Efficiency, productivity and returns to scale in ports: a comparison of data envelopment analysis and econometric estimation with application to Caribbean Small Island Developing States

Abstract

The paper investigates for the first time the evolutionary technical efficiency and productivity of ports in Small Island Developing States (SIDS). The focus is primarily on Caribbean SIDS, benchmarked against two comparator groups, major top ports and ports in other SIDS in the Pacific and Indian Oceans. Both a non-parametric DEA and the previously under-utilised econometric COLS Malmquist Productivity Index (MPI) model are used. The results reveal a significantly lower level of technical efficiency in the Caribbean when assessed against top ports, but higher when compared to other SIDS. Caribbean ports achieved a higher rate of productivity improvement compared to top ports, and thus partially closed the efficiency gap, but not as much as ports in other SIDS. This was primarily the result of higher returns to density due to the rising traffic levels of the latter group.Analysis of both SIDS groups against the top group revealed that there exist 'two tiers' in terms of port development, with those in the first representing a level of efficiency that those in the second simply cannot match. In some respects therefore, this 'inefficiency' may be better described as port under-utilization. This suggests that, for these ports, maximizing the efficient use of existing capacities and only thereafter considering port expansion could be the most feasible option for improving technical efficiency in the medium to long-run.

Combining aggregate production function and technical efficiency: Indonesian case study

“…The second is forming the frontier production function equation by adding the constant of the initial production function equation with its maximum error term. The third is measuring the technical efficiency by dividing the initial production function with the frontier equations (Alarenan et al, 2019;Julien et al, 2020;Vasanthi et al, 2017).…”

“…All of them are approaches to measuring technical efficiency. However, the COLS approach applies a least square estimation, which is a parametric approach, while the SFA is a maximum likelihood estimation, which is a parametric-stochastic approach, and the DEA is a linear programming estimate, which is a non-parametric and non-stochastic approach (Julien et al, 2020;Vasanthi et al, 2017).…”

Investigating the Competitive Factors of Container Ports in the Mediterranean Area: An Experimental Analysis Using DEA and PCA

“…Therefore, the most used indicator is the annual container throughput, measured in TEUs. In this study, the choice of the input variables was based on the literature reviews about DEA studies on port efficiency provided by Schøyen and Odeck [26], Julien et al [27], Iyer and Nanyam [16] and Fancello et al [28]. In detail, three capital inputs, Yard Area (Yard) -m 2 , Berth Length (Berth) -m, Maximum Depth (Depth) -m, two non-capital inputs, no.…”

Evaluating the efficiency of Mediterranean container ports using data envelopment analysis

“…In the last years, the related literature has shown a growing interest into port competitiveness. Reviews of previous studies concerning container port efficiency show an increasing use of Frontier methods, such as Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA), for measuring and benchmarking port performance and efficiency (Iyer & Nanyam, 2020;Julien et al, 2018;Kutin et al, 2017;Serra & Fancello, 2020). In this study, DEA was chosen for a number of reasons which, in the authors' opinion, make it preferable in applications on port competitiveness.…”