1. Tomaso Duso (DICE – Henrich-Heine University) Measuring Competition Policy and its Effectiveness Paolo Buccirossi (LEAR) Lorenzo Ciari (European University Institute & LEAR) Tomaso Duso (DICE – Henrich-Heine University) Giancarlo Spagnolo (University of Rome Tor Vergata, SSE, & CEPR) Cristiana Vitale (LEAR) ACLE Conference – Competition Policy for Emerging Economies: When and How? May 20, 2011

2. Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 2 Tomaso Duso (DICE) Motivation competition policy ¤ Is competition policy socially beneficial? competition policy is ineffective  Crandall and Winston ’ s (JEP 2003) provocative paper claims that competition policy is ineffective deterrence  Baker (JEP 2003) and Werden (JEP 2003) disagree and point to the social benefits of competition policy as a deterrence mechanism competition policy (deterrence) broad policy ¤ The key empirical question is how to measure competition policy (deterrence) and how to identify the effectiveness of such a broad policy quantify  We propose a way to quantify the quality of competition policy effectiveness of competition policy  We provide a framework to measure the effectiveness of competition policy in providing higher efficiency and productivity legal institutions  We analyze the interactions and complementarities between competition policy and legal institutions

3. 1. Measuring the quality of competition policy Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 3 Buccirossi, P., L. Ciari, T. Duso, G. Spagnolo, and C. Vitale (2011), “ Measuring the Deterrence Effect of Competition Policy: The Competition Policy Indexes, ” Journal of Competition Law and Economics, 7, 165-204 Tomaso Duso (DICE)

4. Competition Policy: Definition and Objective competition policy ¤ The term competition policy refers to:  Competition legislation:  Competition legislation: A set of prohibitions and obligations including merger control provisions that firms have to comply with enforcement  Its enforcement: An array of tools for policing behavior and punishing any violation deter firms no over-deterrence ¤ Competition policy should deter firms from undertaking any behavior that reduces social welfare by distorting competition, while not frightening any behavior that improves social welfare (no over-deterrence) Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 4 Tomaso Duso (DICE)

5. Deterrence: Competition Policy Variables optimal level of deterrence sanctionsdetectionconviction errors ¤ The optimal level of deterrence is determined by 1) the size of the sanctions 2) the (perceived) probability of detection and conviction, and 3) the (perceived) probability of errors (Becker, JPE 1968; Polinsky and Shavell, JEL 2000) ¤ The following policy variables affect these three factors: formal independence  the formal independence of the CA with respect to political or economic interests degree of separation  the degree of separation between the adjudicator and the prosecutor quality of the law  the quality of the law on the books level of losssanctions  the level of loss (sanctions) that firms (and their employees) can expect to suffer as a consequence of a conviction type of investigative powers  the type of investigative powers held by the CA financial and human resources  the amount and quality of the CA ’ s financial and human resources (the budget and skills of the CA ’ s staff) Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 5 Tomaso Duso (DICE)

6. The Competition Policy Indexes (CPIs) tailored questionnaires13 jurisdictions country studies CAs ’ own websites ¤ We submitted a set of tailored questionnaires to the CAs in 13 jurisdictions and integrated them with information from the OECD country studies and from the CAs ’ own websites policy variables separately hard-core cartelsabusesother infringementsmergers 1995 to 2005 ¤ We obtained information on each of the six policy variables identified as determinants of deterrence, separately for each type of possible competition law infringement (hard-core cartels, abuses, other infringements) and for mergers over the years from 1995 to 2005 score/weight agreed best practice ¤ Each piece of information at each step of the aggregation process was assigned a score/weight on a scale of 0-1 against a benchmark of generally agreed best practice equal weightsrandom weightsfactor analysis ¤ We tested the sensitivity of this weighting scheme to alternative ones using 1) equal weights, 2) random weights, and 3) factor analysis Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 6 Tomaso Duso (DICE)

7. [Details] Details Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 7 Tomaso Duso (DICE)

8. Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 8 [Figure] Figure Tomaso Duso (DICE)

9. The CPI: Discussion legal framework institutional settingsenforcement tools ¤ The CPIs embody both formal and practical aspects of a competition policy regime by combining key information on the legal framework, the institutional settings, and the enforcement tools best practices ¤ This information is evaluated against a benchmark of best practices and then aggregated ¤ There is room for further refinement of the CPIs developing  Cover a longer time period, as well as more (developing!) countries level of the sanctions that are effectively imposed  Include more detailed data on the enforcement features (more extensive information on the level of the sanctions that are effectively imposed) transparencyaccountability assessment analysis ¤ Exercise serves as a tool of transparency and accountability for public enforcement agencies, as well as a starting point for an assessment analysis Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 9 Tomaso Duso (DICE)

10. 2. Measuring the effect of competition policy Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 10 Buccirossi, P., L. Ciari, T. Duso, G. Spagnolo, and C. Vitale (2011), “ Competition Policy and Productivity Growth: An Empirical Assessment, ” CEPR DP 7470 Tomaso Duso (DICE)

11. Theoretical Background deterring anti- competitive behavior ¤ The aim of Competition Policy is to protect social welfare by deterring anti- competitive behavior ¤ By making markets more competitive, competition policy affects static and dynamic efficiency through a variety of channels (e.g., Aghion and Schenkerman, EJ 2004; Aghion et al. QJE 2004, ReStat2009). The causal link is thus:  Competition Policy  Competition  Efficiency / Productivity direct link between competition policy efficiency ¤ Given our main research focus, we look at the direct link between competition policy and efficiency to identify and estimate the causal effect of the policy (e.g. Pavnick, RES 2002; Nicoletti and Scarpetta, EP 2003) ¤ The deterrence effect of competition policy also depends on external factors such as the institutional environment (Aghion and Howitt, JEEA 2006) judiciary system  We look at the complementarities between the policy and the judiciary system Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 11 Tomaso Duso (DICE)

12. The Empirical Application: The TFP Model ¤ As a measure of efficiency/productivity we use Total Factor Productivity (TFP) growth three-dimensional (country, industry, time) panel data ¤ The proposed specification is derived from a standard neoclassical aggregate industry production function Y ijt =TFP ijt F ijt (L ijt, K ijt ) (e.g. Griffith et al. REStat 2004). It is estimated with a three-dimensional (country, industry, time) panel data approach: productivity frontierproductivity gap where TFP Ljt is the TFP level in the country-industry on the productivity frontier, (TFP ijt /TFP Ljt ) represents the productivity gap to the technological frontier, X and Z are sets of control variables (R&D, PMR, human capital, trade openness, and the quality of institutions) and are country- industry and time fixed effects Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 12 Tomaso Duso (DICE)

13. Identification endogeneity ¤ The identification problem derives from the possible endogeneity of the CPIs because of omitted variables and two-way causality  Aggregate  Aggregate the features of the competition policy regime (institutional inertia)  Lag  Lag the potentially endogenous explanatory variables time invariant unobserved individual heterogeneity  Panel data allows us to control for time invariant unobserved individual heterogeneity controls  We include all possible controls based on the existing literature on the determinants of TFP growth endogeneity ¤ We take two further steps (more formal) to tackle the problem of potential endogeneity:  Instrumental Variables political and institutional variables policy in neighboring countries  Instrumental Variables (IV) approach: We use political and institutional variables (government type and its attitudes towards regulation, economic planning, market economy) policy in neighboring countries as instruments non-linearitiesbetter legal institutionsnot strongly regulated  Look at non-linearities: CP has a stronger impact in countries with better legal institutions and in industries which are not strongly regulated Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 13 Tomaso Duso (DICE)

14. TheData The Data 13 jurisdictions EU 1995 to 2005 ¤ We selected 13 jurisdictions (Canada, Czech Republic, France, Germany, Hungary, Italy, Japan, Netherlands, Spain, Sweden, the UK, EU, and US) over the years from 1995 to 2005 part of the EU ¤ For the countries that are part of the EU, we built a set of indexes that incorporate information on both national as well as EU competition policy regimes 22 industries ¤ For each country, our sample includes 22 industries based on the definitions of the International Standard Industrial Classification (ISIC) KLEMS consortium Groningen Growth and Development Center ¤ Data on TFP growth is drawn from the KLEMS consortium and from the Groningen Growth and Development Center STAN MEIPMR database ANBERD WGI ¤ Other data come from the OECD Structural Analysis (STAN) database, the OECD Main Economic Indicators (MEI) database, the OECD PMR database, the OECD Analytical Business Enterprise Research and Development (ANBERD) database, and the World Bank Worldwide Governance Indicators (WGI) database Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 14 Tomaso Duso (DICE)

15. Results positive impact statistically significant ¤ Competition policy has a positive impact on TFP growth, and this is statistically significant at the 1% level [Table] [Figure]TableFigure political variables policy in neighboring countries ¤ We can reject the hypothesis of the policy being endogenous by using political variables (government types and their ideological position) or policy in neighboring countries as instruments [Table]Table [] ¤ Using different weighting schemes for the Aggregate CPI, labor productivity instead of TFP, different lag structures, long run/aggregated effects, different sets of control variables, sub-samples, structures for the error terms does not change our qualitative results [Table]Table Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 15 Tomaso Duso (DICE)

16. institutions ¤ Controlling for institutions (contracts enforcement and quality of the judiciary/law) does not alter our results. Yet, good institutions have a positive impact on TFP growth [Table]Table interactionslegal institutions ¤ We explore the interactions between legal institutions and competition policy. CP has a significantly larger impact in countries with: enforcement of contracts  Low cost for the enforcement of contracts (Doing Business), rule of law  high rule of law (Fraser) manufacturing sectors ¤ We further look at the effect of competition policy in different industries. CP has a much stronger impact in the manufacturing sectors  Sector specific regulations  Sector specific regulations in services industries (e.g. electricity, gas, water, communication, financial intermediation) may make them to a lesser extent subject to the ex-ante antitrust scrutiny identification ¤ Both results improve our identification strategy Results - Heterogeneity Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 16 Tomaso Duso (DICE)

17. Conclusions dimensions of a good competition policy hard data accountable ¤ We identify important dimensions of a good competition policy, which can be quantified using hard data. Along these dimension, CAs should be kept accountable (keep track of their activities) institutional details  Especially in developing country, clear data on institutional details should be collected enforcement  Keep track of the policy ’ s enforcement (cases & sanctions)! significant and positive impact ¤ Good competition policy appears to exert a significant and positive impact on efficiency. But, results are based on OECD countries complementarities ¤ We find complementarities between good judiciary institutions and competition policy institutional design  Very important for institutional design. One size fits all is not a good policy Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 17 Tomaso Duso (DICE)

18. Results - Robustness Checks [Table] Table robustness checkspositive and significant ¤ To verify the robustness of our results we perform several robustness checks with which we consistently obtain a positive and significant effect of the CPI  Employ an alternative CPI in which information is aggregated using Factor Analysis  Use 1,000 sets of random weights and generate a distribution of coefficient estimates [Figure]Figure  Include the EU dimension of the policy  Use alternative measures of productivity (non corrected for mark-ups and labor productivity)  Use long-run rates of TFP growth  Test the sensitivity of our results to the exclusion of specific countries- industries [Figure]Figure robustness ¤ All results point to the robustness of our main findings Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 18 Tomaso Duso (DICE)

19. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 19

20. Tomaso Duso Fig 1: The Aggregate CPIs [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 20

21. Tomaso Duso Fig 1: TFP Growth and the Aggregate CPIs [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 21

22. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 22

23. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 23

24. Tomaso Duso Results – Low-level indexes [Table] Table Institutional CPI Enforcement CPI ¤ Both the Institutional CPI and the Enforcement CPI coefficient are positive and significant yet the former has a much larger impact Antitrust CPI Merger CPI ¤ Both the coefficient estimate for the Antitrust CPI and Merger CPI are positive and strongly significant, yet the former has a larger impact quality of the lawpowers held by the CAs ¤ A positive impact on the intensity of competition is suggested for the quality of the law and for the powers held by the CAs during the investigation resources ¤ Also the resources held by the CAs have a positive impact on TFP growth Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 24

25. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 25

26. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 26

27. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 27

28. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 28

29. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 29

30. Tomaso Duso [back]back Measuring Competition Policy and its Effectiveness 20 May 2011 | ACLE| 30