Delays in the construction of power plants from electricity auctions in Brazil
Description
This database was compiled gathering information published by the Brazilian Power Trade Chamber (CCEE), the Brazilian Electricity Regulatory Agency (ANEEL) and the Independent System Operator (ONS). Basic information on project characteristics, commercial operation date, and main PPA features were obtained from the combination of two CCEE's monthly updated publications: the “Consolidated Auction Results” (CCEE, 2021a) and the “Market Information Bulletin - Individual Project Data” (CCEE, 2021b). Additionally, we accessed auction notices and PPA drafts to verify, for each case, the auction design features employed, such as technological banding (technology-specific or multi-technology products), preliminary phase for project elimination by connection point and allocation of transmission delays risks to project developers. The resulting database contains, for each project, information regarding energy source, geographic location, installed capacity; date of effective start of commercial operation, project status (in operation, delayed or cancelled); auction type, date and relevant design features; contract modality (quantity or availability); start and end date of the contracted supply period; bidding price and average energy traded per year. We focus on auction rounds related to the construction of new power plants and for which deadlines had been met by December 2021. Regarding wind power and solar photovoltaic plants, based on information extracted from ONS (2022), 660 projects were grouped into 130 complexes, each considered as a single project, as they usually share the same facilities, developers, licensing and construction process. For projects that sold energy in more than one auction, we considered only the first auction in which the project was awarded a contract. In the end, 523 obs. remained in the grouped database , which comprises all bid winners, including projects that have been officially cancelled or those with extremely large delays and low probability of realization. For the econometric analysis developed in the paper, we restricted the sample in 4 steps. First, we dropped 83 fossil-fuel thermal plants and kept just the most relevant sources in terms of number of projects and installed capacity: Wind, Solar PV, Biomass and Hydropower (large and small) plants. Second, we excluded 50 officially cancelled projects. Third, we removed 28 projects with extremely small implementation times (less than one year), which does not seem compatible with construction times for the technologies. Finally, we filtered out 23 projects that have been either delayed or anticipated by more than 3 years from the auction deadline, which indicates high probability that the schedule of such projects is related to specificities beyond the project features and auction design elements under analysis. Thus, the final sample for statistical analysis contains 339 projects. We added a Dictionary with descriptions on the main variables of the databases.