Spatial patterns in local taxation: tax mimicking or error mimicking?

This paper tests for mimicry in local tax setting, by using a panel data set of the English non-metropolitan districts in the 1980s — when property tax rate variability across districts was highest. The results confirm the presence of large and significant spatial interactions among districts. Even after allowing for district-specific and time effects, a district appears to be significantly affected by its neighbours' policies. A 10% increase in the local property tax rate of a district's neighbours leads to an increase of 4-5% in its own property tax rate. On the other hand, the absence of correlation in tax rates between lower tier (district) authorities and upper tier (county) authorities suggests that the spatial pattern in district tax rates is not simply driven by spatially auto-correlated shocks.