A spatial merger estimator with an application to school district consolidation

Municipal consolidation and local government behavior: evidence from Japanese voting data on merger referenda

“…A municipality is likely to receive a large financial benefit from consolidation and choose to merge when disparity of per capita expenditures between the merging municipality and the merged one is large. The estimation result is consistent with the predicted theory and the existing empirics (Alesina, Baqir, and Hoxby, 2004;Brasington, 1999Brasington, , 2003aGordon and Knight, 2009;Sorensen, 2006). Second, with regard to size effects, large and small municipalities are more likely to merge.…”

“…The literature finds evidence for trade-off between scale economy and heterogeneity of preference, particularly with respect to income and race (e.g., Alesina, Baqir, and Hoxby, 2004;Alesina and Spolaore, 2003;Austin, 1999;Brasington, 1999Brasington, , 2003aBrasington, , 2003bGordon and Knight, 2009;Nelson, 1990;Sorensen, 2006). 4 These empirical studies employed a discrete choice model to investigate the role of economics of scale, heterogeneous preference, and tax base gains in merger decisions.…”

“…To avoid the drawbacks associated with this difficulty, Austin (1999) and a series of papers by Brasington (1999Brasington ( , 2003aBrasington ( , 2003b estimated by a bivariate probit model developed by Poirier (1980) or a similar model. Gordon and Knight (2009) developed a simulation-based estimator derived from a calculation of the probability of a merger in matching games. Still, the less informative nature of data on individual preference in consolidation continues to gives rise to a difficulty in rigorous estimation of local government behavior.…”

“…However, there would likely remain some shortcomings in these approaches since the bivariate probit model suffers from lack of information, and an estimation model using a matching game imposes various strong restrictions on the justification of the theoretical framework. Moreover, because merger with more than two municipalities is allowed in Japanese consolidation, an estimation based on a one-sided matching game, such as that proposed by Gordon and Knight (2009), could not be applied to Japanese cases. On the other hand, the share data this paper employs allow for distinguishing preferences of distinct municipalities, further enabling a measurement of the strength of preferences in each municipality.…”

“…Additionally, the study used data pertaining to local referenda on municipality partitioning to investigate the median voter's behavior in boundary reform. This is in contrast with most empirical works that have studied the mergers of school districts and local governments, such as Brasington (1999Brasington ( , 2003aBrasington ( , 2003b, which used a data set of about 300 urban communities in Ohio that belonged to a consolidated school district or provided schooling independently, and Gordon and Knight (2009), which employed 12,234 samples at the school district border level, including 50 merger cases in Iowa during the 1990s. 6 Some studies state that unconditional grants, called Local Allocation Tax (LAT) in Japan, are calculated following the non-arbitrary formula, and may be considered as a lump sum transfer (Mochida, 2006;Reschovsky, 2007).…”