Smoothed bootstrap Malmquist index based on DEA model to compute productivity of tax offices

• We study productivity of tax offices using DEA-based Malmquist Productivity Index.
• In addition a smoothed bootstrap technique is used to provide confidence intervals.
• Mann Whitney U test is utilised to study the influence of specific context variables.
• No evidence that population or number of municipalities affect productivity.
• Results reveal an increase in productivity due to technology and resource management.

This paper analyses the productivity growth of the SUMA tax offices located in Spain evolved between 2004 and 2006 by using Malmquist Index based on Data Envelopment Analysis (DEA) models. It goes a step forward by smoothed bootstrap procedure which improves the quality of the results by generalising the samples, so that the conclusions obtained from them can be applied in order to increase productivity levels. Additionally, the productivity effect is divided into two different components, efficiency and technological change, with the objective of helping to clarify the role played by either the managers or the level of technology in the final performance figures.