The drivers of antitrust effectiveness

Hacienda Pública Española / Revista de Economía Pública, 185-(2/2008): 69-88 © 2008, Instituto de Estudios Fiscales The drivers of antitrust effectiveness* JOAN-RAMON BORRELL Departamento de Política Económica, Instituto de Economía Aplicada (IREA) Universitat de Barcelona JUAN LUIS JIMÉNEZ Departamento de Análisis Económico Aplicado, Grupo de Economía de las Infraestructuras y el Transporte Universidad de Las Palmas de Gran Canaria Recibido: Julio, 2007 Aceptado: Febrero, 2008 Abstract This paper shows what drivers the effectiveness of antitrust policy, by using internationally collated data on the perception of effectiveness of competition policy. It concludes that average antitrust effectiveness depends on per capita income and supranational policy leadership, such as the one at the core of the EU. Additionally, it shows that some aspects of competition policy design have a significant impact on policy results. Effectiveness is driven by using an economic approach to judge dominance and abusive practices. We show that antitrust is sounder when the legal mandate on merger policy focuses on competition in markets, rather than on more broadly defined public interests. Antitrust effectiveness is also spurred by taking an active stance against cartels and especially by introducing a leniency programme to enforce the prohibition of cartels. Finally, it is important that an independent antitrust authority has the final say on prohibiting competition restrains. Keywords: Antitrust; Policy Effectiveness; Political economy. JEL classification: D7; L4; O4. ∗ This paper has benefited from helpful comments and suggestions by Tomas Serebrisky, Germà Bel, Javier Campos, Andrés Gómez-Lobo, Pilar Socorro, seminar participants at the Department of Applied Economics in the University of Las Palmas de Gran Canaria and Real Colegio Complutense at Harvard University, and two anonymous referees. Juan Luis Jiménez thanks support from Universidad de Las Palmas de Gran Canaria, research grant UNI2005/19. A previous version of this paper has been published as Working Paper nº 363 in the Fundación de las Cajas de Ahorros (FUNCAS) collection. The usual disclaimer applies. 70 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ 1. Introduction Competition policy is at the core of a set of policies designed to foster product rivalry in the goods and services markets. Promoting competition has become a policy objective enforced in order to pursue the ultimate goal of improving economic performance. Policy-makers claim that, by designing and enforcing competition legislation and antitrust institutions that promote and protect competition effectively in the product markets, welfare improvements are rendered. However, Bertrand and Ivaldi (2006) emphasise the lack of research into identifying the distinct features of antitrust that drive the effectiveness of promoting competition in the domestic markets. Competition policy has very gradually been introduced in countries around the world; Canada passed the first antitrust statute in 1889. The US has had a federal antitrust policy since the Sherman Act of 1890; however, according to ABA Section of Antitrust Law (2003) 13 states had already enacted some form of antitrust statute. In 1889 Kansas enacted the first generally applicable state antitrust law. Even Senator Sherman described the federal law he was promoting as supplementary to the states’ antitrust statutes. Most European countries, including the EU since its inception with the Treaty of Rome, have had antitrust policies since the mid twentieth century; other countries have adopted antitrust regimes much more recently. Antitrust acts have been included in market orientated reforms packages in South East Asia and Latin America, and also through regional trade agreements; i.e. by Eastern Europe prior to acceding to the EU. Commitments to pass antitrust legislation have been included also in multilateral agreements such as the Uruguay Round and in the accession agreements of the new members of the WTO. Around 100 nations currently have competition laws. Merger policy was established in the USA in 1914, but prior notifications of mergers have only been compulsory since 1976. The EU established merger controls with a system of prior notification with a 1990 Council regulation. Although antitrust statutes have been enacted worldwide, antitrust law design and implementation differs substantially across countries. It is an open question as to the type of institution and as to which policy practices drive sounder enforcement. The aim of this paper is to shed light on whether some of the distinctive characteristics related to the way antitrust is enforced generally spur policy effectiveness. In order to do so, we have coded qualitative information on objective features of competition policy from several countries around the world. Next we offer a comprehensive insight into what drives the effectiveness of competition policy, from national variations in antitrust design and implementation, to variations across policy areas. The paper is organised as follows. After this introductory section, Section 2 reviews the literature on the drivers of antitrust effectiveness. Section 3 describes the data and the empirical strategy. Section 4 shows and discusses the results, and the paper concludes with Section 5. 71 The drivers of antitrust effectiveness 2. Literature review According to Aghion and Schankerman (2004), competition enhancing policies, reforms and certain types of physical infrastructure facilitate welfare gains in three ways, through cost reduction by incumbent firms, market selection of more efficient firms and entry. First, competition provides an incentive for existing firms to reduce their production costs. Static efficiency gains are achieved through reducing slack in the use of inputs, and via improved resource allocation in response to higher competitive pressures; i.e. cost reduction or restructuring. Second, competition moves market shares from high-cost to low-cost firms, and this reduces the industry average production costs; this is market selection of the efficient incumbent firms. Third, post-entry competition reduces the incentives for new high-cost firms to enter the market, but it encourages entry by low-cost potential entrants. Furthermore, competition might boost dynamic efficiency by spurring innovation and the faster diffusion of new technologies, along the lines suggested by Aghion and Howitt (1996). Although there is a consensus in the literature that competition provides welfare gains, it is less clear cut how the effective enforcement of antitrust law fosters competition in the market place, and whether antitrust policy spurs welfare. There is no common consensus on how antitrust should be designed and enforced, in order to effectively promote competition. There is a large literature on regulatory effectiveness. Recent papers such as Gutiérrez (2003), Stern and Trillas (2003), Levine, Stern and Trillas (2005), Stern and Cubbin (2005) and Cubbin and Stern (2006) discuss theoretically and empirically what drives effectiveness in regulation of telecoms and energy. By contrast, less effort has been devoted to analyse the drivers of antitrust effectiveness. Additionally, it is far from established how the effectiveness of competition policy and the impact of antitrust and welfare across countries should be measured. There are two main avenues of research. On one hand, Serebrisky (2004) and Voigt (2006) use synthetic indicators to obtain a comprehensive picture of the strength of competition policy, based on objective indicators that rank countries with a set of characteristics that are assumed to drive antitrust effectiveness. Voigt (2006) estimates the significant and positive impacts on productivity of policy characteristics that are assumed to strengthen antitrust. However, he shows that the estimated impacts are not particularly robust, when the indicators for the general quality of institutions are included. On the other hand, some papers such as Dutz and Hayri (2000) and Borrell and Tolosa (2008) rely mostly on subjective indicators that try to assess the effectiveness of competition policy across countries, in terms of their respective growth and productivity. Both these papers show that antitrust has a strong positive impact on growth and productivity. Subjective indicators measure the perceived effectiveness of policy, using surveys of mainly business people. There are two measures of perceived effectiveness of competition policy. The World Economic Forum (WEF) used the Executive Opinion Survey to compile the Global Competitiveness Report.1 It asks executives to rank their country according to the following statement: “Antimonop- 72 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ oly policy in your country is 1=is lax and ineffective at promoting competition, 7=effectively promotes competition.” The other indicator is the one included in the World Competitiveness Yearbook that is compiled by the International Institute for Management Development (IMD).2 IMD asks, “Is competition legislation efficient in preventing unfair competition?” According to Voigt (2006), the two subjective measures of perceived competition are highly correlated (over 80%), and they are both also correlated with general policy effectiveness. Neither the WEF nor the IMD indicator of competition policy effectiveness offer data that separates the effect of national competition policy from the EU competition policy; therefore, it indicates the overall effect of both policies for the EU member states. Synthetic indicators are comprised of information from different policy aspects, like the presence or absence of certain legal provisions, the rules adjudicating antitrust cases, the severity of penalties, and the resources available to the competition authority and its autonomy. The indicators show the variance across countries in the design and implementation of antitrust policy. For instance, rules adjudicating antitrust cases differ strongly among the countries and policy domains. In some countries, some competition restrictions, such as price fixing agreements, bid rigging or imposing minimum resale prices, are per se illegal. For example price fixing agreements are illegal per se in Australia, Greece, Mexico; and the USA defines, this conduct as unlawful, regardless of its effect on competition. In 1990 Italy also joined the group of countries that define price-fixing as per se illegal. By contrast, in most countries competition restrictions are judged according to the rule of reason, even price fixing tends to be adjudicated on a case by case basis in the European Union and in most European jurisdictions. Conduct is judged lawful or unlawful, depending on its effect on competition, as determined by the evidence in each case. As Motta (2004) has highlighted, article 81(3) of the Treaty of the European Union states that all agreements among competitors do not fall under a per se rule of prohibition. Conversely, Gual et al. (2005) and Ahlborn, Evans and Padilla (2004) state that Europe tends to use the per se illegality rule, when adjudicating the use of exclusionary practices such as foreclosure, predatory pricing, price squeezes, fidelity rebates or tying by dominant firms. Similar cases are judged, using the rule of reason, on a case by case basis in the USA and in other jurisdictions. The differences among countries are not only restricted to adjudicating rules, but also exist in matters such as the degree of independence of the antitrust authorities, the use of criminal sanctions, the way plaintiffs can seek redress before civil courts, whether to adopt an economic or legal approach to dominant position abuses, merger policy objectives and enforcement. There have been different efforts to systematically compare competition rules and institutions across countries. Nicholson (2004) reviews the different efforts made at quantifying an- 73 The drivers of antitrust effectiveness titrust regimes. CUTS International has reviewed around 100 jurisdictions worldwide.3 Global Competition Review, the official research partner of the International Bar Association, publishes accounts of the specific legal and enforcement aspects of competition policy in different domains, such as cartels, abuses of dominant position and mergers in developed countries.4 Voigt (2006) has undertaken the most comprehensive quantitative analysis of competition laws. He gathered information regarding the basis of competition laws, legal regimes and economic approaches; i.e. per se rules versus the rule of reason, the structure of competition authorities and de jure and de facto independence. He constructed indicators regarding these aspects of competition policy, and has made assumptions on how these elements have made competition policy effective. Subjective and objective measures both have strengths and weaknesses. The subjective measures depend strongly on the expectations of the business people surveyed, regarding their views on effective competition policy, whereas the objective indicators depend on how they are constructed. A higher value of the indicator can be achieved by optimizing certain areas, while other very important elements are replaced with less decisive or incomplete components; however, this can blur the robustness of the whole policy when being evaluated. By making the nature of the relationship between subjective measures of effectiveness and the objective antitrust design indicators and enforcement more explicit, we can understand the strength and weaknesses of comparative antitrust analysis across countries better. 3. Empirical Strategy and Data In this paper the data will explain the relationship of the perceived effectiveness of antitrust versus a set of objective characteristics of competition policy design and enforcement. Our starting point is that average antitrust effectiveness depends mostly on three different sets of covariates. Firstly, we would expect the average policy effectiveness to increase when there is a specialized administrative or judicial body in charge of antitrust law enforcement. Secondly, we expect antitrust statutes to be more effective when implemented selectively, and when tough prohibitions and sanctions are imposed on the more damaging competition restrains. Finally, we would expect the effectiveness of antitrust to be correlated with mean income, as is generally the case for all public policies. As to who should enforce antitrust, Carlton and Picker (2007) proffer the principal-agent approach found in modern political science literature on legislative bargaining. They analyse the pros and cons of government regulation versus antitrust, whether, either by special agencies or by the judiciary, as a means of controlling competition. They conclude that, historically, the increased enforcement of antitrust law has brought benefits to consumers. Sometimes the benefits came after the government regulation had been replaced by antitrust, but on other occasions the two need to be used in tandem. 74 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ With respect to the substantive features of cartel policy, monopolization policy and merger policy, we are going to empirically analyse them in accordance with recent literature on antitrust effectiveness. Baker (2003) and Kaplow and Shapiro (2007) have shown what effective antitrust enforcement is for a consensus of lawyers and economists. In the domains of prosecution and deterrence of cartels, the most important policy is to fight against high profile international cartels that cause substantial harm. With respect to monopolization, the key issue is to prosecute dominant firms that engage in the more harmful exclusionary practices. And, merger policies have to be tailored to remedy the harmful effects of a tiny proportion of horizontal mergers in concentrated markets. This could eventually lead to unilateral price increases and mergers that affect competition in the development of new products. We have gathered qualitative information regarding the institutional setting up of antitrust, cartel policy enforcement, the handling of monopolization and dominant positions, and the features of merger policy worldwide including developed and developing countries. We have coded information for 13 distinctive objective policy features, that is, 611 items of data. We also have information on subjective antitrust effectiveness indicators and mean income. We will analyse the data using cross-section variation. Most covariates are not expected to vary much from one year to another, so it would not be more informative to have a panel data with year observations around 2004. And, unfortunately, the available data does not allow us to analyse the effect of long-run governance changes on antitrust effectiveness using panel data techniques as in Gutiérrez (2003) for telecoms regulation and in Cubbin and Stern (2006) for electricity. Table 1 shows the definition of the coded data, and table 2 shows the data’s descriptive statistics. Table 3 shows the distribution of the detailed features of the setting up, design and implementation of antitrust policy across countries. Table 1 VARIABLES IN THE DATASET Name Acronym Source Description Antitrust effectiveness effectiv WEF Anti-monopoly policy in your country (1 = is lax and not effective at promoting competition, 7 = effectively promotes competition) Independence on antitrust decisions indcard Authors 1 = There is an independent competition authority; 0 = In other case Government prosecution mininst Authors 1 = There is an competition authority but a dependent ministry of the executive files complains; 0 = In other case Cartel per se illegal cartperse Authors 1 = Cartels are per se illegal; 0 = Cartels are adjudicated using the rule of reason Published guidelines for cartel cartelguide Authors 1 = Competition authority has published some enforcement cartel guidelines; 0 = Otherwise Criminal sanctions penalcart Authors 1 = Cartels are criminal felonies; 0 = In other case Punitive damages danoscart Authors 1 = It is possible to claim for punitive damages; 0 = Otherwise Leniency programs leniency Authors 1 = There is a leniency program; 0 = Otherwise 75 The drivers of antitrust effectiveness Table 1 VARIABLES IN THE DATASET (Continued) Name Acronym Source Description Dominance abuses per se illegal dompos Authors 1 = Per se for dominant position; 0 = Rule of Reason for dominant position Dominance defined by market thresdom GCR 1 = It exist a threshold obove which a position is share qualified as dominant; 0 = In other case Dominance threshold levthresd GCR 1 = If it exists, level of the threshold dominant position; (0 or 20 to 70%) 0 = In other case Published merger guidelines mergerguide Authors 1 = Competition authority has published some merger guidelines; 0 = Otherwise Government has the last say findecmerg Authors 1 = Government has the last say on mergers; mergers 0 = Competition authority takes decisions regarding mergers Protecting competition in objectimerg Authors 1 = Legal mandate for merger control is protecting merger law competition; 0 = It is protecting the public interest in general Per capita GDP cgdp Penn GDP (current US dollars), 2003 Tables Table 2 DESCRIPTIVE STATISTICS Acronym Obs Mean Std. Dev. Min Max Antitrust effectiveness effectiv 47 4.66 0.89 2.8 6.1 Independence on antitrust decisions indcart 47 0.94 0.25 0 1 Government prosecution minist 47 0.13 0.34 0 1 Cartel per se illegal cartperse 47 0.34 0.48 0 1 Published guidelines for cartel cartelguide 47 0.23 0.43 0 1 enforcement Criminal sanctions penalcart 47 0.36 0.49 0 1 Punitive damages danoscart 47 0.23 0.43 0 1 Leniency programs leniency 47 0.47 0.50 0 1 Dominance abuses per se illegal dompos 47 0.32 0.47 0 1 Dominance defined by market share thresdom 47 0.70 0.46 0 1 Dominance threshold (0 or 20 to 70%) levthresd 47 28.36 21.15 0 70 Published merger guidelines merguide 47 0.62 0.49 0 1 Government has the last say mergers findecmerg 47 0.43 0.50 0 1 Protecting competition in merger law obtectimerg 47 0.91 0.28 0 1 Authority Independence ind 47 0.87 0.28 0 1 Active stance in cartel policy cartel 47 1.64 1.24 0 5 Economic approach in dominance law dominance 47 –0,74 0.56 –1,75 0 Competition focussed merger policy merger_pol 47 2.11 0.87 0 3 Per capita GDP cgdp 47 19,164.81 9,111,38 3,212,53 37,313.33 EU-15 eu15 47 0.30 0.46 0 1 EU-Enlargement 2004 eu15-25 47 0.21 0.41 0 1 Bulgaria & Romania eu-25-27 47 0.04 0.20 0 1 76 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ Table 3 DISTRIBUTION OF OBJECTIVE FEATURES OF COMPETITION POLICY Variable Value Countries for which the variable takes the value shown Independence on antitrust decisions 0 Colombia, Costa Rica, Venezuela Government prosecution 1 Belgium, France, India, Latvia, Spain, Malta Cartel per se illegal 1 Australia, Austria, Chile, Costa Rica, Slovenia, France, Greece, Hungary, Italy, Korea, Latvia, Mexico, Romania, South Africa, USA, Venezuela Published guidelines for cartel 1 Canada, Greece, Netherlands, Ireland, Japan, Korea, enforcement Malta, New Zealand, Sweden, United Kingdom, USA Criminal sanctions 1 Canada, Czech Republic, Estonia, France, India, Ireland, Island, Israel, Japan, Korea, Mexico, Norway, Slovak Republic, Slovenia, Taiwan, United Kingdom, USA Punitive damages 1 Canada, Croatia, Czech Republic, Hungary, New Zealand, Norway, Slovak Republic, Slovenia, Taiwan, USA Leniency programs 1 Canada, Cyprus, Czech Republic, Finland, France, Netherlands, Hungary, Ireland, Korea, New Zealand, Norway, Poland, Romania, South Africa, Slovak Republic, Sweden, United Kingdom, USA Dominance abuses per se illegal 1 Costa Rica, Croatia, Czech Republic, Denmark, France, Greece, Hungary, India, Ireland, Italy, Latvia, Mexico, Romania, South, Venezuela Dominance defined by market share 0 Australia, Belgium, Chile, Cyprus, Colombia, Costa Rica, Finland, Japan, Kenya, Luxemburg, Mexico, New Zealand, Portugal, Romania, Spain, Venezuela Published merger guidelines 1 Australia, Austria, Belgium, Brazil, Canada, Czech Republic, Finland, Germany, Greece, Ireland, Japan Korea, Latvia, Lithuania, Malta, Netherlands, New Zealand, Norway, Poland, Romania, South Africa, Slovak Republic, Spain, Sweden, Switzerland, Taiwan, United Kingdom, USA, Venezuela Protecting competition in merger law 0 Argentina, Poland, Portugal, Taiwan Government has the last say on mergers 1 Argentina, Belgium, Cyprus, Colombia, Costa Rica, Finland, France, Germany, Greece, India, Israel, Italy, Malta, Norway, New Zealand, Poland, Portugal, Romania, Spain, Taiwan OCDE countries in italics, and EU countries bolded. Our aim is to identify and quantify the impact of a set of policy features on mean antitrust effectiveness. We use the World Economic Forum (WEF) “effectiveness of antitrust policy” as the dependent variable that needs to be explained. Table 2 shows that antitrust effectiveness ranges from 2.8 to 6.1, and it also shows that the sample average is 4.66. Firstly, we will estimate the impact of all the detailed qualitative coded information on the effectiveness of antitrust features controlling per capita income and EU membership. We do not expect to obtain precise estimates from this first empirical specification, because sound policy is rarely based on single feature of a policy, but we expect the data to highlight some salient characteristics. Secondly, we will construct four indicators based on what the literature has 77 The drivers of antitrust effectiveness identified as sound antitrust policy; they are the independence of the authority, an active stance with respect to cartel policy, an economic approach to dominance law and competition focused merger policy. We then estimate how important having a sounder policy design in each domain is for overall antitrust effectiveness, again controlling for per capita income and EU membership. Thirdly, we group the averages of these four policy domain indicators into two summary indicators using weights based on the principal components analysis. This type of analysis allows us to identify to what extent sound policy in one domain correlates with sound policy in another, but it is uncorrelated with sound policy in yet other domains. This can help us to understand the transitional path to more effective antitrust within the cross-section. Income is measured by using GDP per capita in 2003 in US dollars; this is the most recent data available for all the countries in the Penn World Table sample.5 Table 2 shows that income ranges from US$3,212 in India to US$37,313 in the USA; average income is US$19,164. EU countries constitute 55% of our sample and table 2 also shows that 30% are EU15 countries, and 21% are the new EU members. As stated before, we include income in all the estimations, because we want a control set for the omitted variables that are heavily correlated with income; examples are the general quality of public policies. Voigt (2006) has already shown that it is difficult to separate the effects of better antitrust from the quality of other policies. Countries with higher incomes would be expected to have better staffed and better resourced antitrust authorities. Those countries probably have better detection, sanctioning infrastructure and technologies. Additionally, as we expect antitrust to spur income, we have estimated the coefficients using instrumental variables techniques. By including income per capita, we also take into account a set of national characteristics that are strongly correlated with income per capita, and make antitrust enforcement easier; examples are the degree of rule of law in the country, the policy measures available to control corruption, the overall quality of economic regulation, the mechanisms in place to make politicians accountable, the efficiency of the legal system, the intensity of competition in local markets, the degree of firm and corporate power decentralization and the openness to international trade. Using data from the World Bank Institute, the World Economic Forum and Transparency International we have computed coefficient correlations between per capita income and these types of national characteristics ranging from 91% to 60%, see table 4. Therefore, controlling just for income allows us to have a parsimonious specification. Most of these income related covariates would also be endogenous; however, we prefer to only include conveniently instrumented income, in order to get unbiased and consistent estimates, rather than including those endogenous covariates in the regression analysis. Finding the appropriate instruments for all these covariates would have proved very difficult, and estimates could easily have turned biased and inconsistent. The objective features of competition policy across countries come from the information provided by CUTS International, Global Competition Review, the International Competition Network, the annual OECD reports and the national government agencies web pages and sta- 84 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ Table 10 shows our inference of the impact on antitrust of all the distinctive objective competition policy features. It details the main drivers of the previously outlined results, although the large number of covariates makes our results slightly less precise. However, as previously mentioned, the correlation among the covariates is not so strong as to make the estimates inaccurate. Table 10 IMPACT OF DETAILED FEATURES OF COMPETITION POLICY OF EFFECTIVENESS log(effectiveness) (1) OLS (2) OLS (3) IV (4) IV Intercept –0.89*** (5.50) –0.87*** (8.89) –2.66*** (6.18) –1.92*** (5.07) Independence on antitrust decisions 0.26*** (3.13) 0.26*** (4.05) 0.10* (1.70) 0.19*** (3.15) Government prosecution –0.03 (0.43) –0.03 (0.56) –0.01 (0.24) –0.01 (0.22) Cartel per se 0.02 (0.30) 0.03 (0.58) 0.00 (0.06) 0.01 (0.29) Cartel guide 0.02 (0.41) –0.04 (0.81) –0.04 (0.93) –0.06* (1.73) Criminal sanctions 0.07 (1.22) 0.07 (1.19) 0.09*** (2.12) 0.07 (1.64) Punitive damages –0.05 (0.75) –0.01 (0.18) –0.09* (1.77) –0.05 (0.88) Leniency damages 0.06 (1.16) 0.06 (1.37) 0.06 (1.55) 0.07** (2.06) Per se dominance abuses rules –0.12 (1.61) –0.13** (2.23) –0.05 (1.11) –0.08** (1.98) Dominance defined by market share –0.24*** (2.57) –0.20*** (2.66) –0.11 (1.36) –0.14* (1.88) Dominance threshold (0 or 20 to 70%) 0.005*** (2.54) 0.004** (2.25) 0.002 (1.04) 0.003* (1.73) Published merger guidelines 0.06 (0.96) 0.05 (1.19) 0.04 (1.04) 0.04 (1.34) Government has the last say on mergers 0.07 (1.16) 0.03 (0.70) 0.03 (0.82) 0.02 (0.72) Protecting competition in merger law 0.19 (1.45) 0.19** (2.35) 0.15* (1.78) 0.17** (2.56) Log per capita GDP 0.20*** (4.20) 0.12*** (2.81) EU-15 0.11** (2.06) 0.04 (1.02) EU-Enlargement 2004 –0.11* (1.93) –0.12*** (2.74) Bulgaria & Romania –0.29*** (5.39) –0.22*** (4.38) R2 0.53 0.75 0.68 0.78 F test F(13,33)=12.96*** F(16,30)=27.14*** F(14,32)=11.37*** F(17,29)=33.20*** Hansen Over-id J test Chi-sq(1)=1.44 Chi-sq(1)=0.46 1. The table report the t statistics robust to heterocedasticity in parenthesis. The table reports the values of the F test and the Hansen Overid test. *, **, ***: for 10%, 5%, and 1% significance level. 2. The null hypothesis of the Hansen J overidentification test is that the instruments are orthogonal with respect to the residual of the regression. Columns 3 and 4 show that the test value is smaller than the critical values, so we do not reject the null hypothesis that the instruments are valid instruments. Table 10 shows that the independence of the authority when adjudicating cases drives antitrust effectiveness. The authority’s independence when adjudicating cases of cartel and dominance does make antitrust more effective even taking into account per capita GDP, EU membership and the other covariates. On average, Colombia, Costa Rica and Venezuela have worse competition policy; these are the three countries in the sample that lack an independent authority. Having governmental power of investigation has a negative but not statistically significant impact on effectiveness; here the relevant issue appears to be having an independent authority calling the shots. We have previously seen that active stances on cartel policy in isolation doesn’t have a significant impact on effectiveness, but leniency programmes do stand out as having an im- 85 The drivers of antitrust effectiveness pact in isolation. Countries that have recently implemented innovative leniency programmes are the ones with more effective antitrust policies. Interestingly, it appears to be unimportant whether cartels are prohibited using either the rule of reason or the per se illegality rule. This result suggests that, as Borrell (2007) stated, there is no single legal rule that fits all countries in the domain of cartel policy. Or alternatively, that actual enforcement of cartel policy is quite similar across countries irrespectively of the formal legal rule. Table 10 also shows why the approach to dominance policy is important for antitrust effectiveness. Countries that enforce dominance policy using a per se illegality rule have a significantly worse antitrust effectiveness of around 8%. Countries that set a fixed market share to define dominant positions also have a worse competition policy effectiveness of around 14%. However, antitrust effectiveness improves, as the market share that defines the dominant position gets larger. Finally, what appears to be important in merger policy is to have a competition orientated legal mandate, rather than having a government or agency that decides on mergers. On average, countries that aim to protect competition when authorising mergers have antitrust policies that are 17% more effective. 5. Concluding remarks The effectiveness of antitrust policy varies strongly across countries. This paper shows that different institutional and policy characteristics drive the effectiveness of antitrust, and concludes that antitrust cannot be effective in a vacuum. Competition policy effectiveness is driven by per capita GDP and EU membership. It also shows how European Eastern countries are finding it particularly difficult to enforce effective antitrust policies. Antitrust depends on income and EU membership, but is also more effective in some countries than in others. Having an economic approach to dominance law clearly makes antitrust more effective. Investigating dominance using an economic approach avoids the per se illegality rules and underscores market share as potential source of market power; it also makes antitrust sounder to have a competition orientated merger policy. Furthermore, engaging in innovative policies such as leniency makes antitrust more effective. Finally, authority independence regarding decisions on prohibiting cartels and dominant position abuses seem to be a prerequisite to sound antitrust; but it is not so important whether an independent agency or the government conducts the antitrust investigations. Notes 1. See Porter, M. E. et al. (2005). 2. See IMD World Competitiveness Yearbook in http://www.imd.ch/research/publications/wcy/index.cfm. 3. See Mehta (2006). 4. See Davis (2006), Rowley and Low (2006), and Wessely (2006). 86 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ 5. See Heston, Summers and Aten (2006). We have not been able to use the 2004 data on per capita income because the per capita income of Brazil, Colombia, India and Latvia were missing from the Penn World Table 6.2 at the time of writing this paper. 6. See Mehta (2006), Davis (2006), Rowley and Low (2006), Wessely (2006), ICN, OCDE and nacional competition authorities websites. 7. The Hansen J overidentification test is the GMM equivalent to the Sargan test. It is a test robust to the presence of heterocedasticity. The joint null hypothesis is that the instruments are valid instruments: the instruments are orthogonal with respect to the residuals of the regression. Under the null, the test statistic is distributed as chi-squared in the number of overidentifying restrictions. A rejection casts doubt on the validity of the instruments. The tables show that the test values are always smaller than the critical values, so we do not reject the null hypothesis that the instruments are valid instruments. See Baum, Schaffer and Stillman (2007) References ABA Section of Antitrust Law (2003), “State Antitrust Enforcement Handbook”, American Bar Association - Section of Antitrust Law, Chicago. Ahlborn, C.; D. S. Evans and A. J. Padilla (2004), “The antitrust economics of tying: a farewell to per se illegality”, Antitrust Bulletin, 49(1&2), 287-341. Aghion, P. and M. Schankerman (2004), “On the welfare effects and political economy of competition-enhancing policies”, The Economic Journal, 114, 800-824. Aghion, P. and P. Howitt (1996), “A model of growth through creative destruction”, Econometrica, 60, 323-51. Baker, J. B. (2003), “The Case for Antitrust Enforcement”, Journal of Economic Perspectives, 17(4), 27-50. Barros, P. P. (2003), “Looking behind the Curtain - Effects form Modernization of European Union Competition Policy”, European Economic Review, 47, 613-624. Baum, C. F.; M. E. Schaffer, S. Stillman (2007), “ivreg2: Stata module for extended instrumental variables/2SLS, GMM and AC/HAC, LIML and k-class regression”, http://ideas.repec.org/c/boc/bocode/s425401.html. Berges-Sennou F.; F. Loss; E. Malavolti-Grimal and T. Verge (2002), “Modernisation de la politique communautaire de concurrence: Regime d’ autorisation ou d’exception legale? (European Competition Policy Reform: Autorisation or Exemption Regime?)” Revue Economique, 53, 437-47. Bertrand, O. and M. Ivaldi (2006), “European Competition Policy in International Markets”, IDEI Working Paper, 419, November. Borrell, J. R. (2008),: “Fashioning Antitrust Regimes”, in Wayne D. Collins (ed.) Issues in Competition Law and Policy, vol. 1, chapter 3, ABA Section of Antitrust Law, forthcoming. Borrell, J. R. and M. Tolosa (2008), “Endogenous antitrust: Cross-country evidence on the impact of competition-enhancing policies on productivity”, Applied Economics Letters, forthcoming. Borrell, J. R. and J. L. Jiménez (2007), “The drivers of effectiveness in competition policy”, Documento de trabajo de FUNCAS, 363/2007. Carlton D. W. and R. c. Picker (2007), “Antitrust and Regulation”, NBER Working Paper, 12902. 87 The drivers of antitrust effectiveness Conway, P. and G. Nicoletti (2006), “Product Market Regulation in the Nonmanufacturing Sectors of OECD Countries: Measurement and Highlights”. OCDE. Economic Department Working Paper, 530. Cubbin, J. and J. Stern (2006), “The Impact of Regulatory Governance and Privatization on Electricity Industry Generation Capacity in Developing Economies”, World Bank Economic Review, 20(1), 115-141. Davis J. (Ed.) (2006), Merger Control. The international regulation of mergers and joint ventures in 61 jurisdictions worldwide, published by Global Competition Review – The Official Research Partner of the International Bar Association, London. Demouguin, D. and C. Fluet (2004), “Deterrence vs Judicial Error: a Comparative View of Standards of Proof”, Centre Interuniversitaire sur le Risque, les Politiques Économiques et l’Emploi, Working Paper 04-18. Dutz, M. A. and A. Hayri (2000), “Does more intense competition lead to higher growth?”, World Bank Policy Research Working Paper, 2320 and CEPR Discussion Paper, 2249. Gal, M. A. (2003), Competition Policy for Small Market Economies, Havard University Press, Cambridge, MA. Gual, J.; M. Hellwig; A. Perrot; M. Polo; P. Rey; K. Schmidt and R. Stenbacka (2005), “An Economic Approach to Article 82”, Report of the EAGCP. Gutiérrez, L. H. (2003), “The effect of endogenous regulation on telecommunications expansion and efficiency in Latin America”, Journal of Regulatory Economics, 23(3), 257-286. Heston A., R. Summers and B. Aten (2006), Penn World Table Version 6.2, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, September. Kaplow, L. and S. Shavell (1994), “Accuracy in the determination of liability”, Journal of Law and Economics, 37, 1-15. Kaplow, L. and C. Shapiro (2007), “Antitrust”, NBER Working Paper, 12867. Levine P.; J. Stern and F. Trillas (2005), “Utility price regulation and time inconsistency: comparisons with monetary policy”, Oxford Economic Papers, 57(3), 447-478. Mehta, P. S. (Ed.) (2006), Competition Regimes in the World. A Civil Society Report, CUTS (Consumer Unity and Trust Society) international – CUTS Centre for Competition, Investment an Economic Regulation, Jaipur. Motta, M. (2004), Competition policy. Theory and practice, Cambridge University Press, Cambridge. Nicholson, M. W. (2004), “Quantifying Antitrust Regimes”, FTC Working Paper, 267, Washington DC. Nicoletti, G.; S. Scarpetta and O. Boylaud (2000), “Summary Indicators of Product Market Regulation with an Extension to Employment Protection Legislation”. OCDE. Economic Department Working Paper, 226, april. Polinsky A. M. and S. Shavell (1989), “Legal error, litigation, and the incentive to obey the law”, Journal of Law, Economics and Organization, 5, 99-108. 88 JOAN-RAMON BORRELL AND JUAN LUIS JIMÉNEZ Porter, M. E. et al. (2005), The Global Competitivenes Report 2004-2005 CD-ROM. Results of the Executive Opinion Survey 2004, World Economic Forum, Palgrave Macmillan, Hampshire. Rowley J. W. and M. Low (2006) (Ed.), Cartel regulation. Getting the fine down in 31 jurisdictions worldwide, published by Global Competition Review – The Official Research Partner of the International Bar Association, London. Serebrisky, T. (2004), “What do we know about competition agencies”, World Bank Policy Research Working Paper 3221, Washington DC. Stern, J and F. Trillas (2003), “Independence and discretion in telecommunications regulation: lessons from independent central banks”, Utilities Policy, 11(4), 191-201. Stern J. and J. Cubbin, J. (2005), “Regulatory effectiveness: the impact of regulation and regulatory governance arrangements on electricity industry outcomes”, World Bank Policy Research Working Paper, N. 3536., Washington DC. Voigt S. (2006), “The economic effects of competition policy: cross-country evidence using four new indicators”, Social Science Research Network Working Paper, 925794. Available at http://ssrn.com/abstract=925794. Wessely, T. (2006). (Ed.), Dominance. The regulation of dominant firm conduct in 37 jurisdictions worldwide, published by Global Competition Review – The Official Research Partner of the International Bar Association, London. Resumen Este trabajo muestra qué guía la efectividad de la política de competencia utilizando una base de datos internacional sobre la percepción de la efectividad de dicha política. El estudio concluye que la efectividad media depende de la renta per cápita y de liderazgos supranacionales como el ejercido desde el núcleo de la Unión Europea. Adicionalmente, el trabajo muestra que algunos aspectos del diseño de la política tienen un impacto significativo en sus resultados. Así, la efectividad mejora con una aproximación económica a los casos de dominancia y prácticas abusivas. Se demuestra que la efectividad mejora cuando el mandato legal para la política de concentraciones se centra en la competencia en los mercados, más que en el interés general. Además, una política de competencia más proactiva contra los carteles mejora la efectividad y, en particular, son muy valiosos los programas de clemencia. Finalmente, es importante que la autoridad independiente de la competencia tenga la última palabra en cuanto a la prohibición de prácticas restrictivas de la competencia. Palabras claves:Política de competencia; Efectividad de la política; Económica política. Clasificación JEL: D7; L4; O4.