To address the company-branch structure in data envelopment analysis (DEA), Färe and Zelenyuk (2021, Annals of Operations Research) proposed a sequential DEA (sDEA) framework, which accounts for both company-level management uniformity and the district specific influence on the production process from a branch-level perspective. To account for the varying levels of branch independence, in this work we extend the flexibility of sDEA by introducing two decentralized estimators, accommodating the variability of branch-level efficiency and frontier. By broadening the scope from spatial to temporal sequences, we also show how the sDEA estimators can provide solutions for panel data problems in the realm of DEA. Through an empirical application on the public hospitals in Queensland, Australia, we illustrate that the canonical and the decentralized sDEA estimators can be adapted to effectively address the efficiency estimation in panel data. In contrast, the conventional DEA estimators, overlooking the time effects, may yield underestimated hospital-level efficiency.
