Severe Flaws of Returns to Scale in Data Envelopment Analysis

In data envelopment analysis (DEA), returns to scale (RTS) are a widely accepted instrument for a decision-making unit (DMU) to reveal its activity scaling potentials. In the case of increasing returns to scale (IRS), a DMU learns that upsizing activities improves its productivity. For decreasing returns to scale (DRS), the instrument likewise should depict a downsizing force, again for improving productivity. Unfortunately, here the classical RTS concept shows misbehavior. Under certain circumstances, it is the wrong indicator for scaling activities and even hides respective productivity improvement potentials.In this paper, we study this phenomenon and illustrate it via little numerical examples and a real-world application consisting of 37 Brazilian banks