Determinantes da eficiência técnica no setor de supermercados brasileiro: uma análise utilizando métodos de fronteira de produção
[thesis]
Tricia Thaíse e Silva Pontes
Pontes, T. T. S. (2016). Technical efficiency determinants in the Brazilian supermarket sector: an analysis using methods of production frontier. Tese de Doutorado, Faculdade de Economia, Administração e Contabilidade, Universidade de São Paulo, São Paulo. This study investigates the determinant technical efficiency factors in the Brazilian supermarket sector based on the application of different production frontier methods that allow a comparison of performance between companies. A sample of
more »
... supermarket chains operating in Brazil between 2006 and 2015 was used. The research methodology was based on three approaches of estimation of the frontier of production: parametric, non-parametric and semiparametric approach. In the first one, the efficiency is estimated through the analysis of the stochastic frontier with panel data, and different specifications for production function. In this step, models with heterogeneity are also estimated in order to find the determinants of supermarket efficiency. Then the non-parametric frontier methods are classified as to the organization of the data, first the DEA-Malmquist was used for panel data, and then the traditional DEA technique for each year separately. Finally, the efficiency scores obtained by the DEA are organized as a panel and used as the dependent variable in the semiparametric approach, which estimates a Tobit model to find the determinants of efficiency. The results of all approaches point to the inefficiency of the Brazilian supermarket sector, which represents an alert for companies to improve their production processes. In addition, a drop in the sector's technical efficiency was identified in recent years, while the productivity showed a positive variation, largely due to technological change. The comparative analysis of the model results showed that the choice of the approach did not influence the construction of the efficiency ranking, on the other hand, the magnitude of the efficiency scores varies significantly between the approaches. SFA models with heterogeneity and two-stage DEA provided different results regarding efficiency determinants; however, in the comparison of the average efficiency it was possible to find evidence of the impact of size, market power and mergers and acquisitions on the supermarket inefficiency, confirming three of the hypotheses proposed in the research. These findings represent a significant contribution to literature on retail efficiency, which is still little explored in Brazil.
doi:10.11606/t.12.2017.tde-30032017-161808
fatcat:5fay4d5tube5zasb7ox5mw6aua