Exploiting the common pool or looking to the future? A study of free-riding leading up to the 2007 municipal amalgamations in Denmark

ABSTRACT

Municipal amalgamations can create incentives for opportunistic behaviour. Several fairly recent studies have examined this on amalgamations in Denmark and Sweden using the so-called law of 1 over n. However, they have yielded inconclusive results and I argue that one plausible explanation is a theoretical deficiency in the law, as it does not account for how future political representation can mitigate the incentive to free-ride. I examine this using large-scale amalgamations in Denmark in 2007. This case is quite unique as the amalgamations were implemented in such a manner that they constitute a quasi-experiment, and because an extra fiscal year was added to the election period where the composition of the new councils was known with certainty, while the municipalities retained some decision-making power. The study’s findings are consistent with the argument that the incentive to free-ride depends on the decision makers’ future political stake in the new, amalgamated municipality.

KEYWORDS:

• Local government
• amalgamations
• public spending
• law of 1 over n
• re-election

Supplementary material

Supplementary data for this article can be accessed here

Acknowledgements

This study is a refinement of ideas first developed in my dissertation initiated in 2006 and concluded in 2009 (Hansen 2011). It has benefited greatly from thoughts and comments from Professors Poul Erik Mouritzen (Department of Political Science, Aarhus University) and Robert Klemmensen (Department of Political Science and Public Management, University of Southern Denmark) and many other persons who shall remain unnamed for the sake of brevity.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Saarimaa and Tukiainen (2015, 143) instead use the tax base. Hansen (2014) instead uses the number of amalgamating municipalities, and Jordahl and Liang (2010, 165) also examine this alternative measure.

2. Augmented component-plus-residual plots (for models specified as model 2 in Table 1) show that this yields an approximately linear relationship between size and budget overrun. However, this is not the case when using the untransformed measure instead, either as defined by Blom-Hansen (2010) or Jordahl and Liang (2010) (though not Hinnerich 2009 where the relationship is also approximately linear).

3. In some cases, the mayor of the newly amalgamated municipality was newly elected as s/he was not a mayor in any of the amalgamating municipalities. In these cases, the mayor is therefore not re-elected in any of the amalgamating municipalities.

4. 2002 is chosen as reference year as there had not been much talk of reform at this point.

5. As for hypothesis 2, the expectations are analogous to those from Equation 2. Expressed in terms of Equation 4, it is firstly expected that increases in size adversely affects budget overrun in 2006 in amalgamating municipalities (a) whose mayor is not re-elected as mayor of the new amalgamated municipality (${\beta }_{13}>0)$; or (b) when the mayor of the newly amalgamated municipality is newly elected (${\beta }_{13}+{\beta }_{29}>0$). Secondly, that increases in size do not affect budget overrun in 2006 in amalgamating municipalities whose mayor is re-elected as mayor of the new amalgamated municipality (${\beta }_{13}+{\beta }_{25}=0$).

6. P values reported in Tables 1 and 2 – as well as those referred to in the text – are from one-sided significance tests, whereas the significance tests reported in the supplementary material are two-sided.

7. It should also be noted that ${\beta }_5$ is statistically insignificant ($P=0.39$ in model 3), which means that the effect of size is not statistically significantly different from when the mayor is not re-elected. However, this is not inconsistent with hypothesis 2 which expects size to adversely affect budget overrun both when the mayor is not re-elected and newly elected.

8. I also re-estimated models 1 and 2 as fixed effects models that only examine differences between municipalities within each amalgamation. This yielded results consistent with models 1–2 (the output can be obtained from the author).

9. This insignificance may be due to multicollinearity as the model includes interactions terms which increases the chance of Type II error and hence make test statistics overly conservative.

10. An alternative dependent variable is actual tax-financed current net operating expenditures. When the control variables are included, these models yield results that are consistent with the models on budget overrun, except for newly elected mayors where the results are inconclusive. Second, defining common pool size using the total budgeted tax base (e.g., as does Saarimaa and Tukiainen 2015, 143) rather than population size yields very similar results (which is not surprising as the two measures correlate very highly with Pearson’s $r\approx 0.99$). For both, see supplementary material.

Additional information

Notes on contributors

Sune Welling Hansen

Sune Welling Hansen is associate professor at the Department of Political Science and Public Management, University of Southern Denmark. His primary research interests are municipal democracy and fiscal management, with a special focus on the fiscal and democratic consequences of municipal amalgamations.