1. Cartel screening in the Brazilian fuel retail market
- Author
-
Douglas Silveira, Silvinha Vasconcelos, Paula Bogossian, and Joaquim Neto
- Subjects
L41 ,Gaussian ,Autoregressive conditional heteroskedasticity ,Economics, Econometrics and Finance (miscellaneous) ,Retail market ,symbols.namesake ,Margin (machine learning) ,Benchmark (surveying) ,0502 economics and business ,Econometrics ,Economics ,ddc:330 ,Fuel retail market ,050207 economics ,L95 ,HB71-74 ,050205 econometrics ,Price dynamics ,05 social sciences ,Cartel ,Identification (information) ,Economics as a science ,C63 ,Cartel Screen ,symbols ,Sale price ,C22 - Abstract
We aim to evaluate two different econometric screens for identifying anti-competitive behavior in the fuel retail market: (i) The Markov-Switching GARCH (MS-GARCH) Models; (ii) The Local Gaussian Correlation (LGC) approach. Using the gasoline cartel judged and condemned in Brasilia as a benchmark, our results indicate that the LGC model, based on the correlation of the resale price margin and price variability, may provide a biased likelihood as well as an incorrect identification of cartel behavior over time. The MSGARCH model, based only on the log deviation of the average gasoline sales price, showed better accuracy in cartel detection.
- Published
- 2021