Modelling severe material deprivation rates in EU regions using fractional response regression

Using Eurostat data for 2016, this study assesses the impact of various economic factors on severe material deprivation rates (SMDR) in European Union (EU) regions. As values of the analysed response variable range between 0 and 1, the study applies
of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Regional Statistics, Vol. 9. No. 2. 2019 Online first Dudek –Sedefo ğ lu 1–18; DOI: 10.15196/RS090210 Modelling severe material deprivation rates in EU regions using fractional response regression Hanna Dudek  Warsaw University of Life Sciences, Poland E-mail:   Gül ş ah Sedefo ğ lu Istanbul University,  Turkey E-mail:   Keywords: European Union, NUTS, severe material deprivation rate (SMDR), odds ratios, fractional response regressionUsing Eurostat data for 2016, this study as-sesses the impact of various economic factors on severe material deprivation rates (SMDR) in European Union (EU) regions. As values of the analysed response variable range between 0 and 1, the study applies fractional regression as well as commonly used linear regression. Results of the extended RESET test indicate that the linear model suffers from misspecifi-cation, while there is no reason to reject the hypothesis regarding the adequacy of fraction-al response models (FRM). Therefore, to as-sess the unitary impact of explanatory varia-bles on expected values of SMDR, a fractional response regression model with logit link function is used, which enables interpretations of odds ratios. It was found that SMDR is af-fected by regional-level factors such as median equivalised disposable household income, at-risk-of-poverty rate (ARPR), gross domestics product (GDP) per capita, long-term unem-ployment rate, and country-level drivers such as relative median at-risk-of-poverty gap, in-come quintile share ratio, and share of social protection expenditure in GDP. Introduction In recent years poverty has commonly been recognised as a state of multidimen-sional deprivations (Alkire et al. 2015; Asselin 2008). In order to identify poverty,  various indicators that capture this phenomenon are incorporated into diverse anal-yses, depending on the context of the country or purpose of the research. Among others, a multidimensional approach is used in the European Union (EU), where three indicators (ARPR, SMDR, people living in very low work intensity house-holds) are taken into account to assess the effect of implementation of the Europe 2020 strategy on the risk of poverty or social exclusion (Eurostat 2018a). These three indicators, which form the headline indicator, represent related but distinct aspects of poverty or social exclusion. It should be stressed that such an approach is  2 Hanna Dudek – Gül ş ah Sedefo ğ lu Regional Statistics, Vol. 9. No. 2. 2019 Online first Dudek –Sedefo ğ lu 1–18; DOI: 10.15196/RS090210 a result of political coordination involving a series of compromises between political and policy preferences and traditions of member states (Maître et al. 2013).  The first measure, ARPR, is the share of people with an equivalised disposable income below the at-risk-of-poverty threshold, which is set at 60% of the national median equivalised disposable income. The next, SMDR, is an indicator adopted by the EU Social Protection Committee that measures the percentage of the population that cannot afford at least four of the following nine items (Eurostat 2017, Guio et al. 2012): 1) to pay their rent, mortgage or utility bills; 2) to keep their home adequately warm; 3) to face unexpected expenses; 4) to eat meat, fish or a protein equivalent every second day; 5) to go on a week-long holiday away from home; 6) to have a television set; 7) to own a washing machine; 8) to have a car; 9) to own a telephone. The last indicator is the share of people aged 0–59 years living in households in which the members of working age (18–59 years, excluding students) worked less than 20% of their total potential during the past year. The definition of the EU’s poverty and social exclusion headline target is based on a combination of these three indicators. Thus, it extends the traditional concept of income poverty to cover material deprivation and labour market exclusion, reflecting the multiple aspects of poverty and social exclusion across Europe. It should also be noted that the SMDR reflects absolute poverty, while the ARPR captures relative poverty. This is because absolute measures define poverty on the basis of a normative judgement of, for example, what qualifies as basic needs,  wherein the level of fulfilment of considered needs is not compared to the level of fulfilment of needs of other society members. Alternatively, relative measures of poverty fix an arbitrary threshold relative to a typical standard in society (Boarini– D'Ercole 2006).  This study analyses the 2016 data as part of the mid-term review for Europe 2020. We focus on the issue of material deprivation. There are many reasons for our interest in this issue. First, the indicator ‘people living in very low work intensity households’ is a measure of social exclusion in the area of the labour market and should not be included in the analysis of poverty, as some experts suggest (Panek– Zwierzchowski 2016). Furthermore, the use of the subsequent indicator ARPR in regional analysis has some limitations. In our opinion, the comparison of living condi-tions in different EU regions on the basis of ARPR raises doubts due to the fact that the poverty threshold is computed separately in each country. As the poverty threshold changes over time and also varies by country, the ARPR approach does not permit a constant benchmark of poverty to be set, which would allow compari-sons of poverty across time and space (Panek–Zwierzchowski 2014). Moreover, according to the Eurostat glossary (2017), the ARPR does not measure wealth or pov-erty, but low income in comparison with other residents in that country, which does not necessarily imply a low standard of living. In contrast to ARPR, the SMDR is an absolute indicator enabling regional comparison within the EU. If only relative na-  Modelling severe material deprivation rates in EU regions using fractional response regression 3 Regional Statistics, Vol. 9. No. 2. 2019 Online first Dudek –Sedefo ğ lu 1–18; DOI: 10.15196/RS090210 tional poverty thresholds were considered, the risk of poverty would seem to be rather similar among the entire EU, masking great differences in living standards among the population (Israel–Spannagel 2013). Instead, considering the inability to afford a particular item or activity is taken to represent the same level of deprivation irrespective of how many other people in the same country are in that situation (Nolan–Whelan 2010). Moreover, the material deprivation indicator seems to take the actual standard of living that people enjoy into account better as it measures the inability to consume goods and services that are seen as basic necessities in the EU conditions. Lastly, analysis of items capturing the same aspect of deprivation in the entire EU enables comparability between regions or countries (Guio 2009).  The current study examines the relationships between SMDR in EU regions and  various regional- and country-level economic factors. It focuses on regional-level analysis because regions, not countries, are the key elements of EU policy (Becker et al. 2010). Thus, knowledge of regional differences in material deprivation are crucial for targeted anti-poverty policies (W  ęź iak–Bia ł owolska 2015). Our study also veri-fies econometric methodology by comparing the appropriateness of fractional re-sponse regression and commonly used linear regression. It applies Papke and  Wooldridge (1996) methodology suggested to handle fractional response data.  The rest of this paper is organised as follows. First, we briefly review the re-search on correlates of material deprivation in the EU. In subsequent sections, we describe applied data and methods. Next, we present obtained results and discus-sion of these. Finally, in the last part of the paper, we summarise the results and provide concluding remarks. Literature review  As this study analyses material deprivation in the context of the Europe 2020 strategy, in this section we focus only on the literature related to modelling this phenomenon in the EU.  The first group of studies analyses correlates of material deprivation bases on micro-data analysis describing a households’ behaviour. Studies in this field use mainly logit or probit models, in which the binary variable usually assumes the value of 1 if material deprivation occurs and 0 otherwise. For example, Nelson (2012), Rezanková and Želinský (2014), Bárcena-Martín et al. (2014), Šoltés and Ulman (2015), Israel (2016), Bruder and Unal (2017), and Saltkjel (2018) examine the impact of various socio-demographic and economic factors on probability of being deprived. In particular, most of the aforementioned studies confirm the role of such characteristics as the household structure, place of residence, income situation, as  well as educational achievements and labour market status of the household’s head.  The second group of studies concerns the modelling of SMDRs at the country level. For example, using simple two-dimensional analyses, Israel and Spannagel  4 Hanna Dudek – Gül ş ah Sedefo ğ lu Regional Statistics, Vol. 9. No. 2. 2019 Online first Dudek –Sedefo ğ lu 1–18; DOI: 10.15196/RS090210 (2013) reveal negative dependence of the median of households’ equivalent income and positive dependence of households’ income inequality. Acar et al. (2017) and B. Kis–Gábos (2016) find a negative relationship between the material deprivation rate (MDR) and GDP per capita. Moreover, Nelson (2012) highlights the role of social assistance benefit levels in the decrease in MDRs, and Whelan and Maître (2012) find a negative impact of government social expenditure as a percentage of GDP. Using more advanced econometric analyses, Blatná (2017) applies an autoregressive distributed lag (ADL) model to analyse EU MDRs for 2005–2015. The author finds that the material deprivation rate in the EU-28 depends on the proportion of people living in households with very low work intensity and of people with a secondary or lower level of education. Calvert and Nolan (2012) and B. Kis et al. (2015) use linear panel data models. In the former, the authors highlight the meaningful role of medi-an income and income inequality in explaining the rate of material deprivation. Fur-thermore, grouping countries into three clusters classified according to the median level of households’ income and estimating the models for each group separately, Calvert and Nolan (2012) find that the impact of both these variables are statistically significant only in low-income countries. B. Kis et al. (2015) consider a wide set of potential drivers, wherein they reveal a significantly positive relationship between SMDR and indicators of income poverty as well as the share of young people, while there is a significantly negative association with households’ average income, house-holds’ savings rate, and employment rate. It is also worth mentioning Dudek’s study (2019), in which the generalised estimating equations (GEE) method is applied to analyse country-level panel data. The author finds that SMDR    in 2008–2015 was affected by such factors as median equivalised disposable income, relative median at-risk-of-poverty gap, long-term unemployment rate, GDP per capita, the share of social protection expenditure in GDP, and income inequality indices . Moreover, she demonstrates that GEE models for a fractional   response variable exhibit better goodness of fit than linear models.  There is a lack of studies relating to the analysis of material deprivation at re-gional level. The few exceptions include Želinský (2012), who finds that there are significant differences in the material deprivation rates between and among Czech and Slovak regions. In particular, the level of deprivation is found to be higher in Slovakia, and deprived households are highly concentrated in the eastern part of this country. In terms of studies relating to SMDR in EU regions, using 2014 data, Dudek (2018) reveals the role of regional-level factors such as long-term GDP per capita, unemployment rate, median equivalised disposable income, and ARPR. This study aims to provide a deeper insight into the issue of regional severe material depri- vation by considering both regional and country-level drivers. Moreover, it consid-ers 2016 data, while Dudek (2018) examines the situation referring to 2014 data.  Modelling severe material deprivation rates in EU regions using fractional response regression 5 Regional Statistics, Vol. 9. No. 2. 2019 Online first Dudek –Sedefo ğ lu 1–18; DOI: 10.15196/RS090210 Data  The current study aims to analyse data on poverty indicators at a regional level using the most detailed Nomenclature of Territorial Units for Statistics (NUTS) level. The NUTS classification is a hierarchical system for dividing the economic territory of the EU 1 . The NUTS classification, specified in Regulation (EC) No 1059/2003 of the European Parliament and of the Council, comprises three levels covering levels 1, 2, and 3 from larger to smaller areas for each member state. However, depending on the country size, not all levels are available for each country (Haldorson 2019, Pathy 2017). For the size of NUTS regions, minimum and maximum population thresholds are defined in the regulation as a principle of the NUTS classification (Brandmueller et al. 2017). Another principle is that regions are composed by aggre-gating smaller administrative regions, as there is no administrative layer in the mem-ber states corresponding to a particular level (Eurostat 2018b). In this study, the analysis at the regional level is restricted to countries for which relevant data is available. For the NUTS 1 and NUTS 2 level regions, Austria, Bul-garia, Cyprus, the Czech Republic, Denmark, Estonia, Greece, Spain, Finland, Croa-tia, Hungary, Italy, Lithuania, Luxembourg, Latvia, the Netherlands, Poland, Roma-nia, Sweden, Slovenia, and Slovakia are included in the analysis; regions in Belgium, Germany, France, the United Kingdom, Ireland, Malta, and Portugal are excluded 2 .  The empirical analysis is based on the following variables measured at the re-gional- and country-level. To variables measured on regional-level belong:  –    GDP per capita expressed in thousands of purchasing power standard (PPS) units (GDP per capita);  –    long-term unemployment rate, meaning the number of persons unemployed for 12 months or longer as a percentage of the active population (L_unemployment);  –    the proportion of the population aged less than 60 years living in households  with very low work intensity (Low_work_intensity);  –    median equivalised disposable household income expressed in PPS (Income);  –     ARPR.  The following are country-level variables:  –    the relative median at-risk-of-poverty gap, which is calculated as the differ-ence between the median equivalised net income of persons below the at-risk-of-poverty threshold and the threshold itself, expressed as a percentage of the at-risk-of-poverty threshold; this is set at 60% of the national median 1  For more details, see Eurostat (2016). 2  For Cyprus, Estonia, Luxembourg, and Latvia, no NUTS 2 level regions are described depending on their country sizes. For Lithuania, only two regions are described in NUTS 2 level, however, data is available only at country-level. Thus, the aforementioned countries are included in the analysis at NUTS 1 level with other NUTS 2 level regions. The total number of regions included in the analysis is 162.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!