https://orcid.org/0000-0002-9890-6091
![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
The decentralization of funding management poses a conceptual challenge to the study of political criteria in the allocation of funding in the EU Cohesion Policy, as existing research often assumes unidimensional actor constellations and motivations. Combining insights from distributive politics and multilevel party politics, this article uses a unique data set of beneficiary data at the local level from Polish Regional Operational Programs (2007–2013) to explore the ability of regional authorities to engage in vote-seeking and explores how institutional and political multilevel structures affect these strategies. We demonstrate that the vote-seeking of sub-national authorities is constrained in two ways. Competition between regional and national authorities limits the possibility of regional governments that are politically opposed to the national government targeting their electoral strongholds. In contrast, partisan harmony between different institutional levels incentivizes a vote-seeking strategy that takes into consideration electoral dynamics at both the regional and national level.
Acknowledgements
The author would like to thank Arndt Wonka, Adrienne Héritier, Ingo Rohlfing and the anonymous reviewers for their comments that helped to improve this paper in its different stages.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 This does not preclude that political actors may rely on different strategies when targeting swing electoral districts/regions (Cox Citation2009).
3 The data only excludes investments into airports, projects outside a region, financial instruments and technical assistance, as these either have no geographical location, or disperse benefits over the whole region. Project start dates determine the assumed year of allocation.
4 According to decomittment rules, project expenditures were reimbursed up to 2015. For the Holy Cross Region, no funding was allocated in 2015.
5 We refrain from the usage of a multilevel model with three levels given the small size of clusters (Stegmueller Citation2013) at the regional/municipality level (16/9). Panel approaches might suffer from changes in the sample, methodological changes of the control variables and the importance of time-invariant variables for allocation outcomes.
6 The power link (0.35) was chosen based on the results from Pearson correlation, Pregibon Link, and Modified Hosmer–Lemeshow test.
7 See the supplementary files for further details on variables.
8 In election years, we rely on the result of the previous elections, as elections usually took place in Poland in October/November.
9 Data and replication files are available upon request.
10 The results related to the hypotheses are also robust when using a random-effects panel tobit model instead of the two-part model, split samples (model 3), and the removal of the top 5% of observations with regard to vote differences (model 4). For further details see supplementary files.
11 As standard errors might be correlated within regional entities, standard errors are clustered for periods determined by the change of political alignment or elections at the region level (N = 60). Results also appear robust when clustering by periods of political alignment (N = 31).
12 Thereby it limits deviations based on second-order effects (e.g. punishing the government for national incumbency) or region-specific voting behavior. Calculated based on official election data.