STRUCTURAL FUNDING AND INTRASTATE REGIONAL DISPARITIES IN POST-COMMUNIST COUNTRIES *

Abstract The intent of this paper is to provide empirical evidence about associations between structural fund allocation in the Czech Republic’s micro-regions for the programming period 2007 to 2013, and intrastate regional disparities by considering the three traditional pillars of sustainable development. The results have mixed evidence, both supporting and not supporting various associations. Economically disadvantaged micro-regions had less structural fund allocation in the economic pillar of sustainable development. To the contrary, socially disadvantaged micro-regions were allocated more structural funds in the social pillar of sustainable development. Therefore, a compensatory effect in sustainable development was observed between economic and social pillars. Results in the environmental pillar were insignifi cant. The results in this paper provide an ambivalent conclusion regarding the contention that spatial distribution of structural funding actually contributes to a reduction in intrastate regional disparities at the micro-regional level. However, micro regions with better economic conditions, e.g. agglomeration economies, better human capital and patent activities – received more structural funding in total.


Introduction
Central and Eastern European (hereafter referred to as CEE) countries have undergone profound economic and political reforms since the 1990s (see, e.g., Ezcurra, Pascual and Rapún, 2007).The transition from central planning to a market economy was accompanied by various societal impacts.These include uneven regional development caused by regions having diff erent abilities to capture benefi ts and avert various losses from reforms (see, e.g., Artelaris, Kallioras and Petrakos, 2010; Barjak, 2001;Smetkowski, 2013;Novák and Netrdová, 2011;Maier and Franke, 2015).In this regard, the success of particular regions was infl uenced, among other factors, by their agglomeration economies, industrial structure, infrastructure endowment, research and technology, human capital, migration, entrepreneurship and SME development, foreign direct investment, and spatial interactions (see, e.g., Banerjee and Jarmuzek, 2010; Ezcurra, Pascual and Rapún, 2007; Blažek and Netrdová, 2012).
Increasing regional disparities within CEE countries stimulated interest in regional policies in the late 1990s.This interest was further strengthened by the eastern enlargement of the EU in 2004 (see, e.g., Pénzes, 2013) because EU structural funding (hereafter referred to as SF) has become the main source for fi nancing regional development (see, e.g., Wokoun, 2007).SF has decisively infl uenced both national and regional structural policies (see, e.g., Czyz and Hauke, 2011).Consequently, research on associations between regional disparities within CEE countries on the one hand, and SF allocation on the other, is fully substantiated from the political viewpoint.The crucial question is whether SF allocation is channeled to 'rich' or to 'poor' regions (see, e.g., Blažek and Macešková, 2010 for this question) or, in other words, whether effi ciency or equity is the fi rst priority (see, e.g., Bański, 2010).Regarding this, Smetkowski (2013), and Czyz and Hauke (2011) point out the shift in emphasis from regionally balanced development to regional competitiveness.This paper is embedded in the aforementioned context concerning regional distribution of SF in the Czech Republic in the programming period 2007-2013.Hence, SF is explored in relation to regional disparities among Czech micro-regions.In this regard, Novosák et al. (2015) did not fi nd evidence for signifi cantly higher SF in lagging regions.However, SF allocation was not thematically divided in this study.Novosák (2016) tackled this drawback and showed more complex associations between SF allocation and regional disparities when considering only the theme of entrepreneurship and SME development.This paper extends our knowledge of associations by focusing on economic, social and environmental pillars of sustainable development.Moreover, more recent data is used.Fundamentally, the goal of this paper is to provide empirical evidence on associations between SF allocation in the Czech Republic's micro-regions in the programming period 2007-2013 and intrastate regional disparities, considering the three traditional pillars of sustainable development.
The paper is structured as follows: the second section provides the theoretical framework.The third section presents data and research methods.The fourth section summarizes results that are discussed in the following section.The last section presents conclusions.

Theoretical framework
The theoretical framework of this paper rests on the literature dealing with determinants of public funding allocation.The special focus is on SF and regional disparities.In this regard, several comments have been made about public funding allocation and its association to regional disparities.The traditional approach to regional policy emphasizes the use of public funding to achieve balanced regional development (see, e.g., Schout and Jordan, 2007).The largest part of public funding should be, therefore, spent in lagging regions (see, e.g., Churski, 2005).Crescenzi (2009), and Crescenzi, De Filippis and Pierangeli (2015) claim, in this context, that public funding ought to be allocated in order to compensate structural disadvantages of various territories.However, several recent theories of regional growth (e.g., theories of regional endogenous growth, new economic geography) question the effi ciency of this approach to regional policy, claiming that spatial concentration could be the preferred strategy to achieve both balanced regional development and also effi ciency (see, e.g., Puga, 2002).De Propris (2007) points out, in this respect, a potential confl ict between equity-oriented and competitiveness-oriented policies.
The associations between SF allocation and regional disparities have been examined by various authors (see, e.g., Crescenzi, 2009;Schraff , 2014; Crescenzi, De Filippis and Pierangeli, 2015; Bouvet and Dall'Erba, 2010; Kemmerling and Bodenstein, 2006;Dellmuth, 2011).According to these studies, more fi nancial resources were spent in less developed regions, indicating concordance between public funding allocation and balanced regional development.However, the fi ndings of these studies were largely predetermined by the choice of SF, choosing the EU15 countries and the NUTS2/NUTS3 spatial levels of analysis.This is because NUTS2/NUTS3 region eligi-bility for SF interventions was directly related to their level of development.On the contrary, the focus of this paper is on micro-regions between LAU1 and LAU2 levels in a post-communist country.The eligibility for SF intervention of almost all micro-regions is crucial because of the direct competition for SF among micro-regions.The associations between public funding allocation and regional disparities are less obvious (see, e.g., Novosák et al., 2015;Hájek et al., 2014;Blažek and Macešková, 2010;Czyz and Hauke, 2011).Concurrently, the knowledge of these associations is highly important for any strategy that deals with regional disparities within countries (see, e.g., Novák and Netrdová, 2011).Note that this is a common issue for all post-communist EU countries.
The research on the associations between SF allocation and regional disparities also has various themes.Dall'Erba (2005), and Mancha-Navarro and Garrido-Yserte (2010) note four themes that are highly relevant for thinking about SF allocation and regional disparities.These themes are: entrepreneurial environment and innovations, human capital and equal opportunities, environment, and transport accessibility.The associations between SF allocation and regional disparities may diff er according to the thematic fi eld.Thus, among others, Kaufmann and Wagner (2005), and also Crescenzi and Rodriguez-Pose (2011) point out that less developed regions often have insuffi cient capacity to absorb funds intended for innovations.Likewise, Novosák (2016) notes less fi nancial intervention intended for lagging regions on more progressive themes.However, there is scarce systematic empirical research on the associations between SF allocation and regional disparities in various thematic areas.This paper contributes to this research by focusing on the three traditional pillars of sustainable development.
Spatial distribution of SF is further infl uenced by other determinants than just regional disparities.These especially include the absorption capacity of the regions, political interests and spatial interactions.Hence, several studies point out the low capacity of lagging regions to absorb SF for at least two reasons -the capacity of regional actors to prepare, submit and administer projects supported by SF, and also the capacity of regional actors to co-fi nance projects supported by SF (see, e.g., Zaman and Georgescu, 2009;Milio, 2007;Tosun, 2014;Dellmuth and Stoff el, 2012).Crescenzi (2009), and Zaman and Georgescu (2009) note that less developed regions have a weaker negotiating position relative to their more developed counterparts.Finally, Dellmuth (2011) claims that low absorption capacity discourages decision-makers from spending SF in less developed regions, because the failure to absorb SF provides an argument against being funded (see, e.g., Dellmuth and Stoff el, 2012; Zaman and Georgescu, 2009;Milio, 2007).
Several insights may be gained on the associations between SF allocation and political interests from literature.Schraff (2014), and Dellmuth and Stoff el (2012) claim that regions with strong support for certain government parties may be preferred 'to reward loyalty'.However, Schraff (2014), and Dellmuth and Stoff el (2012) also note that SF may rather be allocated to regions with a narrow election victory to 'improve' the position of government parties.Dellmuth (2011), and Kemmerling and Bodenstein (2006) suggest, in addition, that leftist and Eurosceptic regions may receive SF for ideological reasons.Finally, spatial interactions also infl uence spatial distribution of SF (see, e.g., Crescenzi, 2009;Camaioni et al., 2013).These interactions may either be positive through spatial cooperation or negative through spatial competition and compensation (see, e.g., Camaioni et al., 2013).

Data and methods
This section describes the methodology used throughout this study.Firstly, dependent variables are introduced.Secondly, explanatory and control variables are defi ned.Thirdly, the methods of data analysis are described.All data refer to the socalled administrative districts of municipalities with extended power, and also to the territory of the capital city of Prague (hereafter referred to as micro-regions; 206 units in total).These micro-regions are administrative territorial units in the Czech Republic, corresponding to the areas between level LAU1 and LAU2.

Data -dependent variables
All dependent variables used in this study relate to SF allocation (in CZK).To this eff ect, the following steps were completed to obtain these variables.Firstly, a list of projects was compiled, containing those submitt ed for SF in the Czech Republic during the programming period 2007-2013, which excluded projects submitt ed under the European Territorial Cooperation objective.The source of information was official data published by the Ministry of Regional Development of the Czech Republic (hereafter referred to as the MRDCR) in October 2015.This also included the location of particular projects (micro-regions).Subsequently, the projects were classifi ed into two groups.The fi rst group involved all projects approved for SF support, while the second group consisted of all projects NOT approved for SF support.Only the fi rst group of projects was used to construct dependent variables.Note that several grant schemes administered by the Ministry of Industry and Trade of the Czech Republic (hereafter referred to as the MITCR), and also by the Ministry of Labor and Social Aff airs of the Czech Republic (hereafter referred to as the MLSACR) were spatially disaggregated in order to bett er tackle spatial distribution of SF.
Both supported and unsupported projects were then thematically classifi ed into one or more categories (see Table 1 for the list of categories).This was achieved by qualitative content analysis of the projects, done by four independent researchers.Thereafter, the categories were linked to the three traditional pillars of sustainable development (see Table 1), which enabled SF allocation to be calculated for each micro-region and each pillar of sustainable development.SF allocation for each micro-region was then divided by its population (annual mean for the period 2007-2015) in order to check the size diff erences of geographic units.Finally, the log-transformation was applied to reduce the infl uence of outliers.Altogether, four dependent variables for further analysis were computed: 1. SF allocation for the economic pillar of sustainable development (ALLOC_ ECON); 2. SF allocation for the social pillar of sustainable development (ALLOC_SOC); 3. SF allocation for the environmental pillar of sustainable development (ALLOC_ ENVI); and 4. SF total allocation (ALLOC_ALL).

Data -explanatory and control variables
Explanatory variables and control variables used in this study are dated back to 2007 or earlier to avoid the endogeneity problem (see, e.g., Crescenzi, 2009; Crescenzi, De Filippis and Pierangeli, 2015, for this approach).Explanatory variables are related to intrastate regional disparities.In this regard, eight indicators were chosen to construct three indexes of regional disparities -the index of economic disadvantage (INDEX_ECON), the index of social disadvantage (INDEX_SOC) and the index of environmental disadvantage (INDEX_ENVI).The choice of the eight indicators was motivated by infl uential theories on regional development.
Population density (DENSITY) is the fi rst indicator used in this study.It is understood as a proxy of spatial concentration of socio-economic activities and agglomeration economies (see, e.g., Rodríguez-Pose and Tselios, 2012).New economic geography, as well as other theories, illustrates the importance of this factor for regional development (see, e.g., Krugman, 1998).The indicator refers to the year 2007 and the Czech Statistical Offi ce (hereafter referred to as CSO) is the source of information.The indicator is log-transformed to improve the normality of its distribution.
The share of tertiary educated people in the population above the age of 15 (TER-TIARY), the share of people with primary school education or with incomplete pri-mary school education in the population above the age of 15 (PRIMARY), and unemployment as the share of unemployed people in the population aged 15-64 years (UNEMPLOY) are perceived as three indicators related to human capital.A number of studies show the impact of human capital on regional development (see, e.g., Rodríguez-Pose and Crescenzi, 2008; Crescenzi, De Filippis and Pierangeli, 2015).Moreover, human capital plays a major role in theories of regional development, including endogenous growth theories (see, e.g., Lucas, 1990) and others.In this study, the indicators TERTIARY and PRIMARY are calculated as the mean of two values from the 2001 and 2011 Censuses.This is due to the missing data in the period between these years.The indicator UNEMPLOY refers to the mean value for the period 2002 to 2007.All data is taken from CSO statistics.
Patent activity (PATENT) is the fi fth indicator used in this study.It is understood as a proxy of innovations that are an essential ingredient for current regional development (see, e.g., Sternberg, 2000).In addition, innovations are embedded in several theories of regional development including regional innovation systems (see, e.g., Tödtling and Trippl, 2005), industrial districts (see, e.g., Becatt ini, 1991), innovative milieu (see, e.g., Maillat, 1998) and others.In this study, the indicator PATENT was carried out as follows.Firstly, patent applications were compiled from the Industrial Property Offi ce of the Czech Republic (hereafter referred to as IPOCR) in the period 2002 to 2007.Secondly, each patent application was assigned to one or more micro-regions according to the applicant's headquarters (weight 50%; evenly distributed among applicants) and the inventor's home (weight 50%; evenly distributed among inventors).Thirdly, a value for each micro-region was summed up and divided by 100,000 inhabitants of the micro-region.Finally, the indicator PATENT was log-transformed to improve the normality of its distribution.
The share of employers and self-employed people in the economically active population (ENTREP) is the sixth indicator used in this study.It is understood as a proxy of entrepreneurship (see, e.g., Fotopoulos, 2012).In this regard, a number of studies point out the positive relationship between entrepreneurship and regional development (see, e.g., Doran, McCarthy and O'Connor, 2016 for a review).In this study, the indicator ENTREP is calculated as the mean of two values from the 2001 and 2011 Censuses because of the missing data in the period between the two Censuses.Data is taken from CSO statistics.
Finally, the last two indicators are net migration per 1,000 inhabitants (MIGRA) and the coeffi cient of ecological stability (ECOL_STAB).Net migration, calculated as the mean value for the period 1996-2007, is understood as a proxy of the quality of life (see, e.g., Navarro-Azorín and Artal-Tur, 2015).The coeffi cient of ecological stability expresses the ratio between environmentally stable and environmentally unstable land-use categories (see, e.g., Buček et al., 2015).This indicator refers to the year 2007.Data is taken from CSO statistics.Principal component analysis (PCA) was employed to reduce the data dimensionality, to extract the most infl uential factors as the indexes of regional disparities, and to eliminate the problem of multicollinearity.In this way, three components with an eigenvalue greater than or almost one (the Kaiser criterion) were extracted, which explains almost 75% of the total variance.Table 2 shows the rotated component matrix and factor loadings.The fi rst principal component (INDEX_ECON) is interpreted as relating to important factors of the economic pillar of sustainable development -agglomeration economies, human capital and innovations.The second principal component (INDEX_SOC) is interpreted as refl ecting important factors of the social pillar of sustainable development -jobs and unemployment, quality of life and human capital.The third principal component (INDEX_ENVI) is interpreted as relating to the most relevant environmental factor -ecological stability.Lower values mean higher economic, social and environmental disadvantage, respectively.Finally, principal component scores were computed from the original variables and used in further analysis (see, e.g., Crescenzi, 2009; Crescenzi, De Filippis and Pierangeli, 2015 for this approach).
Control variables relate to the absorption capacity of micro-regions and to political interests.Absorption capacity is understood as the capacity of regional actors to prepare, submit, administer and co-fi nance projects supported by SF (see, e.g., Tosun, 2014;Dellmuth and Stoff el, 2012).This defi nition is implemented in the following way: the fi rst variable is the total requested SF allocation for projects not supported by SF, according to each inhabitant, and this is log-transformed to improve the normality of distribution (ABSORPTION).The second variable is calculated as the share of projects supported by SF in ratio to the number of all projects submitt ed for SF support (QUALITY).Moreover, the two variables of absorption capacity are also calculated for the three pillars of sustainable development.Political interests are controlled by a dummy variable GOVERNMENT.The value of '1' is assigned to a micro-region if government parties won more than 50% of votes in the Parliamentary elections in 2006 and 2010.Otherwise, the value of '0' is assigned to the micro-region.Finally, the non-eligibility of micro-regions for the convergence objective is controlled by a dummy variable OBJECTIVE 2. The value of '1' is only assigned to the territory of the capital city, Prague.

Methods
Associations between SF allocation and intrastate regional disparities are explored using standard methods of descriptive and inferential statistics, and also regression modeling.In this regard, four regression models are estimated either by OLS (ordinary least squares) or by GMM (generalized method of moments), following the approach suggested by Anselin et al. (1996) and others.The models are in the form: where y i is SF allocation (ALLOC_ALL, ALLOC_ECON, ALLOC_SOC and ALLOC_ ENVI) in a micro-region i; EXPL li is an explanatory variable l in a micro-region i; CONTROL mi is a control variable m in a micro-region i, and u i is the error term.
Regression assumptions were checked by performing residual diagnostics on the four models.Kolmogorov-Smirnov and Shapiro-Wilk tests were computed to verify the normality assumption, and also White's test to verify the homoscedasticity assumption and Moran's I to verify the presence of spatial autocorrelation were done.The multicollinearity assumption was tested by obtaining Variance Infl ation Factor statistics.The diagnostics results did not reveal any violation of the normality and multicollinearity assumptions.The violation of the homoscedasticity assumption was corrected by computing White heteroscedasticity robust standard errors.Finally, spatial autocorrelation was corrected by spatial eff ects modelling, as suggested by Anselin et al. (1996) and others.Therefore, spatial autoregressive coeffi cients ρ and λ were computed for spatial lag and spatial error models respectively.Note that row-standardized fi rst-order queen contiguity weights were used for calculation purposes.

Figure 1 illustrates spatial distribution of SF among micro-regions in the Czech
Republic.Two initial insights may be gleaned from the fi gure.Firstly, a mosaic spatial patt ern of SF allocation is apparent.However, some spatial clusters of both high and low values of SF allocation may be identifi ed.Secondly, the fi gure indicates the presence of two upper outliers.This is caused by the fact that the most fi nancially expensive infrastructure projects are located in these micro-regions.The factor of very large infrastructure projects should therefore be controlled in regression models.Consequently, a dummy variable LARGE_PROJECTS was defi ned in this study.The value '1' was assigned to any micro-region where a project with funding of more than 500 million was implemented.Otherwise, the value of '0' was assigned to each micro-region.In addition, the variable was also decomposed according to the three pillars of sustainable development.
Table 3 extends our research by exploring descriptive statistics of SF allocation for the three pillars of sustainable development.The following fi ndings are noteworthy in this regard: fi rstly, standard deviation values show that SF allocation is more dispersed in the economic pillar (ALLOC_ECON) than in the two other pillars (AL-LOC_SOC, ALLOC_ENVI).However, the lower standard deviation value of total SF allocation (ALLOC_ALL) indicates a compensatory eff ect of the pillars.Secondly, Moran I's values for the three pillars of sustainable development (ALLOC_ECON, ALLOC_SOC and ALLOC_ENVI) are positive and statistically signifi cant.Adjacent micro-regions therefore tend to have similar SF allocation for particular pillars.Moran I's value of total SF allocation (ALLOC_ALL) is nevertheless, statistically insignificant.Hence, a compensatory eff ect of pillars is again observed in the three pillars of sustainable development.Table 4 provides Moran I's values of the three indexes of intrastate regional disparities (INDEX_ECON, INDEX_SOC and INDEX_ENVI).All the values are positive and statistically signifi cant, indicating similar values of the indexes for adjacent micro-regions.This concurs with the above-mentioned fi ndings concerning SF allocation.However, Moran I's values of the three indexes are much higher than those reported for SF allocation (see Table 3).Regional disparities are, therefore, more spatially concentrated than SF allocation.Others have said that SF allocation follows a more spatially balanced location framework than the patt ern of regional disparities, and SF cannot fully compensate for micro-regional disadvantages (see, e.g., Crescenzi, 2009 for similar interpretation of results).Table 5 reports correlation coeffi cients between the measures of SF allocation, and also in intrastate regional disparities.Three associations are statistically signifi cant at the 0.01 signifi cance level.Therefore, the index of economic disadvantage (INDEX_ ECON) is positively associated with SF allocation for the economic pillar of sustainable development (ALLOC_ECON), and also with the total SF allocation (ALLOC_ ALL).The index of social disadvantage (INDEX_SOC) is negatively associated with SF allocation for the social pillar of sustainable development (ALLOC_SOC).Thus, economically disadvantaged micro-regions have relatively low SF allocation for the economic pillar of sustainable development and for the total SF allocation, whereas socially disadvantaged micro-regions have relatively high SF allocation for the social pillar of sustainable development.These fi ndings further support the contention that SF allocation of the three pillars of sustainable development is compensatory in nature.In addition, the essence of this compensation is revealed -SF allocation for the economic pillar of sustainable development is more often channeled to 'rich', while SF allocation for the social pillar of sustainable development is more channeled to 'poor' micro-regions.

Discussion
The preceding section has presented introductory results concerning associations between SF allocation and intrastate regional disparities in the Czech Republic by considering the three traditional pillars of sustainable development.The results indicate that SF allocation for the economic pillar of sustainable development does not favor economically disadvantaged micro-regions, while SF allocation for the social pillar of sustainable development does indeed favor socially disadvantaged micro-regions.Therefore, there is a compensatory nature of SF allocation for the economic and social pillars of sustainable development.Moreover, the total SF allocation is positively and signifi cantly infl uenced by the index of economic disadvantages.This section discusses the relevance of these results, checking for other factors associated with SF allocation -absorption capacity of regions, political interests and spatial interactions.Therefore, regression models were estimated for four dependent variables -SF allocation for the economic pillar of sustainable development (Model 1), SF allocation for the social pillar of sustainable development (Model 2), SF allocation for the environmental pillar of sustainable development (Model 3), and the total SF allocation (Model 4).
The regression model estimates are given in Table 6.The parameters and model statistics support the fi ndings detailed in the preceding section.At the 0.01 signifi cance level there is a positive and statistically signifi cant impact in the index of economic disadvantage (INDEX_ECON) on SF allocation for the economic pillar of sustainable development (ALLOC_ECON) (Model 1).The impact in the social disadvantage index (INDEX_SOC) on SF allocation for the social pillar of sustainable development (ALLOC_SOC) is, on the contrary, negative and statistically signifi cant at the 0.01 signifi cance level (Model 2).Moreover, model 4 illustrates the compensatory nature of SF allocation for the economic and social pillars of sustainable development.Hence, the impact in the index of economic disadvantage (INDEX_ECON) on the total SF allocation (ALLOC_ALL) is still positive and statistically signifi cant at the 0.01 signifi cance level.Nevertheless, the regression coeffi cient is lower than the one obtained earlier (Model 1).The impact in the index of social disadvantage (INDEX_SOC) on the total SF allocation (ALLOC_ALL) is still negative but statistically insignifi cant.
The sign and statistical signifi cance of the estimated coeffi cients of control variables provide additional insights into understanding the determinants of SF allocation.Firstly, absorption capacity of micro-regions -i.e. the capacity to create (AB-SORPTION) competitive (QUALITY) projects -is the crucial factor in determining spatial distribution of SF, regardless of the sustainability pillars.The association between absorption capacity of micro-regions and SF allocation is positive and statistically signifi cant at the 0.01 signifi cance level with one exception.Hence, high SF allocation required for unsupported projects (ABSORPTION) and the high percentage of accepted projects (QUALITY) tend to be prerequisites for high SF allocation.Secondly, the presence of very large infrastructure projects (LARGE_PROJECTS) generally has positive impacts on SF allocation.The impact is statistically signifi cant at the 0.01 signifi cance level for the social pillar of sustainable development (ALLOC_ infrastructure projects and checking unequal absorption capacities of micro-regions.Consequently, a compensatory eff ect of the economic and social pillars of sustainable development is observed.However, a positive and signifi cant impact of the economic pillar of sustainable development on the total SF allocation is revealed.Micro-regions with bett er economic conditions -agglomeration economies, human capital and patent activities -receive more SF in total, in other words.Overall, the results in this paper provide ambivalent evidence in the debate that spatial distribution of SF contributes to reduction of intrastate regional disparities at the micro-regional level.
The results in this paper are relevant in political practice.Firstly, it has been shown that thematic decomposition is important for understanding complex associations between SF allocations to that of intrastate regional disparities.Aggregate fi gures may mask diff erences across issue areas, providing incomplete information for political decision-making.This relates to the argument that there is a compensatory eff ect of economic and social pillars in sustainable development.Earmarking SF for these two pillars may substantially infl uence regional distribution of SF.Furthermore, the way of defi ning regional disparities may completely change conclusions concerning associations between SF allocation and intrastate regional disparities.
The results in this paper are further relevant for considering the coherence of regional policies and sector policies, or the links between the goals of regional policies and sector policies.Hence, these results may relate to the discussion about the equity and effi ciency goals of regional policies.The question is whether SF is allocated to lagging regions or to the most competitive regions.This paper indicates that the answer is not straightforward and that various results may be obtained depending on the sector policy analyzed.Altogether, these associations should be considered in order to have more congruent policies.
Finally, the results are useful for understanding the coherence of regional policies formulated at diff erent spatial levels.This is an important issue for post-communist EU countries with respect to the spatial coherence of EU cohesion policy and national regional policies.In this regard, almost complete territories of post-communist EU countries are eligible for funding under the Convergence objective of cohesion policy.Moreover, SF is recently the main source of fi nancing for regional development in post-communist EU countries.However, their national regional policies usually select assisted regions at a lower spatial level rather than considering cohesion policy, refl ecting intrastate regional disparities.Consequently, SF allocation in assisted regions is crucial in order to evaluate the spatial coherence of EU cohesion policy and national regional policies.

Table 1 :
Categories of themes and pillars of sustainable development -associations Source: Compiled by the authors.

Table 2 :
Rotated component matrix (PCA, Varimax with Kaiser normalising) Compiled by the authors; data from the CSO and IPOCR.

Table 4 :
Intrastate regional disparities -Moran's I Compiled by the authors; data from CSO and IPOCR. Source: