Abstract
We exploit the introduction of an open data online platform - part of transparency program initiated by the Italian Government in late 2014 - as a natural experiment to analyse the effect of data disclosure on expenditure on mayors’ expenditure and output decisions. First, we analyse the effect of the program, by comparing municipalities on the border between ordinary and special regions, exploiting the fact that the latter regions did not participate in the program. We find that mayors in ordinary regions immediately change their behaviour after data disclosure, and that the reaction depends on their relative performance as shown on the online platform. Second, we investigate the effect of mayors’ attention to data disclosure within treated regions using daily data on their access to the platform, which we instrument with the daily publication of newspaper articles mentioning the program. We find that mayors react to data disclosure by reducing spending via reducing service provision, where the latter variable is "hidden" i.e. not disclosed on the online platform. Overall, local mayors seem to target variables that are measured in the program at the expense of variables that are not measured. Our results indicate that the details of the design of data disclosure are important in achieving benefits from such programs.