Title: Decentralization and the shadow economy: Oates meets Allingham–Sandmo
Abstract: Abstract This article studies the impact of decentralization on the shadow economy. We argue that decentralization may decrease the size of the shadow economy mainly through two transmission channels: (1) decentralization enhancing public sector efficiency (efficiency effect), and (2) decentralization reducing the distance between bureaucrats and economic agents, which increases the probability of detection of shadow economic activities (deterrence effect). Using various measures of fiscal, political and government employment decentralization in a cross-section of countries, we find the deterrence effect to be of more importance. The deterrence effect is stronger, the lower the degree of institutional quality. We find no robust evidence of the efficiency effect. Keywords: decentralizationshadow economyinstitutional qualityJEL Classification:: O17H77D78 Acknowledgements We would like to thank Alexander Karmann, Marcel Thum, Christoph Böhm and the participants of brown bag seminar at TU Dresden for helpful comments. Further, we acknowledge the useful comments of an anonymous referee. Gunther Markwardt thanks the CES in Munich for the hospitality enjoyed while working on this article. Notes 1 Alañón and Góómez-Antonio (Citation2005) find that e.g. Spain had a shadow economy of around 20% of official GDP already in the early 2000s. 2 Decentralization may also have opposite effects on the shadow economy, e.g. through corruption as transmission channel, which is beyond the scope of this article. See Dreher and Schneider (Citation2010) for details. 3 The literature about the 'shadow', 'underground', 'informal', 'second', 'cash' or 'parallel' economy is increasing and various topics, such as how to measure it, its causes and effects on the official economy, and the usefulness of shadow economy estimates for economic policy are discussed by e.g. Frey and Pommerehne (1984), Johnson et al. (1997, 1998), Gerxhani (2004), Schneider (2005) and Méon et al. (Citation2011). 4 For a detailed discussion on the strengths and weaknesses of the different estimation methodologies including surveys and discrepancy methods see Schneider and Enste (Citation2000). 5 Several criteria have to be fulfilled to count as a federal country: countries must have at least two levels of government, which share parts of the executive and legislative authority; sub-national governments have a representation in the federal parliament (second chamber); there is a duty to obtain consent on constitutional amendments; a constitutional jurisdiction solves disputes between organs of state; institutions foster collaboration (see Watts, Citation2008). 6 See Table A1 in the Appendix. Table A2 provides summary statistics. 7 Using the 'fiscal freedom index' of the Heritage Foundation as an alternative measure of the tax burden does not change the results. The estimated coefficient of the fiscal freedom index is not statistically significant. 8 For an excellent overview on dos and don'ts in interaction models see Brambor et al. (Citation2006).