How Effective is European Merger Control EC Competition

  • Slides: 24
Download presentation
How Effective is European Merger Control? EC Competition Enforcement Data Amsterdam, April 9 -10

How Effective is European Merger Control? EC Competition Enforcement Data Amsterdam, April 9 -10 2008 Tomaso Duso (Humboldt University and WZB) Klaus Gugler (University of Vienna) Burcin Yurtoglu (University of Vienna) Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 1

Introduction § Economic evaluation of EU merger control decisions. § Research questions: Did merger

Introduction § Economic evaluation of EU merger control decisions. § Research questions: Did merger control achieve the objective of restoring effective competition? Are remedies the best instrument? § We use stock market reactions as an (ex-ante) independent assessment of the concentration as well as the merger control procedure. § Identification assumption: anti-competitive rents generated by the merger should be dissipated by the commission’s decision, if this is effective. § We apply this approach to a sample of 151 mergers scrutinized by the European Commission between 1990 - 2002. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 2

Measuring Effectiveness Idea & Methodology § § § Each merger has two possible effects:

Measuring Effectiveness Idea & Methodology § § § Each merger has two possible effects: § Market power increase (positive for both merging firms and rivals). § Efficiency gains (positive for merging firms, negative for rivals). An effective antitrust decisions should maintain the benefits to consumers generated by increased efficiency and, at the same time, reduce the market power effects of the merger, i. e. all rents generated by a market power increase should be reversed by an effective antitrust decision. Hence, we measure antitrust effectiveness by: 1. Measuring the rents generated by the merger and by the antitrust decision. 2. Relate the two by means of regression analysis. We expect a negative relation between the two. This relation should change depending on the kind of decision. § Main results: Prohibitions perfectly restore competition. Remedies are not always effective. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 3

1. Measuring Rent I § We use information from the financial markets to measure

1. Measuring Rent I § We use information from the financial markets to measure the profitability of a merger as well as the effect of the Commission’s decision. § The event study methodology looks at how stock prices of firms involved in the merger (merging firms and rivals) react to a particular event (e. g. merger announcement, commission‘s decision etc. ). § We measure abnormal returns as the exceptional returns (compared to the market) that a firm realizes around a particular event. These measures should capture the market’s valuation of the event’s effect. § Which events do we use? Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 4

1. Measuring Rent III The EU Merger Control Merger’s effect Duso-Gugler-Yurtoglu Antitrust decision’s effect

1. Measuring Rent III The EU Merger Control Merger’s effect Duso-Gugler-Yurtoglu Antitrust decision’s effect Effectivness of EU Merger Control 5

1. Measuring Rent II Correcting for Expectations § The observed abnormal returns entail the

1. Measuring Rent II Correcting for Expectations § The observed abnormal returns entail the real merger/decision effect but also the market’s prior/update about the antitrust action. § In a first step we estimate the probability of an action by using observable mergers’ characteristics. § We then use these probabilities to correct the estimated abnormal returns in order to obtain clean measures of the merger/decision effect : Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 6

2. The Empirical Implementation § Suppose the agency does not make mistakes. All anticompetitive

2. The Empirical Implementation § Suppose the agency does not make mistakes. All anticompetitive rents should be eliminated by an effective decision (rent reversion). § To measure policy effectiveness we run a basic linear regression of announcement effect on decision effects for merging firms (i=M) and rivals (i=R): constant slope § The a and b-coefficients measure the degree of market power reversion due to the Commission’s decision. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 7

2. Predictions: Rivals PD* BLOCK (Panticomp + Pefficiencies) REMEDIES A 1 = B 1

2. Predictions: Rivals PD* BLOCK (Panticomp + Pefficiencies) REMEDIES A 1 = B 1 A 2 PA* - (Panticomp + Pefficiencies) - (Panticomp) B 2 R 1 R 2 Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 8

2. Predictions: Summary Effective Merger Control Rivals Merging firms a b Blocking 0 -1

2. Predictions: Summary Effective Merger Control Rivals Merging firms a b Blocking 0 -1 Remedies <0 (-1, 0) 0 0 (other remedies and structural) Clearance Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 9

The Data § 151 mergers analyzed by the EU Commission between 1990 and 2002.

The Data § 151 mergers analyzed by the EU Commission between 1990 and 2002. We identify 544 different firms involved in these mergers either as merging entities or rivals. § Almost all Phase II cases (71) and a random sample of Phase I cases (80). § Sources: – EU decisions (rivals, decisions and merger-specific information) – Dow Jones Interactive (announcement date) – Datastream (stock market reactions) – Compustat Global (accounting data) Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 10

Results: Full Sample 1. Prohibitions restore on average effective competition (b. RB=-0. 88 for

Results: Full Sample 1. Prohibitions restore on average effective competition (b. RB=-0. 88 for rivals and b. MB=-0. 72 for merging firms, not statistically significantly different from -1). 2. Blockings are a significant cost for merging firms (a. MB=-0. 21, significant). 3. We cannot reject the hypothesis of full profit reversion for blocking decisions: This constitutes a consistency check for our approach. 4. On average, remedies are not completely successful in restoring effective competition. Our predictions for the merging firms are met, yet they are only partially met for rivals. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 11

Results: Qualification 1. Remedies are, though only partially, effective when they are applied in

Results: Qualification 1. Remedies are, though only partially, effective when they are applied in phase 1. 2. In phase 2, remedies seem less effective. Structural remedies might be seen as a rent transfer from merging firms to rivals. 3. Anticompetitive mergers that are cleared in phase 1 are good news for the rivals. There is no effect for merging firms: Type II errors benefit rivals and don’t hurt merging firms. 4. (Structural) remedies (correctly) applied to anticompetitive mergers have a strong and significantly negative effect on rivals are more effective 5. Remedies applied in pro-competitive mergers have negative and significant effect on merging firms and no effect on rivals: (weak) type I error constitute a cost for merging firms Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 12

Results: Robustness 1. The results remain qualitatively the same also excluding mergers with vertical

Results: Robustness 1. The results remain qualitatively the same also excluding mergers with vertical and/or conglomerate effects. 2. The results are robust to merger wave arguments (1990 -96 vs. 1996 -2002, merger wave industries vs. non merger wave industries). 3. The remedies’ effectiveness is substantially increased in remedies-intensive industries. This suggests that the Commission has learnt over time and in certain industries to implement effective remedies. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 13

Results: Robustness III Ex-Post Evaluation § We use balance-sheet data and compare actual profit

Results: Robustness III Ex-Post Evaluation § We use balance-sheet data and compare actual profit levels two years after the merger with a counterfactual given by the development of profits in the same 3 digit industry as the merging firms or their rivals (Gugler et al. , 2003). § We find a significantly positive relationship between the ex post profit effects and the announcement CAARs and total CAARs. § We then relate the profit effects for the rivals to the merging firms’ profit effects. We do not find a significant relation between the two effects for mergers cleared unconditionally or blocked. We find a significant positive relation for those mergers that were cleared with commitments. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 14

Conclusions § Financial market information (event studies) are useful to assess competition policy effectiveness:

Conclusions § Financial market information (event studies) are useful to assess competition policy effectiveness: § They allow to measure both the merger’s and antitrust decision’s effects separately. § They are easy to implement and require a limited amount of information. § Important to correct for the market’s prior about the antitrust action. § Crucial is to look how these effects are related. § Methodology produces consistent and robust results: § Blockings indeed restore the pre-merger situation. § Remedies are on average only partially effective. § Remedies are more effective when applied in phase 1 and when applied to anticompetitive mergers § Our results are robust to several sub-sampling and they are supported by an alternative methodology based on ex-post evaluation methods. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 15

Back-up slides Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 16

Back-up slides Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 16

Correcting for Market Expectations • The observed abnormal return around the announcement day (ΠA)

Correcting for Market Expectations • The observed abnormal return around the announcement day (ΠA) entails the real merger effect times the market’s prior about the probability of clearance: • The observed abnormal return around the decision day is the market update of the expected value of the Commission’s action: • In a first step we estimate the probabilities of clearance and action by using observable mergers’ characteristics. Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 17

Predictions: Merging firms PD* (Panticomp + Pefficiencies) A 1 A 2 PA* - (Panticomp)

Predictions: Merging firms PD* (Panticomp + Pefficiencies) A 1 A 2 PA* - (Panticomp) R 1 -(Panticomp + Pefficiencies) B 1 R 2 REMEDIES - (Panticomp + Pefficiencies) B 2 Duso-Gugler-Yurtoglu Effectivness of EU Merger Control BLOCK 18

The Main Variables Long run (50 days) cumulative abnormal returns around the merger’s announcement

The Main Variables Long run (50 days) cumulative abnormal returns around the merger’s announcement day. Sum of the cumulative abnormal returns around the Phase I and Phase II decisions (50 days for Phase II and 5 days for Phase I). § Decision: Clearance, Divestitures, Other remedies, Block. § Controls: Year and industry dummies, conglomerate and foreclosure dummies. § Anticompetitive: dummy equals 1 if the rivals’ CAARs around the announcement are positive (Duso, Neven, and Röller, 2007). Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 19

Results: Full Sample Dependent var. : CARs at decision PA Rivals Merging firms Coeff

Results: Full Sample Dependent var. : CARs at decision PA Rivals Merging firms Coeff St. Err. Coeff CLEAR 0. 042 0. 061 -0. 020 0. 093 OTHER REMEDIES 0. 013 0. 037 -0. 095 0. 056 STURUCTURAL 0. 049 0. 028 -0. 033 0. 045 BLOCK -0. 061 0. 041 -0. 213 0. 063 CLEAR * PA 0. 274 0. 076 0. 007 0. 090 OTHER REMEDIES * PA -0. 116 0. 093 -0. 341 0. 172 STURUCTURAL * PA -0. 105 0. 091 -0. 165 0. 077 BLOCK * PA -0. 875 0. 100 -0. 718 -0. 136 Duso-Gugler-Yurtoglu Effectivness of EU Merger Control St. Err. 20

Results: Phase I Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 21

Results: Phase I Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 21

Results: Phase II Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 22

Results: Phase II Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 22

Results: Pro- vs. Anti-competitive Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 23

Results: Pro- vs. Anti-competitive Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 23

Results: Remedy-Intensive Industries Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 24

Results: Remedy-Intensive Industries Duso-Gugler-Yurtoglu Effectivness of EU Merger Control 24