This paper proposes a simple method measuring spatial robustness of estimated coefficients and considers the role of administrative districts and regions' size. The procedure, dubbed "Grid and Shake", offers a solution for a practical empirical issue, when one compares a variables of interest across spatially aggregated units, such as regions. It may, for instance, be applied to investigate competition, agglomeration, spillover effects. The method offers to (i) have carry out estimations at various levels of aggregation and compare evidence, (ii) treat uneven and non-random distribution of administrative unit size, (iii) have the ability to compare results on administrative and artificial units, and (iv) be able to gouge statistical significance of differences. To illustrate the method, we use Hungarian data and compare estimates of agglomeration externalities at various levels of aggregation. We find that differences among estimated elasticities found at various levels of aggregation are broadly in the same range as those found in the literature employing various estimation method. Hence, the method of spatial aggregation seems to be of equal importance to modeling and econometric specification of the estimation.
AbstractThis paper proposes a simple method measuring spatial robustness of estimated coefficients and considers the role of administrative districts and regions' size. The procedure, dubbed "Grid and Shake", offers a solution for a practical empirical issue, when one compares a variables of interest across spatially aggregated units, such as regions. It may, for instance, be applied to investigate competition, agglomeration, spillover effects. The method offers to (i) have carry out estimations at various levels of aggregation and compare evidence, (ii) treat uneven and non-random distribution of administrative unit size, (iii) have the ability to compare results on administrative and artificial units, and (iv) be able to gouge statistical significance of differences. To illustrate the method, we use Hungarian data and compare estimates of agglomeration externalities at various levels of aggregation. We find that differences among estimated elasticities found at various levels of aggregation are broadly in the same range as those found in the literature employing various estimation method. Hence, the method of spatial aggregation seems to be of equal importance to modeling and econometric specification of the estimation.