Productivity of tax offices in Norway

The performance of local tax offices is studied over time using data envelopment analysis to calculate Malmquist productivity indices. The index has the proper homogeneity properties of a total factor productivity index. One input, cost, and six output categories of the main service activities carried out by tax offices, are specified. A bootstrap approach is applied to establish confidence intervals for the individual indices enabling an identification of units that have significant productivity decline, growth, or no change. A novel visual test groups units into these three possible categories. This way of showing consequences of uncertainty should facilitate more tailor-made policies to promote efficiency and productivity improvements. Productivity changes are distributed from a 26 % decline to a 35 % increase over the three-year period with an average growth of 4 %. Inspecting individual unit results, the confidence intervals tend to be wider the larger the units, thus providing more accurate insights than point estimates for actions to improve productivity. Looking at positive and negative changes in cost and productivity together the development of offices is classified into four categories of interest to policymakers; efficient cost increase, efficient cost saving, inefficient cost saving, and inefficient cost increase.