Do Citizens Really Shop between Decentralised Jurisdictions? Tiebout and Internal Migration Revisited

Abstract

Tiebout predicted that differences in service provision and tax rates across regions would lead citizens to migrate to their preferred jurisdiction. This central tenet of the fiscal federalism literature has rarely been explored outside the North American context. This paper delves into this gap in the literature by examining the factors that undermine Tiebout's prediction. It undertakes an international comparative analysis drawing upon recent innovations in the measurement of internal migration that facilitate country comparisons. While some of Tiebout's ideas find limited support, the overall weight of evidence suggests a weak link between internal migration and decentralisation

Acknowledgements

Financial support of the European Research Council under the European Union's Seventh Framework Programme (FP7/2007-2013/ERC grant agreement n° 269868 and project ECO2011-29314-C02-01 of the Spanish Ministry of the Economy and Competitiveness is gratefully acknowledged. We would also like to thank the editor, Ronan Paddison, as well as the anonymous reviewers for their comments.

Notes

Note that ‘decentralisation’ in this context is used to refer to the devolution of powers from a national to a sub-national political and administrative unit, and does not refer to the migration of individuals away from large urban centres (as it does in much literature on migration and settlement patterns).

Courgeau's K can be calculated with the formula

M(n)/P = k ln n

where, M(n) is the number of inter-area migrants observed when the territory is split into n units and P is total national population. K can be interpreted as “an indicator of mobility independent of the administrative settings used for the measurement of migration” (Rees and Kupiszewski, 1999, p. 32).

This is not the same as saying that decentralisation would not lead to internal migration. The MEI measures the degree to which internal migration changes the overall spatial pattern of human settlement. The MEI therefore assumes low values when there is a high level of internal migration but also a high degree of off-setting among regions, which is exactly what Tiebout's theory would predict.

For the GFS-based calculations, the proportion of sub-national expenditure over total expenditure was calculated. While the data taken from Schneider's (2003) study are based solely on 1996 data, this is suitable to compare with Bell and Muhidin's (2009) data because the latter draw upon census data with a median year of 2001. It is worth noting that our two measures of fiscal decentralisation are correlated: the GFS ratio and Schneider's measure of fiscal decentralisation share a Spearman's rank correlation of 0.86 with significance 0.003 for the nine countries included from the Rees and Kupiszewski data.

Bell and Muhidin's (2009) data include five-year Courgeau's K values for Ghana, South Africa, China, Indonesia, Malaysia, Philippines, Vietnam, Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Australia, Canada, Portugal and the USA and lifetime Courgeau's K values for Ghana, Kenya, Uganda, Belarus, Cambodia, India, Indonesia, Philippines, Brazil, Chile, Colombia, Costa Rica, Ecuador, Panama, Portugal, Spain and the USA. Rees and Kupiszewski's (1999) data include annual Courgeau's K's for Czech Republic, Great Britain, Estonia, Germany, Italy, the Netherlands, Norway, Poland, Portugal and Romania.