ABSTRACT
The study of the football transfer market is relevant to understanding the clubs’ managerial capacity when they trade their most important assets. So far, the research on football transfers has focused mainly on the movements between European clubs. Thus, it is necessary to expand the analysis to developing countries, particularly those with a strong football tradition. We study the Brazilian professional football transfer market by analysing the transfers between top Brazilian clubs from 2018 to 2020. We use a two-step Heckman correction to account first for variables that affect the probability of a transfer and later, the actual transfer fee. We include an innovative index to capture the players’ ability to perform according to the specialists’ expectations. Our results show that besides appreciating the specific characteristics of the player such as age and nationality, the Brazilian market also considers the capacity of a player to fulfill previous expectations when deciding who is transferred and for what price. Finally, the remaining time of the contract and the recent transfer history of the player also affect transfer movements in Brazil, along with the seller club’s performance.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The data supporting this study’s findings are openly available in Football-Transfers-BR at https://doi.org/10.1080/00036846.2022.2107989.
Notes
1 Cartola FC is a football fantasy game – played by 1,93 mi people in the first gameweek of 2022 - that gathers data on league games and ranks players by fictional prices based on their performances. Gomide and Gualberto (Citation2020) stored this data in a public repository.
2 The performance score of a player each round is a weighted average of goals, assists, tackles, fouls (committed and suffered), shots (missed and on goal), cards (yellow and red), offsides, missed passes, clean sheets and goals conceded (for defenders and goalkeepers), and difficult saves for goalkeepers.
3 The bias presented by van Ophem and Ruijg (Citation2015) is not accounted for in this research, as we do not weigh the probability of a transfer having its value disclosed.