Merger Efficiency Gains: An Assessment of the French Urban Transport Industry Ariane Charpin & Joanna Piechucka Paris School of Economics - Université Paris 1 CRESSE, June 29, 2016 Thank you for the financial support provided by projet POLICRE - ANR-12-BSH1-0009-01. Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 1 / 14
Introduction Objective and motivation The aim of the paper is to evaluate ex post the merger that took place between two major transport groups in the French urban transport industry Merger efficiency gains are the key focus of the analysis Efficiency gains are often overlooked, both in ex ante analysis of competition authorities and in academic literature Central aspect to the economic motivation behind a merger Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 2 / 14
Introduction Preview of the results We study the effect of the merger on operating costs of transport operators that belong to the merging groups Our results suggest that the merger did not lead to efficiency gains for the merging parties A possible explanation is that the merger proved to be difficult to implement in practice Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 3 / 14
The industry Organizational background Local public authorities are responsible for organizing urban public transport They can provide the service or delegate this task to a private or public-private transport operator The relationship between the operator and the local authority is regulated through a contract Two main contract types: fixed-price and cost-plus Urban public transport in France is highly subsidized Award mechanism Until 1993: automatic renewal of contracts Sapin Law of 1993: mandatory competitive bidding Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 4 / 14
The merger Transport groups and merger timeline Transport groups Keolis (30%, 2010) Veolia Transport (26%, 2010) Transdev (16%, 2010) 3 smaller groups: RATP Dev, CarPostal and Vectalia Brief story of the merger Early 2009: Transdev ends its partnership with RATP July 2009: Transdev discloses its choice of Veolia Transport over Keolis The merger was authorized conditionally by the French Competition Authority in December 2010 and occurred in March 2011 Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 5 / 14
The merger Potential efficiency gains Knowledge sharing between local operators belonging the same group Improved managerial efficiency Spread of one group specific values to the other Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 6 / 14
Data 9-year panel of 131 bus networks in France provided by CEREMA, GART and UTP for the years 2005-2013 Variable Mean Std. Dev. Min. Max. N C Operating costs ( 000 euros) 9381.38 11213.22 378.89 78589.57 883 Y Seat-kilometers 198624.5 224811.2 3576 1245552 883 N Network length (km) 257.71 346.42 13 7274 883 wl Labor price 37.09 5.22 22.33 52.97 883 wm Material price 18.95 6.79 1.69 59.72 883 wo Overhead price 0.71 0.58 0.08 5.75 883 labor part 0.54 0.11 0.13 0.79 883 materials part 0.13 0.04 0.02 0.46 883 cost plus Contract type 0.09 0.29 0 1 883 Original database on participants in tenders for the years 2004-2013 An analysis of the level of competition in the industry is performed (not presented today) Indispensable for the construction of two of our control groups Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 7 / 14
Empirical Strategy Difference-in-differences approach is taken to study the impact of the merger on operating costs of operators of the merging group ln(c nt) = α 0 + α 1 Post t + α 2 Treat n + α 3 Post t Treat n + γ i X int + δ n + e nt Control groups - all containing only networks operated by Keolis, RATP Dev or CarPostal : All networks (C1) Networks where contracts were signed prior to the merger (C2) Networks where there is no competition (C3) Networks where Transdev and Veolia never submitted a bid (C4) i Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 8 / 14
Identification Parallel trends assumption Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 9 / 14
Identification Non-endogeneity of decision to merge The decision to merger was not linked to operating costs Transdev wanted to create a leading international group Political considerations for the choice of Veolia Transport Complementarity between the two groups Shareholder of Transdev also the main shareholder of the group to which Veolia Transport belonged We (strongly) believe the decision to merge was not endogenous Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 10 / 14
Main results DiD estimation Explanation C1 C2 C3 C4 Post Treat Post-merger for treatment 0.019 0.019 0.019 0.017 [0.024] [0.023] [0.035] [0.031] Post Post-merger for all 0.073 0.074 0.079 0.081 [0.017] [0.016] [0.027] [0.023] Treat Treatment 0.040 0.043 0 0 [0.024] [0.032] [.] [.] ln(y ) Output 0.31 0.30 0.25 0.24 [0.051] [0.049] [0.063] [0.059] ln(wl) Labor price 0.57 0.56 0.57 0.57 [0.040] [0.040] [0.056] [0.056] ln(wo) Overhead price 0.30 0.32 0.31 0.31 [0.024] [0.024] [0.036] [0.035] ln(n) Network size 0.040 0.038 0.053 0.052 [0.013] [0.013] [0.016] [0.015] Constant 2.40 2.57 3.12 3.24 [0.62] [0.60] [0.75] [0.69] Network fixed-effects Yes Yes Yes Yes N 763 733 461 501 R2-adjusted 0.82 0.83 0.83 0.83 Standard errors in brackets. p < 0.10, p < 0.05, p < 0.01 The monetary variables C, wl and wo have been normalized with respect to the factor price wm to ensure that the cost function is homogeneous of degree one in input prices. Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 11 / 14
Main results DiD estimation with heterogenous treatment effects Explanation C1 C2 C3 C4 Post Treat Post-merger for treatment 0.017 0.017 0.010 0.011 [0.025] [0.025] [0.039] [0.034] Post Treat CP Post-merger for treatment and cost-plus 0.020 0.064 0.067 0.064 [0.060] [0.038] [0.045] [0.040] Post Post-merger for all 0.073 0.075 0.087 0.088 [0.017] [0.016] [0.031] [0.025] Post CP Post-merger for cost-plus 0.0053-0.054-0.065-0.064 [0.067] [0.055] [0.070] [0.058] Treat Treatment 0.027 0.031 0 0 [0.027] [0.031] [.] [.] Treat CP Treatment and cost-plus 0.063 0.15-0.0044-0.0025 [0.064] [0.070] [0.047] [0.047] CP Cost-plus -0.057-0.10 0 0 [0.051] [0.047] [.] [.] Remaining controls Yes Yes Yes Yes Network fixed-effects Yes Yes Yes Yes N 763 733 461 501 R2-adjusted 0.82 0.83 0.83 0.83 Standard errors in brackets. p < 0.10, p < 0.05, p < 0.01 The monetary variables C, wl and wo have been normalized with respect to the factor price wm to ensure that the cost function is homogeneous of degree one in input prices. Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 12 / 14
A fiasco merger? A merger badly studied and badly conducted. French Court of Auditors (2016) The merger between two equals quickly proved to be a fiasco, because of differences in culture, intertia in decision making and internal rivalries between the executives of the two groups. Les Echos (2013) This marriage is one of the biggest failures of the previous 5 years: the shareholders did not agree on anything, the managers disagreed on everything, clients were neglected and employees were forgotten. Le Nouvel Observateur (2012) Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 13 / 14
Conclusion and extenstions Econometric results show that the merger did not lead to efficiency gains for the merging parties Possible extensions Complementing the analysis with new years Differentiating the effect of the merger Studying the effect of the merger on subsidies Ariane Charpin & Joanna Piechucka (PSE) Merger Efficiency Gains June 30, 2016 14 / 14