What shapes the flypaper effect? The role of the political environment in the budget process

ABSTRACT

This study investigates the spatial heterogeneity of the flypaper effect in a sample of 2,451 Spanish municipalities over the period 2003–2015 by means of Bayesian spatial panel data econometric techniques including municipal and time-period fixed effects. In particular, we analyse how differences in the degree of political competition and the local governments’ monitoring and enforcement effort in tax collection affect the size of the flypaper effect. Our results suggest that municipalities with higher tax collection efficiency and where local governments have more political strength exhibit a lower flypaper effect.

KEYWORDS:

• Flypaper effect
• political competition
• tax collection efficiency
• Spanish municipalities
• Bayesian spatial econometrics

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1. See Hines and Thaler (1995) or Inman (2008) for an overview.

2. The estimation strategy for cross-section spatial models of Acosta (2010) and Bastida, Guillamón, and Benito (2013) is based on the IV estimator, which in the context of spatial regressions has as a main drawback the fact that the coefficient estimate of the spatial autoregressive term may fall outside its parameter space and it is usually less efficient (Elhorst and Fréret 2009).

3. For a comprehensive survey, see Gamkhar and Shaw (2007) or Inman (2008).

4. In the case of Spain, intergovernmental grants are mainly unconditional and account for more than a third of municipal revenues on average.

5. In this setting ${\beta }_i$ is smoothed using a combination of values in neighbouring areas. The value of $\theta$ is based on a cross-validation procedure where the score function $\theta =argmin{\sum }_{i=1}^n{\left[E_i-{\hat{E}}_i\left(\theta \right)\right]}^2$ is minimised.

6. Using data on Peruvian municipalities, he finds that municipalities facing higher tax collection costs are more responsive to additional grants, which should increase the size of the FPE.

7. The use of political variables has forced us to restrict the panel data to the electoral years (i.e., 2003, 2007, 2011 and 2015). Therefore, our $T=4$. This limited time dimension precludes us to employ a dynamic specification given that by construction we would lose one period leaving us with $T=3$, and such a short panel could yield very noisy estimates of both the temporal and the space-time diffusion parameters.

8. The first spatial weights matrix is based on the concept of first-order contiguity, according to which $w_{ij}=1$ if jurisdictions $i$ and $j$ are physical neighbours and $0$ otherwise. The definition of neighbouring regions used here is based on physical contiguity. To ensure that every region has at least one neighbour, we employ the Delaunay triangulation by constructing Voronoi polygons from the centroids of the sample municipalities using the Matlab function xy2cont.m developed by LeSage. See LeSage and Pace (2009, 118) for further details. Secondly, several matrices based on inverse distance with cut-offs at different thresholds of distance in kilometres are introduced. In addition, power distance and exponential decay matrices with cut-offs at the first and second quartile are employed. Finally, k-nearest neighbours computed from the great circle distance between the centroids of the various jurisdictions are also considered.

9. The results are shown in Table A1 in the Appendix.

10. In order to save space, the results are not shown here but available upon request to the authors.

11. We use the tilde notation to distinguish them from the parameter estimates in Equations (5) and (12), given that to simulate the FPE conditional on $Z_t$ we use the total effects implied by the Monte Carlo simulation of the partial derivative of the reduced form of the SDM in Equation (8).

12. A concern might arise as to whether our measure of tax collection efficiency may simply reflect higher per capita income of the local taxpayers, thus affecting the interpretation of the previous result. Nevertheless, this is not the case in our sample of municipalities as there is a negative but relatively weak correlation between these two variables (−0.31 with p-value 0.00).

Additional information

Funding

This work has benefitted from the financial support of the Spanish Government’s Ministry for the Economy and Competitiveness, Project [ECO2016-76681-R].

Notes on contributors

Vicente Rios

Vicente Rios is currently a postdoctoral researcher at the University of Milan, Italy. His research interests are spatial econometrics, regional epidemic modelling, economic development, environmental and energy economics, political economy and local public finance. He has recently published in journals including the Economics Bulletin, the Journal of Policy Modeling and Regional Studies.

Miriam Hortas-Rico

Miriam Hortas-Rico is an assistant professor in the Department of Economics and Public Finance at the Autonomous University of Madrid, Spain. Her main research interests are urban economics, fiscal federalism, local public finance and spatial econometrics. She has recently published several articles in Regional Studies.

Pedro Pascual

Pedro Pascual is Professor of Applied Economics in the Department of Economics at the Autonomous University of Madrid, Spain. His areas of research are public economy and regional economy, with a focus on growth, convergence, regional inequality, decentralisation and public capital. He has published recently in Socio-Economic Planning Sciences, Papers in Regional Science, and Applied Economics.