![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
ABSTRACT
During the past decade, holding local referenda in the Czech Republic has become more frequent. Due to current regulations, however, the direct voice of people cannot be always heard; it largely depends on the number of citizens who express their preferences. The aim of this article is to show how to achieve higher turnout in local referenda such that they become binding on politicians. It hypothesises that the questions on NIMBY (“not-in-my-backyard”) issues should considerably stimulate people's political engagement. This supposition is examined through the empirical analysis of 258 local referenda held between 2004 and 2014 in the Czech Republic.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes on contributor
Michal Nový holds a PhD in Political Science from Masaryk University in Brno, Czech Republic. He also has an MSc in Regional Development from Mendel University in Brno. His research focuses on political behaviour and public administration.
Notes
1. However, we should stress that the NIMBY phenomenon does not only involve renewable energy projects. In the Czech case, it refers to development proposals in general. For example, it includes such facilities as nuclear waste disposal sites or airports because they also meet the NIMBY definition (see below).
2. It should be noted that there is an older tradition of direct democracy. As Kaufmann and Waters (Citation2004, 48) have emphasised, the right to referenda was included in the 1920 Constitution of the Czechoslovak Republic.
3. It is little bit surprising that no referendum according to the main 1992 law took place until 2000. During the 1990s, only referenda on municipal separation and agglomeration (on the basis of the 1990 law on municipalities) were held (Smith Citation2011, 35).
4. Different rules are applied if the referendum is held due to the split of a municipality or its merging with another municipality. In this case, the minimum turnout of 35% together with the requirement concerning a majority that comprises at least 25% of all eligible voters is irrelevant. To be binding, more than 50% of all eligible voters must vote in favour of a given issue.
5. There is a more extensive discussion whether such minimums for referendum validity are reasonable at all. For example, Verhulst and Nijeboer (Citation2007) did not agree with the use of these quorums. They point out that other democratic institutions, such as elections, are not burdened by such quorums. In their view, those who do not want to participate in the political contest should be considered as the individuals who entrusted their political voice to those who participate. In addition to others, counterarguments were presented by a former Czech Minister of the Interior, Stanislav Gross (see Špok et al. Citation2006, 14−15). Before the new 2004 referendum law was passed, he was fearful that if referendum result based on active participation of for example 20% or 30% of people will be binding for local government, the legitimacy of the approved decision would be very low.
6. In the Czech Republic, unfortunately, no surveys on the particular referenda are usually administered. For an illustration of how an analysis of referenda at the individual level might look like, see Dvořák (Citation2013).
7. There is no doubt that the literature cited here has contributed to the discussion on the determinants of voter turnout in local referenda. However, it does not contain broader theoretical arguments on why a particular aggregate-level variable might affect voter turnout. Thus, it makes sense to extend the findings of these works here (using a rational-choice perspective).
8. Of course, previous studies have examined the broader range of the explanatory variables. On the other hand, the explanatory model presented in this article strives to be as parsimonious as possible. In short, Smith (Citation2007) observed the association between referendum turnout and socio-demographic/socio-economic variables (e.g. percentage of people who have completed high school, proportion of people employed in agriculture, etc.; see also Kostelecký Citation2005) or the number of civil society organisations. But some of these variables may be strongly correlated with municipal size (since all of them are associated with the centre-periphery cleavage). In the Czech Republic, countryside municipalities are small, people there are less educated and usually have less developed civic skills, and many of them work in agriculture). Smith (Citation2007) also employed participation rate in the previous municipal election as a predictor of referendum turnout, most likely to show to what extent “participatory culture” is developed in the municipality. We do not follow this approach, because it is doubtful whether such a variable can accurately express something about deeply rooted participation patterns of the citizens. In addition, this article analyses referenda over the past decade, so for some cases, the 2010 municipal elections would be taken into consideration and for others, the 2002 municipal elections would be more relevant. Furthermore, there are other variables that might be influential, but in the Czech Republic, it is very complicated to gauge them (e.g. the intensity of a given referendum campaign, etc.).
9. Due to the existence of minimum turnout quorums, however, the rational calculation might also take some other form. In some cases, citizens who are not in favour of a given issue may opt to abstain to prevent the referendum result from being binding. Such a strategy may be sometimes more effective than to express disapproval through active participation.
10. To justify such a causal chain that represents the indirect effect of institutional incentives (i.e. institutional incentives → short-term political attitudes → turnout), we must add that some scholars assert that if the variable which lies at the beginning of the causal chain, that is, the independent variable, is strongly associated with the mediating variable, and at the same time, the mediating variable is greatly associated with the final outcome, there should also be a strong relationship between the independent variable and the final outcome (see Anderson Citation2007).
11. From time to time, the initiators of LULU projects offer some compensation for the depreciation of the area. However, these attempts to compensate for the situation (“a backyard bonus”) do not usually lead to homeowners’ assent (Schearmur and Shearmur Citation2010).
12. On the other hand, higher turnout in small municipalities might also be determined by aspects of rural living. People in the rural areas know each other rather well and there is higher social pressure on political participation. Therefore, higher turnout may be caused by denser social networks and social capital (see Putnam Citation1993).
13. It should be emphasised that this statement may be somewhat controversial. To this, Blais (Citation2000) notes that the cost of voting is usually perceived to be a sum of two elements: (1) the cost of going to a polling place and (2) the cost of acquiring information to determine which answer to the referendum question is better (in the case of parliamentary elections, analogously, it is the cost of acquiring information to assess which candidate or party would best represent a voter’s views). In the two-in-one situation of concurrent elections, the first element is lower relatively, but the cost of the second element is higher because the voter must obtain information about two different contests.
14. We are thankful to Markéta Kovandová from the Ministry of Interior for the access to data and for additional information about local referenda.
15. Of course, the same referenda that are recorded in both databases are considered only as one case.
16. Czech local referenda before 2004 have been described in detail by Smith (Citation2007, Citation2011). Some other information is also available in the works of Špok et al. (Citation2006) and Kratochvílová (Citation2011).
17. In the Czech case, there are several contests that can be held simultaneously with a local referendum: parliamentary (to the Chamber of Deputies and to the Senate), presidential, local, provincial and/or European elections.
18. A 50% turnout threshold was in effect from 2004 until 1 July 2008 for all types of local referenda.
19. If the referendum concerns the merging or split of the municipality, the majority of all eligible citizens must vote in favour of it to be binding. This implies that referendum turnout must always be higher than 50%.
20. After a legal change that enabled the municipalities to limit or even ban gambling machines (see Law No. Citation300/Citation2011 Coll. which amends Law No. Citation202/Citation1990 Coll., on lotteries), several referenda on these facilities were held. However, these machines as such are not regarded as LULUs because they are not confronted with a “public good versus private bad” conflict typical of NIMBY phenomena. In addition, it should be emphasised that in every country, some topics are exempt from referendum, for example, municipal fees (see Beramendi et al. Citation2008, chap. 2). In the Czech Republic, besides other things, the adoption of a local ordinance cannot be subject to a referendum (see Law No. Citation22/Citation2004 Coll.). According to Law No. Citation202/Citation1990 Coll., on lotteries, the ban on gambling machines must be always included in a local ordinance. Nonetheless, many local referenda that concerned the ban on these machines were held.
21. It is noteworthy that the decisions of local governments on these types of “special” LULU facilities can be considered as controversial if they are based on the results of local referenda. These facilities were planned in areas where the municipalities which conducted the referendum had almost no powers. In other words, despite the fact that these LULUs were not under the control of local self-government, several referenda on their citizens’ acceptance were held.
22. Sometimes, it was not easy to decide whether a particular facility could be related to NIMBY opposition. In addition to the issue of the gambling machines (see footnote above), there are some other borderline cases, for example, fuel storage plants or photovoltaic plants. Due to their low or unclear negative impact on the residents, these facilities are not regarded as LULUs.
23. It is also proper to look very briefly at the remaining (i.e. non-NIMBY) issues that were submitted to local referenda. Most frequently, Czech people decided directly about (1) the changes in municipal plans (in particular, about the enlargement of territories where building is allowed), (2) purchase, sale or lease of municipal property (the land parcels in the first place), (3) investment intentions of the municipality (building of sewage system, swimming pool, sports hall, etc.), (4) municipality split, and (5) modifications of existing municipal facilities (e.g. the reconstruction and merging of schools, relocation of nurseries). After the legislative amendment in 2011, there has also been a substantial increase in the questions considering banning gambling machines in the municipality.
24. To investigate the research question in more detail, the results of OLS regression models are compared with logistic regression models that considered whether the referendum result was binding for the politicians as a binary response variable (1 – yes). As described above, the effect of the referendum on local governments is strongly dependent on the number of participants, and therefore, these additional models can confirm whether or not the factors discussed in the previous paragraphs were really influential. Here, β-coefficients (used for assessing the impact of particular variables) and Nagelkerke R2 (for evaluating how a model reflects the data structure) are displayed (see Field Citation2009). Apart from a few short comments in the next section, these models are presented in full in Appendix 3.
25. The year 2004 is used as the reference category.
26. When comparing means of turnout during the past 10 years, it appears that the time period may have an impact on turnout. See Appendix 2.
27. The lowest quintile of municipal size covers communities of up to 267 eligible citizens. On the other hand, the highest quintile comprises municipalities that have more than 2687 eligible citizens.
28. As can be seen in , the coefficients change to a large extent. Therefore, it is not adequate to mention exact numbers to express the intensity of the effect.
29. The other coefficients introduced in Appendix 3 have the same direction as those in . Besides other things, this appendix confirms that the NIMBY question does not only lead to higher turnout, but it also increases the probability that the referendum result will be binding.
30. The models in Appendix 3 were also supplemented by these extra parameters, but the effect of none of them was statistically significant. This is why these extended models are not introduced in Appendix 3.
31. However, a continuous time variable (which is much more parsimonious than the set of year dummies) was not significant when added into the model instead of time-fixed effects.