Stimuli contributing to local property taxation – with the focus on spatial effects of industry and functional urban areas

A lot of scholars dealing with the issues of tax policy at the level of local governments often pay a close attention to tax mimicking phenomenon, a situation when neighbouring local governments follow the same pattern of tax policy in order to support another kind of their public policy. For instance, the tax mimicking is a part of yardstick competition, as presented by Besley and Case (1995) for the United States on data covering the period 1960–1989, Bordignon et al. (2003) for Italy, Allers and Elhorst (2005) for the Netherlands, Revelli and Tovmo (2007) for Norway, Baskaran (2014) for the German State of North Rhine-Westphalia, Bastida et al. (2019) for Spain, Małkowska et al. (2018) for Poland and Sedmihradská (2013) for the Czech Republic.

Scholars focusing on the public policy might pay a closer attention to dynamics of agenda-settings. Local tax policy as a part of yardstick competition might be considered as an innovation in public policy that has a potential to diffuse within a specific region (Bordignon et al. 2004). It is an evolutionary process when local governments are changing their policy. Thus, the policy of the high local property tax coefficient might be analysed from the point of view Baumgartner and Jones’ punctuated equilibrium model, as John (1999) did it for British local government finance in the early 1990s.

Scholars in social sciences pay a closer attention to ideological motivation of tax competition among the local governments, e.g. Solé Ollé (2003), or they focus on intergovernmental and inter-regional relationship within a framework of fiscal federalism, e.g. Fossen et al. (2017), Buettner and von Schwerin (2016), Gerard et al. (2010) and Bordington et al. (2004). Other scholars focus on an impact of yardstick competition on local prices; for instance, Klien (2015) who scrutinised the impact of tax policy on water prices in Austria.

Rarely, the issue of local tax policy and tax mimicking is analysed in a nexus with other factors contributing to local governments’ decision-making on taxation as presented by Goemine and Smolders (2015). Nevertheless, the question what factors matter in local tax policy is a momentous one. For instance, analysing the panel data set of 104 French local governments from 1989 to 2001, Dubois and Paty (2010) found that “voters reward their incumbent when neighbouring cities that are similar in terms of demographic characteristics have high local taxes”. Also, a level of concentration of population in an analysed region where policy of high local taxes is applied matters in local political decision-making.

As noted by Zdražil and Pernica (2018), there were a lot of municipalities in the Czech Republic that raised the local property tax coefficient whenever their region was affected by a presence of some booming industry. For instance, a number of municipalities in well-visited mountain resorts and municipalities with a significant presence of automotive or power industry opted for the policy of the high local property tax coefficient. Furthermore, Zdražil and Pernica (2018) pointed out that the policy of high local coefficient (LC) is frequently applied in areas positioned close to large cities coping with the burden of sub-urbanisation. For instance, Schneider (1987) explains the taxation in sub-urban zones as an effect of inequalities between the taxes collected, which spreads the pressure on raising the local tax bases.

So, it seems that the policy of higher property taxation, known in the Czech Republic as the policy of the high local property tax coefficient, and systematically applying by ca 10% of municipalities since 2008 is linked mostly with demographic and economic booming areas. In other words, the municipalities take advantage of economic development to maximise their tax revenue, although they have not contributed to decision of private actors to install their property in the municipality. From this point of view, the policy of higher property taxation is hardly to interpret as a yardstick competition, since the effects of this policy, e.g., described in an econometric model by Bastida et al (2019), remains statistically insignificant.

Following those arguments, the municipal decision-making on local taxation should be analysed in the nexus with some spatial pattern of the region because the higher local property tax has been often levied on the real estate, which had been generated as a consequence of regional development in the past. Yet, that is not in accord with theory of yardstick competition because investors allured in the municipality in the past are penalised finally by higher taxation on their property. In that opinion, we assume that the propensity of municipalities to raising the local property tax coefficient in particular regions is more than just a tax mimicking and yardstick competition, as presented by Sedmihradská and Bakoš (2016). Due to the fact that a significant number of municipalities keep continuing the policy of the high local property tax coefficient after the fiscal stress as presented in the Tab. 1, there may be more links between the policy of the high local property tax coefficient and the socioeconomic background of the region where the coefficient was raised.

As presented in the Tab. 1, the municipalities faced a fiscal stress by 2013. In response to a significant drop in revenues, the debt of municipalities was increasing. In order to cover this drop, some municipalities took advantage of policy of the high local property tax coefficient. Property tax revenues doubled between 2008 and 2014, and this tax gained more significance in local budgets. Taking into account just municipalities with the LC 2, 3, 4 or 5, the significance of property tax in their budgets is almost twofold in comparison with other municipalities.

However, the local property tax is levied on the housing development and industry and the level of socioeconomic developed of a region encourages the local governments to tax the population and industry concentrating in the municipality with high property tax. From that point of view, the tax mimicking may be just a spin-off of regional development and spatial distribution of the industry and population in the past. As presented by Zdražil and Pernica (2018), there is a link between the level of standard of living in a municipality and the level of property tax coefficient set by the local government, that is, municipalities applying the policy of the high local property tax coefficient kept the standard of living determined as a combination of population indicators and availability of public services, such as education and medical facilities, communication services, fire brigades, etc. Yet, other municipalities maintain the same standard of living without any need of extra revenues. Therefore, the option for a higher taxation might be determined by an extra factor standing still out of scholars’ attention.

Due to the fact that tiny municipalities, such as Trnavka, Vresova and Nizni Lhoty, opted for the highest LC when a large industry facility, such as a power plant, was operated on their administrative territory, we supposed that a localisation of industry in the municipality inspires the local government to impose a higher taxation on it. In particular, when industry is suspected to produce negative externalities, such as pollution. However, this taxation cannot be considered as a Pigovian tax, because the same decision on higher property taxation was made also by municipalities without any significant location of dirty industry.

On the other hand, densely populated municipalities, such as large cities with a status of regional centres, opted frequently for the policy of higher taxation as well. By 2020, 15 of 26 of such municipalities opted for a rising of the LC; 11 of them continue this policy for more than nine fiscal years. Nonetheless, the regional centres usually opted for the LC 2, that is, lower tax increase per unit. In comparison with municipalities with the localisation of industrial facilities, they utilise concertation of population in order to maximise tax revenue on the low level of the LC.

It means that decision-making on raising of the LC is supposed be motivated by those two patterns of thinking. In order to test an impact of socioeconomic background of the region on the decision-making of local governments on an introduction and continuing of the policy of the high local property tax coefficient, two hypotheses were set:

H1: The localisation of industry stimulates municipalities to increase the local property tax.

Following the aim of this paper, the links between localisation of large industrial facilities and the LC of property is examined in this part. As presented in the Tab. 3, there are 1,713 large industrial facilities that are distributed across 1,442 municipalities, that is, at least one large industrial facility is localised in 23% of Czech municipalities. However, as presented in the Tab. 3, the large industrial facilities are localised in one-half of municipalities that apply the policy of the higher local property tax coefficient. Therefore, the municipalities burdened by large industrial facilities apply the policy of higher property taxation twice frequently than the rest.

As presented in the Tab. 3, the importance of the industry factor is decreasing over time when the peak was in 2010 and the bottom in 2019. Based on the figures above, the presence of the industry should be considered as an inducement for an implementation of the policy of higher local property coefficient during the acute fiscal stress. In particular, it was an easy way how to get extra tax revenue when national economy slowed down due to the financial crises of 2008.

Yet, the stress subsided and other stimuli for taxation of real estate emerged. The policy became inviting to municipalities with a different background. With regard to the development after 2010, the industrial stimuli keep enfeebling although tax mimicking keeps on. In fact, the large industry is localised irregularly in the Czech Republic. The municipalities with its presence in their administrative territories opted for its higher taxation as the first once; however, as mimicking proceeds, than increasing number the policy appliers attract not only municipalities with industrial facilities to join, but also that with a different spatial background.

Going more into detail, we found out that property tax did become significant for some local governments. For instance, property tax is a substitutional source of revenues in the budget of the municipality of Temelin, where a nuclear power plant is installed. According to MFCR (2020), the increased portion of property tax revenues amounted 40% of all budget revenues in 2018. Trnavka, a village of 229 residents, in the close vicinity of the private coal power plant Chvaletice, won 50% of the budget revenues as the increased portion of property tax revenues in 2018 (MFCR 2020). Besides, 25 local governments sticking to the policy of higher (highest) local property tax coefficient in 2018 gain 15% of their revenues as the increased portion of property tax once. As presented in the Tab. 4, those municipalities have to bear the presence of different industries.

On the other hand, the property tax is levied not only on industrial sector, but households are taxed as well. That shows the local governments in a bad light. Yet, the local governments are usually conscious of these consequences. In order to respond to any taxpayers’ criticism, they established munificent programmes of subsidises. For instance, both Trnavka and Temelin provide a special programme supporting just their residents (Latislav 2020; Otta 2009). Was it really an aim to authorise the local governments to build up a few of micro imperio in imperium that are independent from the distribution of tax revenues by central rules of fiscal federalism in fact? However, this analysis has explained just a part of local governments’ motivation to implement a local policy of high property taxation.

Another part of the expected stimuli for increased property taxation is the uneven concertation of population on the state territory. Also, we should analyse the spatial distribution of municipalities that apply the LC as the next step. The reason is to examine the supposed link between the application of LCs and densely inhabited cities with their commuting zones, that is, within FUAs.

Following the OECD definition (2012), there are 1,745 municipalities belonging to FUAs in the Czech Republic. That means there are almost 28% of municipalities that are highly integrated with their regional centres in terms of commuting. As Tab. 5 shows, the share of municipalities applying the policy of the higher property taxation within an FUA is slightly below 50%. As the municipalities in the presence of large industrial facilities alike, the municipalities in FUAs apply the policy of higher local property tax coefficient twice frequently than the rest. Nonetheless, this group of municipalities is much more stable. According to Tab. 5, the share of such as municipalities varies around 46.5% since 2012 onwards. Hence, one can suggest the localisation of a municipality in an FUA is a consequential stimulus contributing to a decision-making of local governments when they opt for higher property taxation.

On the contrary to the preceding analysis taking into account the industrial infrastructure as notable points in the administrative district of a local government, an FUA is composed of the administrative districts in the neighbourhood of a grand city, such as the capital and the seats of regional governments. The FUAs are a continuous areas, instead of irregularly distributed industrial facilities in space; therefore, the effect of tax mimicking should be more obvious between the FUA members. There are 14 FUAs in the Czech Republic and many tax mimicking municipalities belong to some FUA, as presented in Fig. 1. However, the share of such municipalities is stable since 2012, in spite of a significant rise by leaps and bounds from 189 to 280 municipalities opting for high property taxation.

Fig. 1

As presented in the Fig. 1, the number of local governments preferring high property taxation in a FUA varied. There is a chain of municipalities located close to the north-western border to Germany that promote the policy of the higher local property tax coefficient. Those four FUAs are located in the commuting zone determined by the axis Usti nad Labem-Most-Chomutov-Karlovy Vary. Despite the tiny size of these FUAs, they are distinctive by a significant share of municipalities preferring the policy of higher local property tax coefficient. The share varied between 30 (Most) and 35% (Usti nad Labem). However, the MC within particular FUAs comes from to 70 (Karlovy Vary) to 100% (Chomutov and Most).

The largest FUA sprawls around the capital (Praha). Although the share of municipalities applying the higher local property tax coefficient policy is ca 26%, the MC is 87%. The policy is often applied in the southern area of the FUA. The Ostrava FUA in the far-eastern tip of the Czech Republic is similar to the FUA of Praha, that is, 25% local governments opt for the high taxation policy and 95% tax mimickers. In fact, the FUAs of Praha, Ostrava, Usti nad Labem, Chomutov, Most and Karlovy Vary consists of 208 from 599 municipalities (or 280 integrated in FUAs, respectively) that adopted this policy in the Czech Republic.

In other FUAs, the shares of municipalities that apply higher property tax policy are much lower, as a rule; at the bottom, there are FUAs of Hradec Kralove, Pardubice and Olomouc, which counts only between 3 and 4% of appliers. As a result, the MCs are also much lower in these FUAs.

Also, we found another interesting and substantial spatial phenomenon about FUA. In some FUAs, the increasing of LC is mostly applied in the nearest commuting zone around the centre of the FUA, for example, Praha, Brno, Ostrava, Plzen or Usti nad Labem, while in the FUA of Olomouc, Jihlava or Ceske Budejovice, the coefficient is rather applied at the outer borders of FUA. Hence, there are probably different patterns of tax mimicking in the Czech Republic, and this phenomenon should be analysed by further research.

As already presented, both the localisation of industry in a municipality as well as the localisation of municipality in a FUA must be considered as important stimuli for any increasing of the LC decided by local governments. In fact, there are more than two-thirds of municipalities with increased LC, wherein we identified one or both factors (industry and FUA), as it implies from Tab. 6.

As shown in Tab. 6, the share of municipalities with one or both factors affecting the decision-making of the grade of the local property tax coefficient is quite stable since 2013. That finding is in line with the findings presented in previous part of this study. Nonetheless, the simple tax mimicking behaviour is also an important phenomenon. As already presented in Tab. 2, almost three-quarters of municipalities with higher LC were affected by it. Hence, we will proceed an additional method to validate the findings of basic statistics. In particular, we decided to apply the econometrical analysis of Granger causality, which can reveal the links between all three stimuli.

At this point, two facts have to be remind. Firstly, the Granger causation does not mean the unambiguous causation between two variables. It refers to the relationship in which the development of one variable can be better explained by the time-lagged development of other variable than development of its own. Secondly, the results of this procedure are limited, as the volumes we can test are not long enough. Still, the results obtained through this method can help us to understand the possible links and impacts of the property taxation policy.

The results of causality testing summarised in Tab. 7 suggest that there are substantial links among the stimuli under examination. In particular, there is a two-way causation between the large industrial facilities and FUAs. This result is to interpret the existence of cumulative effects of both stimuli that push the municipalities with large industrial facilities located within an FUA to increase the LC of property tax.

The analysis also suggests causation between tax mimicking and increasing of LC in the FUAs. In fact, the mimicking behaviour pushes the municipalities located in the FUAs to increase the LC. On the other hand, the process of increasing the LCs in the FUAs is not intense enough to impact the overall mimicking behaviour significantly. Those results not only validate the previous conclusion about the importance of mimicking in the FUAs, but they confirm also that the tax mimicking is potent behind the borders of FUAs. The analysis suggests causation between tax mimicking and increasing of LC in the FUAs. In fact, the mimicking behaviour pushes the municipalities located in the FUAs to increase the LC. On the other hand, the process of increasing the LCs in the FUAs is not intense enough to impact the overall mimicking behaviour significantly. Those results not only validate the previous conclusion about the importance of mimicking in the FUAs, but they confirm also that the tax mimicking is potent behind the borders of FUAs.

Finally, we found no causation relationships between the tax mimicking and the large industrial facilities in both directions. Similarly, one can consider the process of tax mimicking is not impacted significantly by localisation of industry and vice versa, as mimicking happens in other areas as well. In fact, the large industrial facilities are not strong enough to impact the overall mimicking behaviour significantly. Hence, we can assume that tax mimicking is an independent process that contributes significantly to enlarging the areas of higher property taxation in functional areas of large urban centres.