摘要:Differences in local or regional public spending needs are a serious concern in many countries, boosting the implementation of equalization grants. In order to quantify the amount of grants, assessments of spending needs are needed. One standard technical approach is to rely on econometric estimates of equations in which the dependent variable is spending on the different services and the set of explanatory variables contains both proxies standing in for factors determining spending needs and also control variables. While this is a sound methodology, usual panel data estimators may be biased due to the low within-variation of some observed and unobserved variables. This short paper shows this potential problem and makes a methodological suggestion to overcome it: the use of a new estimator for panel data with time-invariant regressors.