Do what we did last year, but do not stray too far from the pack: A behavioral public finance approach to municipal cash reserves

We leverage a national panel of US municipalities to show that behavioral finance helps explain the number of months of expenses that municipalities save in cash and investment reserves. We hypothesize that municipal managers may be using numerical anchoring based on historical values to target the number of months of savings to hold and that they may also be engaged in social learning to target months of savings based on the behavior of neighboring municipalities. We test for these effects by combining two innovative techniques, a two-stage regression designed to test for anchoring of present financial values based on theoretically unimportant historical values, and a measure of the spatial autocorrelation of savings to test for social learning. The results suggest that, in deciding how much to save, municipal managers are influenced by the levels of savings they held in the past and the savings levels of their neighbors, and that they underreact to changes in theoretically relevant economic fundamentals. Further tests also suggest that the smallest cities by population are more influenced by the behavior of their neighbors than their past savings, whereas the largest cities show the opposite result, effectively choosing themselves as their own role models.