these implications.
A final alternative is that regulation is the outcome of the pressure of multiple constituencies (Peltzman, 1976). On the one hand, the operators want to restrict competition to make more profits. Operators may be able to capture the regulator to extract rents for themselves. According to this view, by restricting competition and increasing profits, regulations serve the interest of operators (Tullock, 1967, and Stigler, 1971). On the other hand, consumers want lower prices. Consistent with this hypothesis, we find that governments tend to favor competition more in more democratic countries, where citizens’ preferences are likely to carry more weight. By contrast, we find that rules appear to be tilted more in favor of producers when these producers are more politically connected.5
Thus far, we only looked at regulatory measures that are easily quantifiable and have been codified by ITU. In more advanced countries (like the United States), a major role in
5 Existing research has documented various channels through which political connections affect firm value. Those include evidence of preferred access to credit in normal situations and in times of financial distress (Faccio, Masulis and McConnell, 2005, Houston et al., 2014, Khwaja and Mian, 2005); the award of government contracts and licenses (Goldman, Rocholl, and So, 2013, Beckman, 1999, Faccio and Hsu, 2016, Bunkanwanicha and Wiwattanakantang, 2009); and leniency in enforcement (Fulmer, and Knill, 2012; Yu and Yu, 2011). A salient aspect that has not been considered in depth is the impact of political connections on competitive outcomes.
 3
shaping competition and prices is played by the Antitrust. Unfortunately, it is more difficult to codify the level of antitrust activism and use it in empirical analysis. Therefore, we resort to a mini-case study of countries at very similar levels of economic development and income per capita, but different levels of antitrust activism. To this end we exploit the Atlantic divide between the United States and Europe. As shown by the EU antitrust case against Google, the European Antitrust Authority is more pro-active than its U.S. counterpart. Thus, we compare the level of prices and quality between the U.S. and the two EU countries with levels of regulation closest to the United States (76.5): Denmark (76) and Germany (78.5).
The United States exhibits much higher average revenues per unique subscriber ($67.6 vs $31.01 in Denmark and $23.28 in Germany)6 and a higher price for a standard basket of mobile phone calls and SMS messages: $35.62 vs $7.50 in Denmark and $17.47 in Germany. If U.S. consumers could enjoy the Danish (German) level of competition they would gain respectively $44bn ($65bn) a year. Yet, unlike the case of pro-competition rules, we do find that higher prices in the United States are associated with better quality. In 2013, in the United States, 4G connections represented 23.1% of the total and 4G coverage was 95.1%. In Denmark, 4G connections represented 9.31% of the total and 4G coverage was 92.37%. In Germany, 4G connections represented 2.7% of the total and 4G coverage was 64.54%. Thus not all of the computed difference represents a pure transfer: part of it might be a compensation for more investments.
To check the extent to which higher prices represent a pure transfer we look at the market capitalization of U.S. operators. Lindenberg and Ross (1981) associate the difference between the market value of assets and the book value of assets to the abnormal profits a firm can earn as a result of some stable market power position. We can apply this logic to the four major U.S. carriers (AT&T, Verizon, T-Mobile, and Sprint) to check whether any part of the transfer is capitalized in the market valuations. If all the $44bn ($65bn) difference went to shareholders, it could explain the entire difference between market and book value of the industry with discount rates between 8.6% and 14.6%. While this is certainly not proof that the entire amount of the tax imposed on consumers through higher prices is transferred to shareholders, it is certainly consistent with this hypothesis.
6 Based on data from © GSMA Intelligence (2015).
 4
The rest of the paper proceeds as follows. Section 1 presents the data used. Section 2 describes the uniqueness of the mobile communication industry and the abnormal price variability across countries. Section 3 analyzes how regulatory choices affect concentration, competition, and prices. Section 4 investigates whether regulatory choices affect the quality of service, investments, and wages. Section 5 investigates the various hypotheses that could explain why some countries make regulatory choices that do not foster competition. Section 6 attempts to estimate the welfare transfer produced by lower competition in the United States. Section 7 concludes.
1. The Data
In this section we present the various datasets used in this study. We restrict our analysis to countries covered in both the ITU and the GSMA databases. Further, we exclude territories. In unreported tests, we verify and confirm that our conclusions are robust to including each and every country or territory with available data.
1.1. Regulatory Data
The regulatory data come from the International Telecommunication Union’s (ITU) ICT
Regulatory Tracker. ITU is the U.N. specialized agency for Information and Communication Technologies. It relies directly on statistics and information provided by national telecommunication agencies. The ICT Regulatory Tracker covers various aspects of policy and regulation in the telecommunication sector in 157 countries and territories (150 countries) for a period of 11 years, starting in 2003. The regulatory data were collected annually through the “World Telecommunication/ICT Regulatory Survey”7 and through information gathered by the ITU via internal research. Data were collected, validated, and harmonized by the ITU, and then published.
The ICT Regulatory Tracker covers 50 indicators that reflect the answers to the questions in the regulatory survey. Each answer is codified on a 0 to 2 scale where 0 is given to the most anticompetitive regulation and 2 to the most competitive one. For example, if a country does not
7 The survey can be retrieved at http://www.itu.int/en/ITU- D/RegulatoryMarket/Documents/ITU_Telecommunication-Regulatory-Survey-2016_E.pdf.
 5
require mobile number portability the answer is coded as zero, if a country requires it but it is not necessarily available to all subscribers, the answer is coded as one. If a country requires it and it is available to all subscribers, the answer is coded as two.
These indicators need to be interpreted broadly, since the questions in the survey include the way property rights are allocated (e.g., number portability and auctions), the openness to foreign competition, the independence of the telecom regulator, and even the degree of competition. Some of these measures regard the telecommunication industry as a whole, not just the mobile sector.
The ITU presents an overall regulatory score, which corresponds to the sum of the 50 indicators. We subtract from the overall regulatory score the scores of the answers to questions 37 through 42. We do so because those questions reflect competitive outcomes, rather than regulatory variables, and/or deal with the (public) ownership of the main fixed line operator. The variables’ definitions are summarized in Table 1, while the summary statistics are provided in Table 2.
Following ITU we group the sub-scores into 4 clusters. Cluster 1, which pertains to the independence of the regulator, includes the answers to questions concerning the existence, independence, and accountability of the regulator (questions 1-10). Cluster 2 pertains to the power of the regulator (vs. the power of the government and/or the operators) (questions 11-21). Cluster 3, which pertains to the economies of scale, includes the answers to questions concerning the ease of entry in the sector, i.e., regulations that enable the sharing of fixed costs (questions 22-36). Cluster 4, which pertains to foreign competition, includes the answers to questions concerning whether foreign ownership in facilities-based operators, spectrum-based operators, local service operators/long-distance service operators, international service operators, internet service providers, and value-added service providers is allowed (restricted, or forbidden), as well as two questions concerning the existence of a concept of market dominance and the criteria used to establish market dominance (questions 43-50). As with the overall regulatory score, the answers to questions 37 through 42 are ignored when constructing the index for Cluster 4.
Finally, we also compute a Regulatory Score of the Mobile Sector by summing the scores given to the answers to four questions that are very specific to the mobile sector: (1) whether mobile number portability is required and available to mobile subscribers, (2) whether users are allowed to make voice over internet protocol (VoIP) or phone calls through the internet, (3)
6
whether secondary trading of the spectrum is allowed, and (4) whether foreign participation or ownership in spectrum-based operators is allowed, limited, or forbidden (questions 34, 35, 31, and 46, respectively).
1.2. Measures of Competition
Data on competition come from Groupe Speciale Mobile Association (GSMA) and ITU. GSMA is an association of nearly 800 operators and more than 250 companies in the broader mobile sector. It provides extensive global mobile data for 237 countries and territories: data cover every mobile operator group, network, and mobile virtual network operator in every country worldwide. Data are updated daily. The type of information available comprises: prices of services, operators’ financial data, data traffic, and market data (e.g., market penetration rates, number of unique subscribers, etc.). GSMA has data available starting in 2000.
©GSMA Intelligence (2015) computes each operator’s market share as “[t]otal connections at the end of the period, expressed as a percentage share of the total market connections.” We use this measure to compute (1) C2, the sum of the market shares of the two largest operators in a given country and quarter,8 and (2) the Herfindahl-Hirschman Index, the sum of the squared market share across all operators in a given country and quarter. The Herfindahl-Hirschman Index ranges from 0 to 10,000, where 10,000 denotes a monopoly.
Additionally, the ITU’s ICT Regulatory Tracker reports a classification of the overall level of competition (i.e., “monopoly,” “partial competition,” or “full competition”) in (i) the local and long distance fixed line services, (ii) IMT (3G, 4G, etc.) services, (iii) cable modem, DSL, fixed wireless broadband, (iv) leased lines, and (v) international gateways (questions 37-41 in the ICT Regulatory Tracker). We add up the answers to those 5 questions to construct a qualitative measure of competition.
1.3. Data on Mobile Phone Prices and Margins
Data on mobile phone prices come from multiple sources. In the main analyses we rely on data from ITU and ©GSMA Intelligence (2015). The ITU data are collected through an annual questionnaire addressed to the government agencies responsible for
8 We do not use the perhaps more popular C4, the sum of the market share of the four largest operators, because the mobile phone industry is highly concentrated and C4 has a mean of 96%.
 7
telecommunication/ICT industry. Price data that should be submitted refers to those offered by the largest national operator in terms of market share. A strict set of rules is provided in order to improve the accuracy and the degree of homogeneity in cross-country comparisons. When there are missing values (especially for countries that do not reply to the questionnaire), the ITU collects the information from government agencies’ websites and from the operators’ annual reports.
From ITU we employ (1) the price of a mobile cellular standard basket consisting of a “monthly usage for 30 outgoing calls per month (on-net, off-net to a fixed line, and for peak and off-peak times) in predetermined ratios, plus 100 SMS messages” and (2) the price of “a mobile- broadband USB/dongle-based postpaid tariffs with 1GB volume of data.” These prices are from the ITU’s “Measuring the Information Society” annual reports. The first price bundle is available since 2008. However, due to a change in the composition of the basked implemented by ITU in 2010, to ensure consistency and comparability we only use data for the 5-year period 2010-2014. The second price bundle is available starting in 2012. A benefit of these price bundles is that they allow comparing similar baskets of mobile services.
The second source for mobile data prices is GSMA. From this source we gather operators’ financial data. We focus on two metrics: (1) ARPU (average revenue per user) is the “[t]otal recurring (service) revenue generated per connection per month in the period. Despite the acronym, the metric is strictly average revenue per connection, not per subscriber” and (2) the EBITDA margin, defined as the “[t]otal operating profit in the period, before interest, tax, depreciation and amortisation, expressed as a percentage of total revenue.” Note that these two metrics only reflect the revenues and costs for the mobile phone segment; additionally, for multinational operators the data are reported separately for each country. The GSMA data are available starting in 2000. These measures reflect the actual demand of mobile services (i.e., they allow for the number of minutes of calls, number of SMS etc. to vary across countries depending on actual usage).
To compare the international variability of mobile prices with that of other goods and services we use Numbeo.9 With 2,687,888 prices in 5,846 cities, entered by 307,465 contributors, Numbeo claims to be the world’s largest database of user contributed data about
9 http://www.numbeo.com
 8
cities and countries worldwide. Data collection relies on users’ inputs and manually collected data from different sources, like websites of supermarkets, governmental institutions, newspapers articles, surveys, taxi company websites, etc. In order to control for noise, Numbeo removes outlier prices observed within a certain area and at a point in time. Numbeo reports prices for dozens of commodities and services. Two prices that are relevant for our study are (1) the “Price of 1 min. of Prepaid Mobile Tariff Local (No Discounts or Plans),” and (2) the “Price of Internet (10 Mbps, Unlimited Data, Cable/ADSL).”
1.4. Quality of Service and Other Outcomes
To measure the quality of the mobile service, we use the percentage of total connections that are 3G and 4G, as well as the percentage of the total market geographic surface area that is, at the end of the period, covered with 3G and 4G services. These data come from GSMA. We supplement them with data on the advertised maximum theoretical download speed, in Mbit/s, associated with a 1GB USB/dongle-based mobile broadband prepaid plan, as reported in the ITU’s “World Telecommunication/ICT Indicators Database,” 19th edition, 2015.
From GSMA we also collect data on a number of additional outcomes such as (1) the “total capital expenditure incurred in the period, including both intangible and tangible assets”, scaled by the “[t]otal revenue generated in the period, including both recurring (service) and nonrecurring revenue”; (2) the average number of “[t]otal deployed, and active, base stations on the network at the end of the period,” scaled by revenues; (3) the number of “[t]otal employed head count (fulltime equivalent) for the telecoms business at the end of the period, and if applicable, only within mobile operations,” scaled by revenues; and (4) “[o]perating expenditure incurred in the period related to the cost of employees, including salary costs,” also scaled by revenues.
1.5. Auction Data
Spectrum auction data are from DotEcon Ltd, a U.K.-based consulting firm founded and owned by two former academic economists.10 As of October 18, 2016, The Spectrum Awards Database assembled by DotEcon covered information on 15,186 licenses from 414 spectrum awards in 96 countries and territories. Those awards cover the frequency bands between
10 http://www.dotecon.com/
 9
300MHz to 300GHz, which include the bands used for mobile telephones (e.g., 2G, 3G, and 4G services) as well as fixed and broadband wireless access. The database includes information on the date, type (e.g., auction, beauty contest, etc.) and duration of each license; the license price(s) in local currency; the spectrum endowment; the names of the license winner(s); a number of details about the award processes; as well as demographic and economic indicators.
The very first spectrum auction was held by the U.S. Federal Communications Commission (FCC) in 1994.11 Not surprisingly, “Auction 1 - Nationwide Narrowband (PCS)” is also the first auction included in the DotEcon database. Based on the information provided in the DotEcon database we estimate that, globally, governments raised over $800bn from the spectrum auctions held through the end of 2014.
To compare the revenues raised by governments we rely on the license prices in local currency as provided by DotEcon. We convert those prices into U.S. dollars using the average exchange rate for the quarter (or year, if prior to 2000) in which the auction was held. We assume that a given government raised no revenues from auctions if DotEcon does not report any spectrum auctions in that given country. Exchange rates for each quarter during 2000-2015 are from GSMA. Average exchange rates for the years prior to 2000 are from the World Bank’s “World Development Indicators” and Datastream.
For each country we determine the total revenues (in U.S. dollars) raised from all spectrum auctions covered in DotEcon through December 2014 (the last year for which we have data on telecom prices data from ITU). We alternatively consider the total revenues raised since 1999 (when the first 3G auction was held) and since 2011 (when the first 4G auction was held).12
1.6. Institutional Variables
As a measure of democracy we use an additive eleven-point scale (0-10) index from Polity IV. It captures the presence of institutions and procedures through which citizens can
11 In the U.S., prior to 1982, the FCC used comparative hearings to allocate spectrum licenses. Lotteries were used between 1982 and 1993.
12 We exclude spectrum actions whose only use is Fixed Wireless Access (FWA). Note that while DotEcon attempts to provide comprehensive coverage of spectrum auctions, it does not cover the revenues raised by governments through other allocation mechanisms, such as lotteries, comparative hearings, or private negotiations. While we are unaware of any database that provides comprehensive coverage of those other allocation mechanisms, the data on capital expenditures from GSMA (which includes tangible and intangible capital expenditures) should reflect those (and other) intangible capital expenditures.
 10
express effective preferences about alternative policies and leaders, the existence of institutionalized constraints on the exercise of power by the executive, and the guarantee of civil liberties to all citizens in their daily lives and in acts of political participation.
As a second, alternative measure of democracy, we use constraints on the executive. This Polity IV variable refers to the extent of institutionalized constraints on the decision-making powers of chief executives, whether individuals or collectivities. Such limitations may be imposed by any “accountability groups” –in Western democracies these are usually legislatures. Finally, Polity IV’s political competition refers to the extent to which alternative preferences for policy and leadership can be pursued in the political arena.
From the World Economic Forum (WEF) we gather data on the prevalence of tax evasion in each country (in 2002). The WEF’s index is based on surveys of corporate executives’ assessments concerning the prevalence of tax evasion in their home countries. We gather data on the corporate income tax rate and on the value added tax rates applicable to mobile operators from Deloitte’s “Global Mobile Tax Review 2011” and Wikipedia. From the “Global Mobile Tax Review 2011” we also obtain data on the overall level of taxes as a proportion of the total cost of mobile ownership. The taxes reflected in this measure include consumer taxes such as VAT, GST and custom duties; telecom and/or mobile-specific taxes such as import taxes on handsets and other mobile devices; taxes for using mobile services; and so forth. The total cost of mobile ownership includes the cost of the handset, the connection cost, any rental expenses, and the cost of calls and SMS usage.
As a measure of corruption we use the Heritage Foundation’s measure of corruption. This is based primarily on Transparency International’s Corruption Perceptions Index (CPI), for 2011. The CPI is based on a 10-point scale in which a score of 10 indicates very little corruption and a score of 0 indicates a very corrupt government. In scoring freedom from corruption, the Index converts the raw CPI data to a scale of 0 to 100 by multiplying the CPI score by 10. For example, if a country’s raw CPI data score is 5.5, its overall freedom from corruption score is 55. For countries that are not covered in the CPI, the Heritage Foundation determines the score by using the qualitative information from internationally recognized and reliable sources. This procedure considers the extent to which corruption prevails in a country. The higher the level of corruption, the lower the level of overall economic freedom and the lower a country’s original corruption
11
score. We rescale the index (=100-original index), so that a higher number denotes higher corruption.
The control variables, such as the logarithm of per capita GDP, the logarithm of population, the population density, and the rate of inflation are all from the World Bank’s “World Development Indicators.”
1.7. Political Connections
Extending Faccio (2006), we measure political connections by computing the fraction of top employees (including executives and board members) of each country’s mobile phone operators who served as heads of state, government ministers, or members of parliament in their country, or worked (in the government) for anybody in those positions. To do so, we first obtain the biographies of individuals covering top corporate positions in any of the “Telecommunication Services” firms covered in Capital IQ.13 The Capital IQ sample includes the biographies of 55,656 unique individuals affiliated with 5,890 firms in the “Telecommunication Services” industry, broadly defined. These include fixed line, mobile, and internet operators, as well as other firms. Capital IQ does not indicate whether a given company is a mobile phone operator. Therefore, we manually match the names of the firms in Capital IQ with the mobile telecom operators in the GSMA database. The matching yields a sample of 6,121 individuals affiliated with 410 mobile telecom operators with bios in Capital IQ, as of July 2015, and financial data in GSMA.
We employ a C# text parsing program to identify whether these individuals have political experience. To identify chiefs of states and government ministers we use the “political titles”
13 The positions are: Chief Executive Officer; Co-Chief Executive Officer; Chairman of Management Board; Co- Chairman of Management Board; President; Co-President; Vice Chairman of Management Board; Co-Owner; Top Key Executive; Chief Financial Officer; Co- Chief Financial Officer; Chief Operating Officer; Co-Chief Operating Officer; Member of Management Board; Chief Investment Officer; Co-Chief Investment Officer; Chief Accounting Officer; Head of Investment Banking; Head of Corporate Finance; Head of Research; Chief Technology Officer; Chief Information Officer; Chief Scientific Officer; Chief Administrative Officer; Head of Investor Relations; Chief Compliance Officer; Chief Legal Officer; Head of Corporate Communications; Head of Corporate Development; Head of Marketing; Head of Sales; Head of Human Resources; Senior Key Executive; Controller; Secretary; Treasurer; Unit CEO; Unit President; Other Key Executive; Assistant Secretary; Assistant Treasurer; Consultant; Administrative Professional; Corporate Communication Professional; Corporate Development Professional; Equity Analyst; Finance and Accounting Professional; Fixed Income Analyst; Human Resources Professional; Investment Banking Professional; Investment Professional; Investor Relation Professional; Legal Professional; Marketing Professional; Operations Professional; Other Analyst; Other Professional; Sales Professional; and Technology Professional.
 12
reported in the “Chiefs of State and Cabinet Members of Foreign Governments” directory published by the CIA (https://www.cia.gov/library/publications/world-leaders-1/index.html). Examples of such titles include Emperor, Eternal General Secretary, Eternal President, Secretary of Commerce, Secretary for Communications, and Supreme Leader. We supplement the CIA database with the names of all Presidents, Chancellors, Chairmen, and Emirs ruling as of or after 1980, identified from http://www.rulers.org/index.html.
The parsing program extracts any sentence listing the political titles described above. We then read each of those sentences to verify that the person in question indeed covered a political position. We include both current and past political roles. Examples of political connections include (1) Laura D'Andrea Tyson (Board Member of AT&T Inc.) who, according to Capital IQ, “is a Member of President Barack Obama's Economic Recovery Advisory Board (PERAB)... [and] served as National Economic Adviser to the President of the United States from 1995 to 1996 and Key Architect of President Clinton's domestic and international policy agenda... [and] is a member of Secretary Hillary Clinton’s Foreign Affairs Policy Board”; and (2) William E. Kennard (also Board Member of AT&T Inc.) who, according to Capital IQ, “is a Member of Secretary of State John Kerry's Foreign Affairs Policy Board and U.S. Department of State Foreign Policy Advisory Board.”
We use this information to build an indicator denoting whether a given individual held a political position at any point in her life. For each country, we then compute the fraction of individuals with political experience across all operators and employ that as a measure of political connections. Note that this variable is static.
2. Uniqueness of the Mobile Communication Industry
2.1 Why the Mobile Communication Industry Is Different
Inside the European Union there is unrestricted circulation of goods, services, capital, and people. In the subgroup of EU countries that signed the Schengen Treaty, even border controls have been removed, so travelling by car from Slovenia to Italy or from France to Belgium appears as seamless as travelling from Massachusetts to Rhode Island. Yet, there is a major difference: every time one crosses a national border in Europe the roaming company of the mobile phone changes. In spite of the European integration process, the mobile communication
13
industry remains segmented at the national level. This is not unique to Europe: throughout the world, the mobile communication industry remains very much segmented by country.
The historical origins of this segmentation are complex. In part, it is a carryover of the national regulation of the wireline communication industry. The mobile communication industry sprouted from the wireline communication in the early 1990s. While the natural monopoly and military strategic considerations of the wireline industry do not apply to wireless, being part of a regulated industry caused mobile communication to be regulated as well.
The governments’ desire to collect much-needed tax revenues by auctioning off the right to use the spectrum played a big role in maintaining regulation of the industry. While the total number of spectrum bands available is determined by physics, the actual amount available (after military and police uses have been taken out) is very much a political decision.
Finally, governments like to retain some control of wireless services to facilitate wiretapping both for military and police purposes.
For all these reasons, the mobile industry tends to be quite heavily regulated. Does this regulation affect the degree of competition? Before answering this question, we want to show that this amount of regulation does indeed make the mobile industry different from most others, at least in terms of price variability across countries.
2.2 International Price Variability
While travelling we have all experienced very different dollar prices for an identical product (a sandwich or a bottle of water) in different countries. One reason for this is the difference between flexible wholesale prices and more sticky retail prices, a difference well illustrated by the The Economist’s “Big Mac” index, which compares the price of the McDonald’s hamburger around the world as a measure of temporary currency over or undervaluation.
Another reason is that while water, bread, and meat are tradable goods, a sandwich or a bottle of water available to individual consumers is a bundle of goods, some tradable (like water, bread, and meat) and some not (like rent and local labor). Thus, differences in labor cost or in the costs of some inputs, possibly driven by specific geographical constraints, is another reason for this price variation.
14
Yet, there is a third reason why prices might be different: the different degree of competition of local markets. If the production cost is similar across countries, competition should lead to greater homogeneity of the retail price for the same product.
To illustrate how competitive conditions vary widely around the world, in Table 3 we compare the international price variability of different product and services, by exploiting the Numbeo dataset. In column 1 we consider the dollar prices. To account for possible differences in purchasing power parity, in column 2 we adjust the prices for differences in purchasing power parity across countries. In column 3 we perform this standardization using the McDonald’s index, i.e., the local price of a Big Mac. We group all of the goods and services reported by Numbeo into three categories: tradable, non-tradable non-regulated, and regulated (which are all non-tradable).
In Table 3 we compute the ratio of the 95th percentile of the distribution of prices for an item with the 5th percentile and then we average these ratios across commodities. Contrary to the Law of One Price, we find that prices –even PPP adjusted prices– vary greatly across countries: the 95th percentile of the distribution is roughly four times the 5th percentile. This is true not only in the tradable sector, but in the non-tradable sector (3.5) as well. Yet, it is even truer in regulated sectors: in these sectors the 95/5 ratio is 6.7, while in the telecom sector it is 7.5.
Thus, in regulated industries (and telecom in particular), prices vary more across countries, even after adjusting for differences in purchasing power parity. Since these differences cannot be easily explained in terms of differences in adjustment costs, they might be due to differences in the degree of competition. We will now try to explore where these differences in the degree of competition come from.
3. Concentration, Prices, and Regulation
In Table 4 we explore how different rules of the game affect concentration, prices, and margins. As dependent variables in the various columns we have several measures of concentration, competition, and prices of mobile phone services. Besides some control variables, as a main explanatory variable we have a measure of how the rules existing in a country promote competition. Our first of such measures is the original 50-items ITU measure of regulation, where we have subtracted the answers to questions 37-41, which concern the assessment on the level of competition, as well as the answer to question 42, regarding the ownership status of the
15
main fixed line operator. The overall regulatory score varies greatly across countries, ranging between 12 and 84.
In all of the cross-sectional models that follow for each variable we compute the average during 2010-2014 (or a shorter period, depending on data availability). To mitigate the impact of outliers, we winsorize the price data and the financial variables at the 5th and 95th percentiles. (We confirm that the results are robust using the raw data.) To increase comparability in the international setting that we are employing, in all the analyses that follow, price data are purchasing power parity adjusted using the World Bank (GDP) conversion factors.
As control variables we use the logarithm of per capita GDP (as a measure of economic development), the logarithm of population (as a measure of the potential size of the market), the level of population density (as a measure of the cost of serving the customers), and the level of inflation. This control is very important when the dependent variable is a nominal price level as in columns 5 and 6. Since in the cross-sectional specifications we are taking 5-year averages, nominal variables tend to appear lower in high-inflation countries. Hence we correct for average inflation. We insert this variable in the other specifications only for consistency, but excluding it does not change our conclusions.
In Table 4.A we report the cross-sectional OLS results. Higher values of the regulatory score (i.e., more pro-competition rules) are associated with lower levels of concentration, whether we measure it as C2 or with the Herfindhal index. The coefficient is statistically different from zero when concentration is measured with the Herfindhal index, but not when it is measured as C2 (probably because the lack of variation in this measure). One standard deviation increase in the regulatory score is associated with 1/7 of a standard deviation decline in the Herfindhal Index.
More pro-competition rules are also associated with higher levels of competition, as measured by the ITU. One standard deviation increase in the regulatory score is associated with a 1/3 of a standard deviation increase of the ITU measure of competition. The ITU convention is to replace an observation with a zero when the answer for a country is missing. We compute the score with and without this convention (columns 3 and 4), with similar results. This correlation is not surprising: it is just a check that ITU is measuring rules and outcomes consistently.
The same pattern is present when we use prices or profit margins of the mobile sector as left hand side variables. Higher values of the regulatory score are associated with lower prices
16
and lower margins. The negative coefficient is statistically different from zero for three of the four measures. One standard deviation increase in the regulatory score is associated with 1/5 of a standard deviation decline in the price of mobile internet services. It is perhaps surprising that the results are significant even for the EBITDA margin, since large corporations tend to redirect their profits into tax havens. For example, the top 4 U.S. operators all have subsidiaries in Singapore and Hong Kong. AT&T has additional subsidiaries in the Bermuda, Ireland, and Switzerland; Sprint has additional subsidiaries in Ireland; and T-Mobile has further subsidiaries in Switzerland and Luxembourg14 –all of which are well known tax havens. Thus, they are in a position where they can potentially hide their profits better.
In Table 4.B we use all of the panel data from 2003 to 2013 (for some specific variables and/or countries the time series is shorter due to lack of data). The specification is the same as in Table 4A, but we include country fixed effects. The pattern of the results is very similar. In fact, the statistical significance of the impact of the regulatory score on concentration and competition is stronger, while that of the impact of the regulatory score on prices and margins is slightly weaker.
In unreported regressions we rerun the same estimations (both in the cross-section and in the panel) by using the various subcomponents of the ITU regulatory score: independence of regulator, power of regulator, economies of scale, and foreign competition. All these sub- components have effects on concentration, competition, prices, and margins similar to that of the overall index. The level of statistical significance varies across the various specifications, but we could not identify any single cluster as more or less important.
We then repeat the same exercise with each single sub-score. The three measures that appear more important in reducing concentration and enhancing competition are (1) the portability of the mobile number, (2) the tradability of the spectrum licenses, and (3) the openness to foreign entry. The three measures that appear more important in reducing prices and margins are (1) the availability of voice over the internet, (2) the openness to foreign entry, and (3) number portability. These results are presented in Tables 3C-3F.
14 Based on information from OneSource.
 17
4. Is Concentration “Somehow” Good?
In the previous section we have shown that the design of the rules of the mobile industry appears to affect the degree of concentration, competition, prices, and margins. If this is the case, is it obvious that every government should try to maximize the ITU regulatory score, i.e., promote the maximum level of competition and the lowest level of prices?
In a static framework, it is well known that any deviation from competition creates a deadweight loss, often referred to as Harberger’s triangle. Yet, Harberger (1964) himself showed that this triangle is generally small. Furthermore, dynamic considerations may make some deviations from perfect competition desirable. For example, one version of this argument –often found in industry lobbying material– is that lower prices lead to lower quality, which hurts consumers. Another variation is that lower prices lead to lower profits, which lead to lower investments, which hurts consumers. Thus, the industry does not dispute the kind of evidence presented in Table 4 (i.e., that rules may reduce concentration, increase competition, and lower prices)15 it disputes only the welfare consequences of this outcome: in spite of low prices consumers are worse off because of the low quality of service.
4.1. Concentration and Quality
It is hard to test the welfare implication of this hypothesis, but we can test the impact that concentration, competition, and prices (or better the changes in concentration, competition, and prices induced by pro-competition rules) have on the quality of service and investments. This is what we do in Table 5. The specifications are similar to the ones used in Table 4. The left hand side variables here are different measures of quality of service: from the percentage of connections (or coverage) that are at least 3G or 4G to the speed of broadband internet. Besides the control variables, our main explanatory variable is alternatively a measure of concentration, competition, prices, or margins. This measure is instrumented by the ITU regulatory score. Thus, we are interested in assessing the effect of the component of competition due to the impact of pro-market rules.
As Table 5.A shows, by and large the effect of concentration on quality is negative. Quality is higher in more competitive markets. We repeat those regressions with the other
15 The result that rules that are more pro-competition lead to lower prices in the telecom industry is not new. In its annual “Measuring the Information Society Report,” ITU (2014) reached a similar conclusion.
 18
measures of concentration, competition, and prices (the results are not tabulated to save space). When we do so we find that the impact of prices and margins on quality is negative and sometimes statistically significant at conventional levels. Thus, contrary to industry claims, markets with higher prices tend to exhibit lower, not higher, quality of services. The results are not super strong. Thus, we would not necessarily conclude that more competition leads to higher quality, but we can certainly reject the opposite claim: that less competition leads to higher quality. In particular, there is no evidence suggesting that a government should tilt the rules in a non-competitive dimension to enhance the quality of services. Thus, this evidence provides little support to public interest theories.
4.2. Concentration and Investments
We reach the same conclusion when we look at the impact of concentration, competition, and prices on investments (Table 5.B). The specifications are identical to Table 5.A, only with a measure of investment intensity (capex over revenues in Table 5.B, or number of base stations over revenues in Table 5.C) as the left hand side variables. As for Tables 4.B and 4.C the measures of concentration, competition, prices and margins are instrumented with the ITU regulatory score. Most of the coefficients in Table 5.B have the opposite sign of what the industry lobbying theory would suggest (i.e., that high profitability and prices lead to more investments), albeit no coefficient is statistically different from zero. As before, we cannot conclude that more competition and lower profits lead to more investments, but we can certainly reject the opposite claim: that less competition and higher profits leads to more investments.
4.3. Concentration, Employment, and Wages.
Another possible reason why governments might want to reduce competition is to increase employment or wages. This hypothesis is especially credible in an industry –like the mobile industry– that has seen dramatic improvements in efficiency and less dramatic growth in revenues over the last 10 years, and hence has started to fire people. Since it is more difficult to fire workers in profitable companies, the government can design rules to limit competition in the hope of preserving employment.
Tables 4.D and 4.E explore this possibility. As with the previous two panels, the measures of concentration, competition, prices, and margins are instrumented with the ITU regulatory score. Once again, most of the coefficients have the opposite sign of what the
19
prediction above suggests (i.e., that high profitability and prices lead to higher employment or higher wages), although no coefficient is statistically different from zero. The results so far are, instead, broadly consistent with a view of regulation as a mechanism to create rents for the incumbents and, possibly, politicians.
5. What Shapes the Regulatory Choices?
Given that in the mobile industry the design of the rules appears to affect concentration, competition, prices, and margins, but does not appear to affect the quality of service, investments, and employment, what explains regulatory choices that do not foster competition?
5.1. Auctions
An obvious reason why a government may want to restrict competition is that it wants to maximize the revenues it obtains from auctioning the spectrum. In the unlikely case mobile operators are able to use first degree price discrimination, this choice is also socially efficient. But even if it is not socially efficient, a government might want to pursue this strategy because it is constrained in its tax-raising or because it thinks this form of tax-raising is less politically costly.
In Table 6 we investigate whether that is the case. In column 1 the dependent variable is the total fees paid for spectrum auctions held in a given country through the end of 2014, scaled by aggregate revenues of the mobile operators. In column 2 we restrict our analysis to auctions held since 1999 (those include all 3G and subsequent auctions). In column 3 we further restrict to auctions held since 2011 (so as to include all 4G auctions).
Contrary to the revenue raising hypothesis for restricting competition, we find that auction revenues are higher in countries that have more competition. To put it differently, allowing competition enables governments to extract higher revenues from the operators through the spectrum auctions.
The reason is pretty simple: auctions do not work well with limited numbers of bidders. Pro-competitive rules increase the number of major players in a market and thus the number of bidders. Hence, the idea that a government should restrict competition to increase its revenues is both theoretically and empirically wrong.
20
5.2. Taxation Hypothesis
A variation of the above argument is that governments allow concentration, high prices, and high margins in order to raise more revenues by taxing the incumbent operators. Taxation does not require multiple operators to be effective, thus this argument has potentially greater validity than the previous one. If this motivation was important, it would more likely play a role in countries where it is difficult to raise other forms of taxes (because of bad administration or tax evasion), or where profits are taxed more than consumption.
We investigate this hypothesis in Table 7, where we employ several proxies for a country’s ability to raise taxes. In column 1 we estimate a regression where the left hand side variable is the pro-competitive regulatory score and the right hand side our standard set of control variables and a measure of tax evasion. Consistent with the “taxation hypothesis,” countries with poorer ability to raise taxes (measured by a higher level of tax evasion) exhibit a lower value of the pro-competition regulatory score, but the effect is not statistically significant.
In column 2, we re-estimate the same specification with the level of sales taxes on telecom operators (inclusive of VAT, GST and custom duties, and mobile-specific taxes) as a measure of the fiscal benefit of higher prices. Contrary to the “taxation hypothesis,” countries with higher sales taxes have more pro-competitive regulation and the coefficient is statistically different from zero at the 1% level.
In column 3, we re-estimate the same specification with the level of corporate taxes as a measure of the fiscal benefit of higher prices. Consistent with the “taxation hypothesis,” countries with higher corporate taxes have less pro-competitive regulation, but the coefficient is not statistically different from zero.
Finally, in column 4, we use the proportion of taxes over the total cost of operating a mobile phone as a measure of the fiscal benefit of higher prices. Contrary to the “taxation hypothesis,” countries with higher sales taxes have more pro-competitive regulation and the coefficient is statistically different from zero at the 5% level. Thus, overall we find no evidence in support of the “taxation hypothesis.”
5.3. Ideology
In the rest of Table 7 we investigate the role of ideology in the choice of regulation. Governments strongly influenced by unions might favor concentration because concentration tends to increase wages and employment. In column 5 we use the percentage of the work force
21
affiliated to labor unions as a right hand side variable (Botero et al., 2004). As expected, the sign of the coefficient is negative (more union, less pro-competition regulation), but it is not statistically significant.
In column 6 as a measure of ideology we use the fraction of years between 1928 and 1995 in which “the chief executive and the largest party in the legislature were leftist or centrist.” As expected, the sign of the coefficient is negative and statistically significant at the 5% level.
In column 7 we restrict the years to the 1975 and 1995 period. As expected, the sign of the coefficient is negative but not statistically significant. Thus, overall we find only weak evidence that the ideology matters for regulatory choices.
5.4. Regulatory Capture
If we analyze the regulatory choice from a political economy perspective (Stigler, 1971 and Peltzman, 1976) we obtain some clear predictions. Stronger connections between mobile industry executives and politicians will favor pro-business regulation, which will tend to limit competition and new entry. By contrast, in a democratic country, politicians can buy consensus by introducing a pro-competitive regulation that lowers prices. Thus, we expect regulation to be more pro-competition in more democratic countries and less pro-competition where there are more connections between mobile industry executives and politicians.
In Table 8 we regress our measures of regulation on our standard set of controls and our political economy variables. In columns 1-3 we use various measures of democracy from Polity IV. Regardless of whether we use the overall measure of democracy, the level of constraints on the executive, or the degree of political competition, we find that democratic countries are more likely to have regulation that favors competition in the mobile telecommunication industry.
This finding is further supported by the fact that 4 of the 10 countries with the most anti- competitive regulatory score (i.e., Kuwait, Belarus, Swaziland, and Cuba) have a democracy score of 0. By contrast, all 10 countries with the most pro-competition score have a democracy score of at least 8, with 8 out of 10 having a perfect score. Thus, the evidence is consistent with the idea that, in democratic countries, politicians cater more to consumers’ interest.
In column 4 we focus on the diffusion of political connections, i.e., the role of operators’ concentrated interests. We find that, when phone operators are more politically connected,
22
regulations are less likely to favor competition.16 A one standard deviation increase in the frequency of political connection reduces the regulatory score by one fourth of its standard deviation. The percentage of politically connected employees of mobile operators is 5.35% in the countries with the most anti-competitive regulations (Belarus ranking first with a percentage of 16.67%), while this fraction is 4.76% in the countries with the most pro-competition regulations (with Poland having the highest percentage, 13.04%).
In column 5 we focus on the role of corruption. We find that more corrupt countries tend to have regulations that are less likely to favor competition. To what extent are corruption and political connections the same thing? To assess whether that is the case, we run a horse race between these variables. Although political connections are positively correlated with corruption, the correlation coefficient is surprisingly low (correlation coefficient = 0.0096). Political connections are on average less prevalent in democratic countries. Further democratic countries tend to be less corrupt. With this in mind, the results of the “horse race” specification, reported in column 6, show that political connections remain significant after controlling for corruption (and democracy).
Overall, the results discussed in this section are most consistent with the view that regulation is the outcome of different political pressures.
6. Antitrust and Competition
6.1 A Case Study
Thus far, we only looked at regulatory measures that are easily quantifiable and have been codified by ITU. In more advanced countries (like the United States), the Antitrust authorities play a major role in shaping competition and prices. It is however more difficult to codify the level of antitrust activism and use it in empirical analysis. Therefore, we resort to a mini-case study of countries at very similar level of economic development and income per capita, but different levels of antitrust activism. To this end we exploit the Atlantic divide between the United States and Europe. As shown by the E.U. antitrust case against Google, the
16 We are unable to investigate the role of lobbying or campaign contributions as those data are unfortunately available only for a very restricted number of countries.
 23
European Antitrust Authority is more pro-active than its U.S. counterpart. Thus, we compare the level of prices and quality between the U.S. and the two EU countries with the level of regulation closest to the Unites States (76.5). These are Germany (78.5) and Denmark (76).
The United States exhibits much higher monthly revenues per unique subscriber ($67.6 in 2015:3 vs $23.48 Germany and $31.01 for Denmark), which implies U.S. cellular phone companies have annual revenues per customers $530 higher than their German counterparts, and $439 higher than their Danish counterparts. One reason for the large difference could be that the U.S. carriers tend to subsidize the headsets, while the European carriers do not. The typical subsidy for an iPhone is $500 dollars (they charge $199 for a phone worth $699). Even factoring in this difference, each U.S. customer pays $280 a year more than a German customer and $189 a year more than a Danish one. Given the number of U.S. customers (233.2 million in 2015), this implies that U.S. operators enjoy a transfer of $65.2bn ($44.1bn) vis-à-vis the German (Danish) benchmark.
Not all of this difference is a pure transfer. We do find better quality of service in the United States, where in 2013 4G connections represented 23.1% of the total and 4G coverage was 95.1%. In Germany, 4G connections represented 2.7% of the total 4G coverage was 64.54%, while for Denmark, 4G connections represented 9.31% of the total and 4G coverage was 92.37%.
To explore to what extent this is a transfer to shareholders we look at the market values of U.S. operators. At the end of May 2016, the total market capitalization of the top four U.S. operators amounted to $486bn versus a total book value of $175bn. Lindberg and Ross (1981) associate the difference between the market value of assets and the book value of assets to the abnormal profits a firm can earn as a result of some stable market power position. If we apply this logic to the four major U.S. carriers (AT&T, Verizon, T-Mobile, and Sprint), the capitalized value of abnormal profits in the U.S. mobile industry should be equal to $311bn.
If we assume that the abnormal profits are equal to the after tax transfer from consumer to producers, we can easily calculate the implicit rate at which these abnormal profits are capitalized, assuming they are fixed in perpetuity. These implicit capitalization rates vary between 8.6% and 14.6%. These are very reasonable rates. Thus, the magnitude of the transfer in welfare from consumer to producers is very plausible.
6.2 What Does the Antitrust Have to Do With It?
24
The United States’ mobile phone market started the new millennium highly fragmented as a result of two political decisions: the AT&T breakup of 1984 and the way spectrum was auctioned off in the late 1990s. As one can see from Figure 1, during the new millennium, concentration greatly increased thanks to three mergers: Bell Atlantic and GTE that formed Verizon Wireless in 2000, Cingular Wireless with AT&T in 2004, and Verizon Wireless Communications with Alltel Corporation in 2008. All three transactions faced the scrutiny of both the Department of Justice and the Federal Communication Commissions and all three had to spin off some assets to comply with the requirement. Thus, the consolidation of the U.S. mobile industry in the 2000s was enabled by a series of antitrust decisions.
The Antitrust was also important for the mergers that did not take place. In 2014 a U.S. industry source (FierceWireless) reported that a declaration by the Department of Justice that “any wireless merger among the four Tier 1 carriers would face heightened scrutiny, a strong hint that any potential deal between Sprint and T-Mobile US would face an uphill battle with the DOJ.”17 Not surprisingly, this merger, previously rumored, did not take place.
It is difficult to establish a causal link between the higher margins of the mobile service industry in the United States and the behavior of its Antitrust authority vis-à-vis the United States. Note, however, that in Denmark an attempt to reduce the number of main operators from 4 to 3 was blocked by the European antitrust in 2015.18 In Germany a similar attempt went through in 2014, but only after intense pressure by Angela Merkel.19 Interestingly, between 2013:3 and 2015:3 Danish ARPU levels dropped by 28%, in Germany by 20%, and in the United States by 3%.
7. Conclusions
Is product market competition determined solely by economic factors or is it affected by political factors as well? In this paper we provide some evidence on the importance of political factors in the degree of competition in the mobile service industry. In cross-country comparisons we show that concentration, competition, and prices are affected by the rules of the game
17
18 “Brussels blocks Danish deal in blow to Three's £10.25bn O2 takeover”, Daily Telegraph 11 Sep 2015.
19 Daniel Thomas and Alex Barker “Telecoms: Europe’s scrambled signal”, Financial Times, June 30, 2014
 http://www.fiercetelecom.com/telecom/report-says-microsoft-close-to-naming-satya-nadella-as-its-new-ceo- chinese-service
25
designed by local governments. In addition, we show that this design seems to be affected more by political lobbying than by economic principles.
One interesting byproduct of our analysis is the lack of a negative relation between market competition and quality of service and level of investment in the mobile industry. Industry conditions under (partial) control of the government have large redistributive effects between consumers and producers. The lack of empirical evidence of any benefit in transferring part of the consumer surplus to the operators should weigh in the current antitrust debate on both sides of the Atlantic.
26
References
Backman, Michael. 1999. Asian Eclipse: Exposing the Dark Side of Business in Asia. Singapore: Wiley.
Botero, Juan C., Simeon Djankov, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2004. “The Regulation of Labor.” Quarterly Journal of Economics 119(4): 1339- 1382.
Bunkawanicha, P. and Y. Wiwattanakantang. 2009. “Big business owners in politics.” Review of Financial Studies 22(6): 2133-2168.
Faccio, Mara, and Hung-Chia Hsu. 2016. “Politically connected private equity and employment.” Journal of Finance forthcoming.
Faccio, Mara, Ronald W. Masulis, and John J. McConnell. 2005. “Political connections and corporate bailouts.” Journal of Finance 61: 2597-2635.
Fulmer, Sarah and April M. Knill. 2012. “Political contributions and the severity of government enforcement.” Working Paper. Florida State University.
Goldman, Eitan, Jorg Rocholl, and Jongil So. 2013. “Political connections and the allocation of procurement contracts.” Review of Finance 13: 1617-1648.
Harberger, Arnold C. 1964. “The measurement of waste.” American Economic Review 54(3): 58- 7.
Hausman, Jerry A., and Agustin J. Ros. 2013. “An econometric assessment of telecommunications prices and consumer surplus in Mexico using panel data.” Journal of Regulatory Economics 43(3): 284-304.
Houston, Joel F., Liangliang Jiang, Chen Lin, and Yue Ma. 2014. Political connections and the cost of bank loans. Journal of Accounting Research 52 (1): 193-243.
International Telecommunication Union. 2014. “Measuring the Information Society Report.”http://www.itu.int/en/ITU-D/Statistics/Pages/publications/mis2014.aspx.
Khwaja, A. I., and A. Mian. 2005. “Do lenders favor politically connected firms? Rent-seeking in an emerging financial market.” Quarterly Journal of Economics 120: 1371-1411.
Krueger, Anne. 1974. “The political economy of the rent-seeking society.” American Economic Review 64(3): 291–303.
Lindenberg, Eric B., and Stephen A. Ross. 1981. “Tobin's q Ratio and Industrial Organization." Journal of Business 54:1-32.
OECD. 2012. OECD Review of Telecommunication Policy and Regulation in Mexico. OECD Publishing. http://dx.doi.org/10.1787/9789264060111-en
Peltzman, Sam. 1976. “Toward a more general theory of regulation.” Journal of Law and Economics 19: 211-240.
Pigou, A. C. 1932. The Economics of Welfare. London: Macmillan and Co.
Stigler, George J. 1971. “The theory of economic regulation.” Bell Journal of Economics and
Management Science 2: 3-21.
Tullock, Gordon. 1967. “The welfare costs of tariffs, monopolies, and theft.” Western Economic
Journal 5(3): 224-232.
Yu, Frank, and Xiaoyun Yu. 2011. Corporate lobbying and fraud detection. Journal of Financial
and Quantitative Analysis 46(6): 1865-1891.
}}