Overlooked effect of negative data on efficiency analysis

Financial data may contain nonpositive values, which can be addressed using logarithm transformation or the addition of constant. However, production scale is important when employing data envelopment analysis (DEA) to assess the efficiency of input-output transformation. Traditional techniques of handling negative data values in DEA, such as adding an arbitrary constant or substituting a small positive value, fail to account for the effect of production scale on efficiency and can produce misleading findings. This study thus proposes a novel method for dealing with negative output values in DEA by revising them based on production scale. Overall, this study enables better informed decision-making by improving the accuracy of efficiency assessments in complex data settings containing both positive and negative values.