The use of propensity score matching methodology for analysis of concession: The case of road networks in Peru

In this paper, a comparison has been made between the average crash rate, cost overruns and time overruns of the road sections concessioned with the estimated level that these sections would have if they had not been concessioned for the Peruvian case. For this exercise, the Propensity Score Matching methodology has been used that guarantees correct comparability between the concessioned and non-concessioned highways. This technique matches the unique characteristics that distinguish the control (to try to make them more similar) and treatment groups. The main result is that the crash rate, number of injuries, number of deaths, cost overruns and time overruns are lower in the concessioned sections than in the non-concessioned sections. In addition to this, the cost of crashes on concessioned and non-concessioned highways for Peru has also been estimated. In this case, the average annual cost per crash on the concessioned highways in the period 2015–2019 would be USD 65.72 million, while for the non-concessioned highways it would be USD 254 million. In other words, if all highways were concessioned, Peru would save an average of US $ 189 million per year from traffic crashes.