DEA-based integrated relationship of returns to scale – an application to road maintenance in Taiwan

Abstract

An Integrated Relationship of Returns to Scale (IRRS) associated with multiple-stage Data Envelopment Analysis (DEA) is proposed for identifying the returns to scale (RTS) among decision-making units (DMUs) appropriately and accurately. The validity and feasibility of the proposed method is tested by using 31 case studies on highway maintenance and construction offices based on the data provided by the Directorate General of Highways (Taiwan). The results show that the multi-stage DEA method with IRRS is more useful than the traditional single-stage DEA for evaluating the status of RTS for each DMU. Among the 31 units evaluated, 14 units are categorized as having increasing returns to scale, 4 have decreasing returns to scale, and no unit has constant returns to scale; the returns for the remaining 13 units cannot be determined.