A multidimensional approach to child poverty in Taiwan

A multidimensional approach to child poverty in Taiwan

Children and Youth Services Review 66 (2016) 35–44 Contents lists available at ScienceDirect Children and Youth Services Review journal homepage: ww...

414KB Sizes 2 Downloads 124 Views

Children and Youth Services Review 66 (2016) 35–44

Contents lists available at ScienceDirect

Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth

A multidimensional approach to child poverty in Taiwan Chao-Hsien Leu a, Ke-Mei Chen b,⁎, Hsiu-Hui Chen a a b

Department of Social Work, Tunghai University, No.1727, Sec.4, Taiwan Boulevard, Xitun District, Taichung 40704, Taiwan Department of Social Welfare, National Chung Cheng University, 168 University Road, Minhsiung Township, Chiayi County 62102, Taiwan

a r t i c l e

i n f o

Article history: Received 6 February 2016 Received in revised form 22 April 2016 Accepted 24 April 2016 Available online 27 April 2016 Keywords: Deprivation Social exclusion Multidimensional poverty Child poverty Fuzzy sets Seemingly unrelated regression (SUR) models

a b s t r a c t To elucidate the multidimensional nature of poverty, this study analyzed child deprivation and social exclusion in Taiwan. First, a fuzzy set approach was used to construct an aggregate poverty index, to measure the levels of perceived necessity, deprivation, and social exclusion experienced by children. The study involved conducting a decomposition analysis to measure the poverty index according to certain dimensions. Second, this study involved analyzing possible determinants of perceived necessity, deprivation, and social exclusion, using seemingly unrelated regression models. We used cross-sectional data obtained from the Household Living Conditions Survey conducted in 2014. The results suggest that over two-thirds of the respondents identified all the items as necessary. Three highest levels of perceived necessity were housing, medical care, and clothing dimensions. Children faced high risks of deprivation and exclusion. The three highest levels of deprivation and exclusion were exhibited in the dimensions of environment, recreation, and education; the lowest two levels of deprivation and exclusion were exhibited in the dimensions of medical care and housing. The dimensions with higher levels of deprivation and exclusion exhibited higher relative contributions to facilitating poverty reduction. Moreover, evaluation of income and expenditure, family income, and family type were significantly related to the degree of perceived necessity and the levels of deprivation and exclusion. Those living in families with a large number of children exhibited a higher level of deprivation. Education of the caregivers was closely linked to social exclusion of children. This paper represents preliminary and small-scale research; however, several implications for methodology and policy can be derived from this study. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Poverty is a complex concept with various definitions (Spicker, Leguizamon, & Gordon, 2007) and cannot be understood appropriately using simple measures referring to income or consumption; diverse methods of measurement are required, because results obtained from only a single measure of poverty lack reliability and validity. Policy responses also vary depending on the type of poverty measure applied. Thus, to identify poverty and draw conclusions, it should be measured using triangulation (Bradshaw, 2001; Bradshaw & Finch, 2003). Over the past few decades, the focus in poverty research has shifted from traditional unidimensional aspects to multidimensional aspects. Prominent concepts such as relative deprivation, proposed by Townsend (1979), and the capability approach, introduced by Sen (1985, 1999), have contributed to the development of a multidimensional perspective on poverty. Several non-monetary indicators have been widely considered as proxies that enable identifying various aspects of poverty; for instance, material deprivation and social exclusion (Bossert, D'Ambrosio, & Peragine, 2007; Jana, Nad'a, & Jana, 2012; Mack & Lansley, 1985; Menchini & Redmond, 2009; Nolan & Whelan, 2010). ⁎ Corresponding author. E-mail address: [email protected] (K.-M. Chen).

http://dx.doi.org/10.1016/j.childyouth.2016.04.018 0190-7409/© 2016 Elsevier Ltd. All rights reserved.

Increasing attention has been focused on identifying and making comparisons between various dimensions of poverty, to understand it more clearly (Bradshaw, 2001; Bradshaw & Finch, 2003; Main & Bradshaw, 2012; Saunders, 2008, 2011). In Australia, Saunders and colleagues employed income, deprivation, and social exclusion approaches to analyze people who experience social disadvantage, detect the similarities and differences between the employed approaches, and explore related factors (Saunders, 2008, 2011; Saunders & Naidoo, 2009; Saunders, Naidoo, & Griffths, 2008). A multidimensional approach has informed comparative studies that have examined and compared poverty across cities, regions, and countries (Batana, 2013; Battiston, Cruces, Lopez-Calva, Lugo, & Santos, 2013; D'Ambrosio, Deutsch, & Silber, 2011; Dewilde, 2004; Roelen, 2014; Waglé, 2005, 2008; Whelan & Maître, 2007; Yu, 2013). Considerable variation has been identified in the dimensions and measures of poverty. Although multidimensional approaches have been commonly applied to poverty research, only a few studies have investigated the relationships among poverty, social disadvantage, and social exclusion (e.g., Lee, 2007, 2011). Most research on poverty in Taiwan has focused primarily on unidimensional poverty measurement, which is related to monetary poverty (e.g. Ho, 2007; Ho, Wang, & Leu, 2003; Leu, 1996, 2010b, 2010c; Wang, Ho, & Liu, 2008). Even in child poverty research, income poverty has often constituted a major approach (e.g. Hsueh,

36

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

2008; Lee & Wang, 2008; Leu, 2010a). Studies that have employed an income-based approach have exhibited data limitations. The income poverty approach has limitations. Income defined as an “indirect” concept of poverty cannot represent the standard of living of a family or an individual (Ringen, 1988). Moreover, the income poverty approach involves assuming intra-household sharing of resources and uses an equivalence scale to adjust for individual needs. Under the assumption of sharing, this approach cannot reflect children's needs and may overestimate poverty among children and underestimate poverty among parents (Saunders, 2010). Household income cannot comprehensively reflect children's material circumstances (Main & Bradshaw, 2012). Therefore, the income-based approach typically fails to adequately identify child poverty. The measures of child poverty are based on a child-centered approach, to ensure that children's perspectives are considered. Children's experiences of poverty differ from those of adults (Grødem, 2008). Their needs are not expressed through their parents; thus, a child-centered approach enables understanding various aspects of poverty from children's perspectives (Main & Bradshaw, 2012). Children are active participants and construct their own understanding of poverty. Evidence from Ireland indicated that children and adults have unique perceptions and experiences of deprivation and social exclusion (Kerrins, Greene, & Murphy, 2011; Swords, Greene, Boyd, & Kerrins, 2011), implying that children's experiences of poverty should be analyzed separately. Consequently, this study employed two types of poverty measure, deprivation and social exclusion, rather than relying on only one measure. Regarding measurements, the methods of multidimensional analysis have been discussed in previous studies (e.g. Alkire & Foster, 2011; Batana, 2013; Belhadj, 2011, 2013; Betti, Cheli, Lemmi, & Verma, 2008; Dewilde, 2004; Giordani & Giorgi, 2010; Neff, 2013; Whelan, Nolan, & Maître, 2014). An analytical method based on fuzzy set theory has been widely adopted in the study of poverty. The fuzzy set approach proposed by Zadeh (1965), as a method of using imprecise data, indicates that poverty measurement should go beyond a dichotomy of poverty (Belhadj, 2011; Betti & Verma, 2008; Betti et al., 2008; Cerioli & Zani, 1990; Mussard & Pi Alperin, 2005; Pi Alperin, 2008). As Belhadj (2011) suggested, poverty is not an absolute dichotomy between poor and non-poor but rather “a matter of degree” (p. 687). Apart from constructing aggregate levels of poverty, the fuzzy set approach can enable deriving a poverty index according to attribute, group, region, and country (Pi Alperin, 2008). In this study, the fuzzy poverty approach enabled estimating relative levels of deprivation and social exclusion. Insufficient research has been conducted on the multidimensionality of child poverty in Taiwan because of the limited availability of data. The purpose of this study was to identify and understand child poverty more holistically by using a multidimensional approach. The specific aims of this study were the following: (1) measuring the levels of perceived necessity, deprivation, and social exclusion experienced by children, using the fuzzy poverty approach; (2) examining possible determinants of perceived necessity, deprivation and social exclusion by applying seemingly unrelated regression (SUR) models. This study constitutes preliminary research, but may have methodological and policy implications. 2. Background During the last half century, the notion that deprivation and social exclusion reflect the multidimensionality of poverty has received increasing attention. Deprivation and social exclusion are not only treated as measures of poverty but are also used as indicators of social disadvantage (Saunders, 2011). The concept of deprivation, initially introduced by Townsend (1979), has emerged as an alternative approach to the analysis of poverty. Poverty exhibits several characteristics: (1) unsatisfied physical and social needs; (2) the lack of resources; and (3) the need to draw a comparison to the lifestyles of others. Thus, poverty is viewed as an objective term that can be understood through the

concept of “relative deprivation.” People are deprived if they live below the socially accepted standard of living (Townsend, 1979). Deprivation, in this context, differs from poverty; deprivation can be experienced without being poor. However, individuals and families can be considered to be in poverty if they lack an adequate standard of living and the ability to fulfill social obligations (Townsend, 1987). Townsend's pioneering work on relative poverty transcended the absolute and narrow definition of poverty. Deprivation defined in a broad manner can be extended to individual lifestyles and the capacity to participate in social activities, implying that the experiences of poverty are diverse. Despite Townsend's considerable contribution to poverty research, he has been criticized for his definition of deprivation. Mack and Lansley (1985) argued that Townsend's concept of deprivation did not distinguish need from choice. The absence of certain necessities may depend on income level or choice. Mack and Lansley (1985) defined deprivation as “an enforced lack of socially perceived necessities” on the basis of a consensual approach, to explain that people may lack socially perceived necessities because of an inability to afford them, rather than a personal choice or preference. Hence, based on the consensual approach, people are considered to be in poverty if they fall below minimum standards of living (p. 34–45). The consensual approach has been widely used to measure deprivation (e.g. Gordon & Pantazis, 1997; Gordon et al., 2000a; Hillyard, Kelly, McLaughlin, Patsios, & Tomlinson, 2003; Swords et al., 2011). It can be used as a headcount measure of poverty, for example, by quantifying material and social deprivation. Moreover, an effective measure of deprivation can identify those who are deprived because of constraints or choices (Gordon, Pantazis, & Townsend, 2000b). Social exclusion has been identified as a principal concern in the study of poverty and become a dominant concept in the United Kingdom, the European Union, and other countries. The concept was introduced in 1974 by Lenoir and means that people are excluded from social protection. In the United Kingdom, the notion of social exclusion was first used in Townsend's studies on poverty, which investigated people's exclusion from normal life (Levitas, 2006). Social exclusion is difficult to define and varies according to country (Levitas, 2006; Silver & Miller, 2003). Although the definition of social exclusion is imprecise, Atkinson (1998) suggested three components to identify social exclusion clearly, namely, relativity, agency, and dynamics. Relativity indicates that exclusion is perceived based on comparisons with the situations of others in a particular place and at a particular time. Agency refers to the notion that social exclusion may result from self-exclusion or the actions of others. Dynamics means that social exclusion has to be examined over time (Atkinson, 1998). To improve operational efficacy, Levitas et al. (2007, p25) proposed the following working definition of social exclusion: …involves the lack or denial of resources, rights, goods and services, and the inability to participate in the normal relationships and activities, available to the majority of people in a society, whether in economic, social, cultural or political arenas. [Levitas et al. (2007), p. 25.] Deprivation and social exclusion have commonalities. The two types of social disadvantage reflect a multidimensional nature and are characterized by relativity and a lack of participation. Socially excluded people feel disconnected from their community and society; however, the causes of social disadvantage vary. Regarding deprivation, nonparticipation is attributed to the absence of resources. About social exclusion, social disadvantage is not inevitably related to the absence of deficient resources and may occur because of other factors such as discrimination or ill health (Burchardt, 2000). The concept of social exclusion involves emphasizing how factors such as relationships, institutions, and behavioral patterns influence the exclusion of people from their community (Saunders, 2010). Identifying the differences among poverty (often measured using income), deprivation, and social exclusion is crucial because various

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

indicators of three concepts partly overlap; for instance, lack of income reflects economic deprivation. Those who experience deprivation may not be poor and those who are poor may not experience deprivation. Poor people may not be socially excluded and excluded people may not be in poverty (Saunders, 2010). Saunders's analytic framework facilitates a clear distinction between deprivation and social exclusion (Saunders, 2008, 2011; Saunders et al., 2008). The framework comprises three stages of identification: (1) identifying the essentials of life by asking whether certain items are necessary; (2) identifying deprivation by confirming that people lack the essentials of life because of an inability to afford them; and (3) identifying social exclusion that occurs by choice, rather than a lack of affordability (Saunders, 2008). Saunders's analytic framework enables identifying the essentials of life and two specific types of social disadvantage. The identification of the essentials is a vital prerequisite for deprivation and social exclusion. 3. Methods 3.1. Study design and sampling The cross-sectional data in the study were derived from the Household Living Conditions (HLC) Survey conducted in 2014. Despite being a small-scale preliminary study, the pilot survey facilitated a basic understanding of multidimensional poverty among children. The survey employed a stratified cluster sampling method and divided the target population living in Chiayi City and Chiayi County into three strata: city, urban townships, and rural townships. A total of 20 primary schools were selected using a simple random sampling process. Subsequently, two classes, randomly chosen from each school, were used as the sample. The unit of analysis indicates children in Grades 5 and 6. Children and their caregivers completed the questionnaire together. Caregivers were defined as parents, grandparents, and, other relatives. Most questions regarding child-specific items and activities were answered by children and questions about demographic statistics, family income, evaluation of income and expenditure, and school participation were answered by caregivers. The random sample consisted of 800 households, and generated a response rate of 97% (n = 778). 3.2. Ethical statement Ethical approval was obtained from the National Cheng Kung University Research Ethics Committee for Human Behavioral Sciences in Taiwan. Written informed consent was obtained from all children and their caregivers. 3.3. Measures The HLC survey comprised 53 items measuring the level of perceived necessity, level of deprivation, and level of social exclusion. The structure of questions was designed as suggested by Saunders (Saunders, 2008, 2011; Saunders et al., 2008). An expert panel comprising various scholars and experts on child poverty assessed the questionnaire regarding its logic, appropriateness, and content. Regarding the reliability of the 53 items, the Cronbach's alpha coefficient of the scale was 0.873 for the essential items and 0.818 for the items of deprivation and social exclusion, indicating excellent internal consistency (Cronbach, 1951). The 53 items were grouped into eight dimensions, listed in the Appendix Table A. The eight dimensions that did not comprise monetary indicators were: diet, clothing, medical care, education, recreation, environment, economic and social relationships, and housing. Most items were child-specific; others were related to the children's families. To measure the level of perceived necessity, respondents were asked if each of these items was essential for them. Regarding the measurement of deprivation and social exclusion, respondents were given five optional answers for each item and asked to choose the one that best described their child or children and family. The five optional answers

37

were: (1) I/my family have the item; (2) I/my family do not have the item because of an inability to afford it; (3) I/my family do not have the item because of inadequate access to opportunities or a lack of time to purchase it; (4) I/my family do not have the item because of a lack of need or want; (5) I/my family do not have the item because of others reasons. The respondents exhibited deprivation when choosing the second answer. Social exclusion related to the items that they were forced or chose not to have/do was indicated when the second to fifth options were selected. In other words, social exclusion involves both voluntary and involuntary exclusion, reflecting subjective choices and objective limitations.1 In this study, the concept of deprivation overlapped with that of social exclusion. The concept of deprivation suggested the relevance of economic resources, whereas social exclusion indicated the involvement of economic and noneconomic aspects. Appendix Table A shows the percentage of respondents who considered each item necessary, those who had each item, and (for those who did not have the item) the various reasons for not having the item. To explore the factors contributing to the levels of perceived necessity, deprivation and social exclusion, several independent variables were included in the analysis, including caregiver education, caregiver participation in school, evaluation of income and expenditure, family income, family type, location, number of children, and child's gender (Table 1); except for caregiver education and number of children, all were discrete variables. In addition, three dependent variables, the level of perceived necessity, level of deprivation, and level of social exclusion, were continuous variables. 3.4. Analytic methods This study involved applying two types of statistical analyses to elucidate multidimensional poverty among families with children to achieve two specific aims: (1) measuring the levels of perceived necessity, deprivation, and social exclusion; (2) analyzing possible determinants of perceived necessity, deprivation, and social exclusion. 3.4.1. Fuzzy set approach A fuzzy set theory, which was first applied to the measurement of poverty by Cerioli and Zani (1990), was used to measure the levels of perceived necessity, deprivation, and social exclusion. The fuzzy poverty approach enables measuring a multidimensional poverty index (MPI), a unidimensional poverty index (UPI) for each of the dimensions, and their contributions to the total poverty index. This has enabled policy makers to alleviate poverty by identifying various dimensions of poverty, including monetary and nonmonetary dimensions (Costa & De Angelis, 2008; Pi Alperin, 2008). The study exclusively focused on non-monetary dimensions and applied the UPI to measure the levels of perceived necessity, deprivation, and social exclusion corresponding to each dimension. This study relied heavily on previous studies employing the concept of fuzzy sets by Costa and De Angelis (2008); Dagum and Costa (2004); Mussard and Pi Alperin (2005), and Pi Alperin (2008). The MPI and UPI are defined as weighted functions. The method of weighting proposed by Cerioli and Zani (1990) is an inverse function of the level of poverty corresponding to a given item, assuming that an item should be given greater weight if it exhibits a low frequency. In other words, more importance is given to the low-frequency items. A wider distribution among the items indicates that people experience stronger feelings of deprivation (Miceli, 1998). The weight for each item is a value between 0 and 1. Regarding material deprivation, for example, the weight refers to the intensity of deprivation. A television, regarded as an essential item, receives a strong weight, because fewer people are deprived of 1 See Atkinson (1998); Barry (2002); Burchardt et al. (1999, 2002b); Burchardt, Le Grand, and Piachaud (2002a) and Le Grand (2004) for a more detailed discussion on voluntary and involuntary nonparticipation/exclusion.

38

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

Table 1 Descriptions of independent variables. Variables

Description

Caregiver education Caregiver participation in school

Education level of caregiver Do you attend parent-teacher activities, a school sports day or field day, and teaching activities? A 4-point response scale ranging from every time (coded as 1) to never (coded as 4) How would you describe your household income and expenditure? 1: Have quiet enough money, 2: Have enough money, 3: Balance between income and expenditure, 4: Have not enough money, 5: Have not quite enough money. 1: Less than 20,000, 2: 20,000–39,999, 3: 40,000–59,999, 4: 60,000–79,999, 5: 80,000–99,999, 6: More than 10,000 Single parent family (reference group), nuclear family, extended family, grandparent family, and other family The location where a household lives in. 1: City, 2: Urban township, 3: Rural township Number of dependent children less than 18 years of age Male (reference group), Female

Evaluation of income and expenditure

Family income

Family type Location Number of children Child's gender

this item; anyone unable to afford a television is likely to be severely deprived. The results of poverty measurement depend on the selected weighting procedures. Furthermore, different weighting functions produce distinct hierarchies (Szeles, 2004). Various weighting approaches, such as the majority necessities index (MNI) and proportional deprivation index (PDI), have been applied to measure poverty. The PDI is more attractive than the MNI and it is more widely used. Although both indices highlight the initial selection of items, the PDI is less sensitive to the initial selection of items. In contrast to the MNI, the PDI is not based on the arbitrary classification of necessary and nonnecessary consumption, which would decrease its sensitivity to individual preferences, and it accounts for the differences between demographic groups (Halleröd, 1994; Halleröd, Bradshaw, & Holmes, 1997). However, the PDI indicates that each item is weighted by the proportion of the population identifying the item as being essential. The inverse function allocates more importance to the low-frequency items. Using the logarithm of the inverse function can prevent overemphasizing the importance of items with very low frequencies (Filippone, Cheli, & D'Agostino, 2001). Moreover, the minimum value of the weight is 0, indicating that no one in the population experiences relative deprivation when each person is classified as the lowest dimension (Filippone et al., 2001; Van der Walt, 2004). However, the inverse function has certain shortcomings. For example, it would be unclear if one of the dimensions provided no contribution to poverty while no one in the population was poor (Filippone et al., 2001; Van der Walt, 2004). 3.4.2. Seemingly unrelated regression models The second aim of this study was to explore the determinants of the perceived necessities, deprivation, and social exclusion. After measuring the levels of perceived necessity, deprivation, and social exclusion, we determined that the level of perceived necessity was related to two types of social disadvantage (Section 4.3). Therefore, SUR models proposed by Zellner (1962) were suitable to formulate two or more regression equations simultaneously and errors in these equations were correlated. The SUR involved applying a generalized least squares (GLS) method to estimate parameters, rather than applying an ordinary least squares (OLS) method; the SUR estimation is more efficient than OLS estimation. 4. Results We used a fuzzy set approach to measure poverty, focusing on the “degree” of poverty; similarly, we applied a fuzzy set approach to

measure the levels of perceived necessity, deprivation, and social exclusion. The study involved conducting a decomposition analysis to measure the poverty index according to certain dimensions. Table 2 lists the results of multidimensional and unidimensional fuzzy poverty measures by dimension. An MPI was used to examine the overall levels of poverty related to the consensus on the necessities and two types of social disadvantage; a UPI was measured according to each of the dimensions. Both MPI and UPI were defined as weighted functions. The overall index and each dimension index can be interpreted as a percentage value. In terms of UPI, these dimension indices also enable us to focus on particular types of deprivation and exclusion. The level of perceived necessity refers to the weighted percentage of children who considered each dimension necessary. The level of deprivation indicates the weighted percentage of children who were deprived in each dimension. The level of social exclusion represents the weighted percentage of children who were excluded from each dimension. Both absolute and relative contributions were calculated. The absolute contribution of each dimension was measured as the difference between the overall poverty index and the poverty index obtained when the contribution of the dimension was zero for all individuals. The absolute contribution also indicated the ratio of the weight of dimension to the total weight multiplied by the UPI. The relative contribution of each dimension was measured as the ratio of the absolute contribution to the MPI, which is the sum of all absolute contributions (Diallo, 2012). Relative contributions provided crucial information for examining which dimension contributed more to reducing deprivation and social exclusion. Using the aforementioned weighted method prevented assigning equal importance to each item. The frequency-based weight, proposed by Cerioli and Zani (1990), is an inverse function that involves assigning more weight to small proportions of an item. In this study, each of the items was assigned a specific weight2, and the weight attached to each dimension equaled 0.125. 4.1. Level of perceived necessity According to Mack and Lansley (1985) and Saunders (2008), items can be defined as “necessities of life” if at least 50% of respondents identify them as essential. Of the 53 items, 50 items were regarded as necessities, implying that a broad range of necessities was covered, including items of a material and social nature. A total of 16 items were considered essential by over 90% of the respondents. Of the 50 necessities of life, half of the child-specific items were considered essential by more than 80% support. In the fuzzy approach, the level of perceived necessity referred to the percentage of respondents who thought of each of these items as necessary. In Table 2, the MPI for the level of perceived necessity indicated that approximately 70% of the respondents defined all the items as necessities of life. The UPI for each dimension is the weighted percentage of respondents who perceived each dimension to be necessary. A broad difference between dimensions indicated a high degree of consensus on the housing dimension (82.3%), but a low degree of consensus on the recreation dimension (41.9%). The three highest levels of perceived necessity were attributed to housing (82.3%), medical care (78.2%), and clothing (75.6%). We calculated the contribution of each dimension to overall index. A high UPI for each dimension represented a high relative contribution to the MPI. The dimensions of housing (14.8%), medical care (14%), and clothing (13.6%) provided the largest contributions, indicating that these dimensions contributed to an increase in the level of perceived necessity. Similarly, considerable variation was observed among the relative contributions, which ranged from 8% to 15%. 2 All items were classified into eight dimensions, and the weights of all items were normalized to unity in each dimension. Consequently, the contribution of each dimension was unaffected by the number of items in the dimension (Pi Alperin & Van Kerm, 2009, p. 3).

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

39

Table 2 Decomposition of poverty indices. Level of perceived necessities Dimensions UPI Diet Clothing Medical care Education Recreation Environment Economic and social relationships Housing MPI

0.7546 0.7559 0.7823 0.6425 0.4188 0.7428 0.6587 0.8233 0.6974

Absolute contributions

Relative contributions

0.0943 0.0945 0.0978 0.0803 0.0524 0.0928 0.0823 0.1029

0.1353 0.1355 0.1402 0.1152 0.0751 0.1331 0.1181 0.1476

Level of deprivation UPI 0.0474 0.0517 0.0313 0.0639 0.0658 0.0683 0.0622 0.0422 0.0541

Level of social exclusion

Absolute contributions

Relative contributions

0.0059 0.0065 0.0039 0.008 0.0082 0.0085 0.0078 0.0053

0.1095 0.1194 0.0722 0.1477 0.1521 0.1578 0.1437 0.0976

UPI 0.1557 0.1472 0.0798 0.2002 0.2089 0.2119 0.1526 0.0849 0.1551

Absolute contributions

Relative contributions

0.0195 0.0184 0.01 0.025 0.0261 0.0265 0.0191 0.0106

0.1255 0.1186 0.0643 0.1613 0.1683 0.1707 0.1229 0.0684

Note: The weight attached to each dimension is 0.125. Because of space limitations, the weights of all items are not listed in the study.

4.2. Levels of deprivation and social exclusion

4.3. Bivariate correlation analysis

Regarding child-related deprivation, the overall index of deprivation, which reflected the weighted percentage of children exposed to deprivation was 5.4%. In other words, around 5% of children experienced deprivation. The UPI provided a measure of relative deprivation for each dimension. These dimension indices could be used to describe each type of deprivation. The environment dimension (6.8%) exhibited the highest level of deprivation, followed by the recreation and education dimensions (6.6% and 6.4% respectively), demonstrating that children were severely deprived regarding the three dimensions. By contrast, the lowest level of deprivation was exhibited in medical care, with less than 4%, followed by the housing dimension (4.2%). Relative contributions provide crucial information for determining which dimensions contribute more to reducing deprivation and social exclusion. High levels of deprivation in the dimensions exhibited high relative contributions which ranged from 7% (medical care) to 16% (environment). The environment (15.8%), recreation (15.2%), and education (14.8%) dimensions play major roles in decreasing poverty, and medical care exerts a minimal effect on the alleviation of deprivation. The aforementioned results reflected children's deprivation relevant to the necessities that they were forced not to have. In this study, social exclusion referred not only to the “inability to afford the essentials,” but also to whether essentials were not wanted or not done. The level of social exclusion represents the weighted percentage of respondents who were excluded. Overall, the level of social exclusion exceeded 15%. This indicated that the level of exclusion varied considerably between certain dimensions. Over 20% of the children were socially excluded from recreation, environment, and education, whereas just 8% were excluded from medical care. The marked difference in the level of social exclusion between environment and medical care was over 13%. Moreover, the results of the three highest levels on dimensions (environment, recreation, and education) and the two lowest levels on dimensions (housing and medical care) were the same as those of deprivation, despite the levels of exclusion on the three highest levels on dimensions being high. Regarding the relative contributions, high levels of exclusion in the dimensions also made high relative contributions. The first three highest levels of exclusion dimensions (environment, recreation, and education) enabled decreasing social exclusion by approximately 50%. Clearly, children in Grades 5 and 6 experienced some degree of social disadvantage. Children faced particularly high risks of deprivation and exclusion in recreation, environment, and education. These dimensions were crucial in reducing deprivation and exclusion. Moreover, the level of perceived necessity was likely related to the levels of deprivation and exclusion, except for the environment dimension. The respondents prioritized necessities and exhibited a higher degree of consensus on necessities; therefore, they might have exhibited lower levels of deprivation and exclusion.

The second aim of the study was to explore the determinants of the level of perceived necessities and two forms of social disadvantage; thus, we employed three regression models. However, because the three dependent variables might be related, it was inappropriate to examine the causal relationships by using traditional regression analysis. The observations of three models were based on the same data set so that they were dependent each other. We investigated the relationships between three dependent variables by using a Pearson productmoment correlation. The initial step enabled choosing an appropriate statistical model to analyze the factors contributing to the level of perceived necessity and two types of social disadvantage. We identified a significantly negative relation between the level of perceived necessity and the level of deprivation (γ = −0.3937) and between the level of exclusion (γ = −0.6502) at the 5% level. It was expected that deprivation and exclusion would be strongly positively correlated (γ = 0.718) because of the overlap between the two variables. The three dependent variables appear to be closely related. Consequently, we applied SUR models to analyze possible factors regarding social disadvantage.

4.4. Determinants of perceived necessities, deprivation, and social exclusion Descriptive statistics of independent variables are provided in Table 3. The SUR enabled simultaneously applying three models having the same independent variables, rather than separately estimating them as a result of correlated error terms. As displayed in the last row of Table 4, the Breusch-Pagan test3 involved a chi-square statistic (λ), to test whether the residuals were contemporaneously correlated between equations. The results presented a rejection of the null hypothesis at a significance level of 1% (λ = 639.204), accounting for contemporaneous correlation among residuals across equations.4 Hence, the SUR models were appropriate for the study. In Model 1, the three factors of evaluation of income and expenditure, family income, and extended family were significantly linked to the level of perceived necessity. More specifically, evaluation of income and expenditure negatively affected the level of perceived necessity, meaning that those who said they had insufficient money exhibited a lower level of perceived necessity. Family income and extended family were positively related to the level of perceived necessity. This implied that high-income families were more able to buy essentials of life than low-income families were. Compared with single parent families, extended families exhibited a higher level of perceived necessity. 3 About the Breusch-Pagan test, see Breusch and Pagan (1980) for more detailed discussions. 4 For the same observations, the correlations of the residuals in the perceived necessity and deprivation equations (−0.2978), in the perceived necessity and social exclusion equations (−0.5846), and in the deprivation and social exclusion equations (0.6654) allowed the null hypothesis to be rejected.

40

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

Table 3 Summarized statistics of independent variables. Variables

N

%

Caregiver education [mean (SD)] Caregiver participation in school Every time Often Rarely Never Evaluation of income and expenditure Have quiet enough money Have enough money Balance between income and expenditure Have not enough money Have not quite enough money Family income Less than 20,000 20,000–39,999 40,000–59,999 60,000–79,999 80,000–99,999 More than 100,000 Family type Nuclear family Extended family Other family Single parent family (reference group) Grandparent family Location City Urban township Rural township Number of children [mean (SD)] Child's gender Male (reference group) Female

772 773 130 198 357 88 767 107 188 230 192 50 761 112 254 176 108 45 66 778 303 278 113 54 30 767 167 205 395 778 778 378 400

11.91 (3.45) 16.82 25.61 46.18 11.38 13.95 24.51 29.99 25.03 6.52 14.72 33.38 23.13 14.19 5.91 8.67 0.39 0.36 0.15 0.07 0.04 21.77 26.73 51.5 2.07 (0.75) 48.59 51.41

Several factors in Model 2 were significantly associated with the level of deprivation: evaluation of income and expenditure, family income, family type, and number of children. In particular, the signs of statistically significant estimates such as evaluation of income and expenditure, family income, and extended family differed from those of Model 1. Those who exhibited a more negative evaluation of their family's economic situation exhibited a lower level of perceived necessity but a higher level of deprivation. Children living in low-income families exhibited a lower level of perceived necessity but a higher level of deprivation. Compared with nuclear, extended, and other family types, children living in single parent families exhibited a high level of deprivation. A high number of children were related to a high level of deprivation. The determinants of social exclusion differed from those of deprivation. Model 3 revealed that children were socially excluded because of self-evaluation, family characteristics, and characteristics of caregivers. The determinants of social exclusion related to self-evaluation and family characteristics were similar to those of deprivation. Evaluation of income and expenditure, family income, and a nuclear as well as an extended family considerably affected children's social exclusion. In addition, the characteristics of caregivers affected children's social exclusion. The educational level of caregivers exerted a negative effect on children's exclusion, implying that the educational level of caregivers was related to lower levels of social exclusion. In other words, children were socially excluded as a result of their caregivers' low level of human capital. 5. Conclusions and policy implications Compared with the conventional framework that defines poverty as a lack of income, a multidimensional approach focusing on nonmonetary indicators elucidates the complexities of children's experience of poverty. However, little attention has been devoted to the multidimensional perspective on child poverty in Taiwan. To identify

Table 4 Seemingly unrelated regression models for determinants of social disadvantage (N = 732). Variables

Model 1

Model 2

Model 3

Perceived necessities

Deprivation

Social exclusion

0.001 (0.002) −0.009 (0.006) −0.028⁎⁎⁎ (0.006) 0.013⁎⁎ (0.005)

−0.001 (0.001) 0.001 (0.004) 0.026⁎⁎⁎ (0.004) −0.008⁎ (0.003)

−0.004⁎ (0.002) 0.01 (0.005) 0.031⁎⁎⁎ (0.005) −0.013⁎⁎ (0.005)

−0.050⁎⁎ (0.016) −0.047⁎⁎ (0.016) −0.039 (0.023) −0.043⁎

Female (Ref. male)

0.035 (0.025) 0.051⁎ (0.024) 0.005 (0.037) 0.044 (0.027) 0.005 (0.007) −0.009 (0.008) 0.027⁎

Constant

(0.011) 0.675⁎⁎⁎

−0.049⁎ (0.020) −0.053⁎⁎ (0.020) −0.030 (0.030) −0.033 (0.022) −0.003 (0.006) 0.007 (0.006) −0.015 (0.009) 0.179⁎⁎⁎

R-squared Breusch-Pagan test (λ)

(0.049) 0.111 639.204⁎⁎⁎

Caregiver education Caregiver participation in school Evaluation of income and expenditure Family income Family type (Ref. single parent family) Nuclear family Extended family Grandparent family Other family Location Number of children

Note: Coefficients and standard errors (in parentheses). ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

(0.017) −0.002 (0.005) 0.013⁎⁎ (0.005) −0.009 (0.007) 0.046 (0.031) 0.171

(0.040) 0.197

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

and understand child poverty more holistically, we conducted a preliminary study based on a multidimensional poverty approach. The aim of this research was to examine the extent of consensus on the necessities of life, identify two types of disadvantage experienced by children and analyze possible determinants of perceived necessity, deprivation and social exclusion. The preliminary findings are summarized as follows. The study involved applying the fuzzy set theory to measure the level of perceived necessity and the levels of deprivation and social exclusion. Our results indicated that over two-thirds of the respondents identified all the proposed items as essential. More specifically, the three highest levels of perceived necessity were housing, medical care, and clothing dimensions. Regarding the aggregate poverty indices, children in Grades 5 and 6 faced high risks of deprivation and exclusion, implying that they experienced some degree of social disadvantage. The three dimensions exhibiting the highest levels of deprivation and social exclusion were material: environment, recreation, and education; the two dimensions exhibiting the lowest levels of deprivation and social exclusion were medical care and housing. The dimensions exhibiting higher levels of deprivation and social exclusion have the potential to effectively facilitate the reduction of poverty. The second aim of this study was to identify possible determinants contributing to the levels of perceived necessity, deprivation, and social exclusion by using SUR models. Self-evaluation, economic factors, and family characteristics were demonstrated to substantially affect the perceived necessities and two types of social disadvantage. Those who indicated that they had insufficient money exhibited a high level of deprivation and exclusion. Low-income families were prone to deprivation. Children who lived in single parent families exhibited high levels of deprivation and exclusion, implying that single parent families had a low number of earners, compared with nuclear and extended families. A high number of dependent children exerted a negative effect on the deprivation of children; however, it had only a minor influence on social exclusion. Deprivation may have been associated with family resource dilution because of large family size. Finally, the characteristics of caregivers played crucial roles in children's social exclusion. Low investment in human capital leads to children being socially excluded. Therefore, a lack of resources and opportunities may discourage children's participation in social life. Several implications for methodology and policy can be derived from this study. First, multidimensional poverty engenders two perspectives providing various insights into child poverty. Multiple measures of poverty involve transcending the one-dimensional approach to poverty and enable obtaining a holistic complete picture of child poverty. Measures of poverty such as deprivation and social exclusion enable identifying various types of social disadvantage, reflecting the problems of vulnerable children. The fuzzy set theory was effective in the analysis of poverty. The approach employed in this study involved estimating the level of deprivation and social exclusion, preventing arbitrary classification based on majority decisions. The fuzzy measure of poverty facilitated the comprehensive use of information about each of the proposed necessities of life such that all relevant factors were considered in the analysis. Second, the results indicate the importance of adaptive preferences. As stated by Wright and Noble (2013), socially perceived necessities may have relevance to people's adaptive preferences. Self-defined economic status was strongly associated with the degree of perceived necessities; in other words, people who had a low economic status reported a low level of perceived necessities. Similarly, the level of perceived necessities decreased with family income, indicating that people's economic status may influence their socially perceived necessities. Furthermore, an obvious relationship was observed between economic constraint and deprivation. The results were in agreement with those reported by Halleröd (2006). In the study, people with fewer economic resources may have higher levels of deprivation (i.e., an absence of affordability) and this may be related to their choices (i.e., they have

41

no want or need, have inadequate access to opportunities, or lack the time to purchase essential items). Another crucial issue requiring clarification is related to the identification of social exclusion. Social exclusion is always deemed involuntary because it arises from factors beyond people's control. However, does this apply to voluntary decisions to withdraw from wider society? The difference between these two positions refers to the distinction between constraints and choices. In the study, social exclusion involved subjective choices and objective limitations, indicating that people were considered socially excluded regardless of whether their exclusion was voluntary or involuntary. Both forms of exclusion are harmful to wider society, even though some people who voluntarily exclude themselves are at the top of the social class hierarchy (Burchardt, Le Grand, & Piachaud, 1999). Empirically separating voluntary from involuntary nonparticipation might be difficult because of problems in ascertaining whether cases of self-exclusion are actually voluntary (Barry, 2002; Burchardt et al., 1999). Situations of nonparticipation relating to individual preferences are not easily distinguished from genuine cases (Saunders, 2011). Moreover, a person's current decisions and choices may rely on past decisions and choices (Burchardt, Le Grand, & Piachaud, 2002b). In other words, a person's current choices may be constrained to some extent by their past choices. Therefore, selfexclusion should be considered when operationalizing social exclusion (Barry, 2002). In addition, the foregoing analysis in the study suggested that a high degree of agreement on the necessities of life may be connected to lower levels of deprivation and social exclusion because people meet what are thought to be essential in advance, according to the results of the bivariate correlation analysis. In other words, the high levels of deprivation and exclusion do not represent a high level of perceived necessity. Thus, the perceptions of necessities should be considered when policy makers assess deprivation and social exclusion. This would be beneficial for the mitigation of child poverty by facilitating accurate identification of poor children and enabling the recognition of various dimensions of non-monetary poverty. Finally, regarding policy implications, emphasizing the multidimensional measurement of child poverty is essential. Income-based measures constitute a major approach to policy analysis in Taiwan, but do not represent an optimal method of identifying child poverty. Children affected by income poverty may not be deprived or socially excluded and children who experience deprivation or exclusion may not be in poverty. Poverty cannot be captured accurately by using only a single poverty approach. As Ravallion (1996) stated, both monetary and nonmonetary indicators should be considered. Policy makers should consider policy objectives before choosing an appropriate approach. The multidimensional approach to poverty measurement is superior in assessing and reducing child poverty. To address poverty related problems of Taiwanese children effectively, connecting research to social policy is crucial. Our findings also have implications for policy targets. Deprivation and social exclusion are concerned with the experiences of individuals situated in specific social contexts (cf. Nolan & Whelan, 1996, 2010; Townsend, 1979), implying that the two types of social disadvantage among children should be examined on the basis of the conditions in which disadvantaged children live. Moreover, the levels and compositions of deprivation and social exclusion would vary among different social contexts. For example, the list of dimensions for deprivation might not be invariably fixed across countries because of analytic purposes in practice (Alkire, 2007). Thus, multiple policy targets are necessary in ameliorating social disadvantage and may reach vulnerable children who are deprived in various dimensions. Resources must be targeted at children and their families according to these deprived dimensions as well. Children's basic rights are protected under the United Nations Convention on the Rights of the Child (CRC). Each child has the right to an adequate standard of living, development, education, and health.

42

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

Deprivation means that children are denied their rights (Bradshaw, Hoelscher, & Richardson, 2007). Furthermore, the CRC enables children to express their views and be listened to. The child-centered approach reflects respect for the views of children, and enables researchers and policy makers to identify socially disadvantaged children. The findings of this research are limited because this was a preliminary and small-scale study; therefore, further research is warranted. Furthermore, because few empirical studies have focused on the novel multidimensional approach to poverty, drawing comparisons with previous studies in Taiwan was problematic. Nevertheless, this paper demonstrated the new approach, elucidating multidimensional poverty in children. Our findings can facilitate the interaction between researchers and policy makers, and may encourage policy innovations in monitoring progress in the reduction of child poverty. The development of

indicators of child poverty largely depends on data collection from children. Future research may involve comparative studies enabling researchers to analyze variations between cities and counties, and identify the most and least deprived and excluded areas in Taiwan. In addition, comparing various types of social disadvantage among children and young people would be beneficial. More importantly, the fuzzy poverty approach can be applied to explore these topics. Acknowledgments This work was supported by the Ministry of Science and Technology in Taiwan (grant number: NSC 101-2410-H-029-058). The authors would like to thank the anonymous reviewers for their helpful comments.

Appendix A Appendix Table A Results for the perceived necessity of each item, the percentage of respondents who had the item, and (for those who did not have the item) the reasons for not having the item. Dimensions

Diet

Clothing

Medical care

Education

Recreation

Environment

Essential items

1. Three meals for a balanced diet 2. At least two meals with fresh fruit and vegetables every day 3. Never going without food or having insufficient food because of a lack of money 4. Dining at a restaurant with family members at least twice a year 5. Dinner with canned foods, instant noodles, or cookies no more than once a week 1. Buying new clothes every 6 months 2. Having at least two pairs of shoes for going out (except for sandals and slippers) 3. Buying clothes for children 4. Having sufficient clothes for various occasions 5. Having sufficient clothes for each season 6. Buying new shoes shortly after old shoes become worn out 1. All family members have valid National Health Insurance cards 2. I would take my child to see a doctor, even if I had no money 3. Seeing a doctor when sick and family members will take care of me 4. Regular dental exams at least twice a year 1. Buying new extracurricular reading materials, on average every 6 months 2. At least 30 children's fiction books 3. Having a personal desk and chair 4. Having a desk lamp or sufficient light 5. Having reference books (e.g., dictionary) 6. Participating in at least one extracurricular activity (e.g., talent classes such as painting, piano, or dancing and leisure activities) during the semester 7. Having a computer (e.g., tablet, desktop, or laptop) and access to the Internet at home 1. Having a 5 days' holiday away from home at least once a year 2. Being able to afford extracurricular activities 3. Going out with classmates once a week 4. Having a bicycle 5. Entertaining classmates at home 6. Participating in at least one extracurricular camp activity during winter and summer vacations 7. Playing with at least three classmates or friends 1. A convenience or grocery store within 1 km of my home 2. A clinic or hospital within 2 km of my home 3. Safe places for children to play near my home

Respondents who perceived the item as necessary (%)

Respondents who have the item (%)

Respondents who did not have the item (%)

Reasons for not having the item Unaffordable (%)

Accessibility or time constraints (%)

Unnecessary or unwanted (%)

98.5 90.9

94.68 85.47

5.32 14.53

60.98 37.5

21.95 26.79

9.76 25.89

7.32 9.82

77.3

82.68

17.32

22.56

24.81

41.35

11.28

70.5

72.43

27.57

27.83

32.55

32.08

7.55

74.7

72.85

27.15

16.35

14.9

54.33

14.42

64.6 76.7

69.86 77.63

30.14 22.37

24.35 27.91

11.3 8.14

58.7 55.23

5.65 8.72

91.4 79.4 95.6 89.9

89.3 78.6 93.88 87.79

10.7 21.4 6.12 12.21

28.05 31.52 55.32 47.87

12.2 10.3 6.38 14.89

41.46 49.7 23.4 22.34

18.29 8.48 14.89 14.89

95.5

96.29

3.71

57.14

7.14

32.14

3.57

84.4

91.15

8.85

34.33

14.93

43.28

7.46

98.1

96.36

3.64

46.43

21.43

14.29

17.86

71.5 56.0

66.84 54.12

33.16 45.88

22.13 25.07

35.57 22.22

30.04 43.02

12.25 9.69

48.3 88.7 86.4 79.8 80.6

50.92 84.93 83.64 79 82.2

49.08 15.07 16.36 21 17.8

25.33 33.04 34.4 28.13 22.06

23.73 16.52 16 20 28.68

39.73 34.78 35.2 40.63 33.82

11.2 15.65 14.4 11.25 15.44

80.5

88.76

11.24

40.7

18.6

31.4

9.3

20.4

22.66

77.34

26.92

48.55

20.27

4.26

85.2 38.8 81.2 65.8 54.5

86.8 43.62 86.05 71.67 56.08

13.2 56.38 13.95 28.33 43.92

67.68 14.86 16.98 22.58 17.26

8.08 44.34 33.96 35.48 37.2

14.14 31.6 33.96 29.95 34.82

10.1 9.2 15.09 11.98 10.71

91.5 87.5

91.79 90.97

8.21 9.03

28.57 24.64

23.81 13.04

25.4 40.58

22.22 21.74

73.9 78.1

69.35 79.63

30.65 20.37

30.51 40

29.24 16.77

12.29 20

27.97 23.23

Other (%)

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

43

Appendix Table A (continued) Dimensions

Economic and social relationships

Housing

Essential items

4. No air, noise, or other environmental pollution nearby 5. Being able to walk to school within 30 min 6. A bus station or train station within 1 km of my home 1. At least three classmates or friends 2. Family members attend weddings and funerals of relatives and friends 3. Regularly saving at least NT$3000 per month for the family 4. Having less than NT$5000 savings 5. Receiving a weekly allowance 6. Discussing schoolwork with classmates 7. Intimate talks with classmates or friends on LINE or Facebook 8. Family members celebrate festive days together 1. At least two bedrooms 2. Having a separate bed 3. Having a flushing toilet 4. No rats and cockroaches 5. Having sufficient water, electricity, and fuel 6. Being able to afford to repair or buy bathroom accessories 7. A clean environment with no unpleasant odors 8. Having a refrigerator, washing machine, and air conditioner 9. Owning furniture 10. Having a dining table large enough for all family members

Respondents who perceived the item as necessary (%)

Respondents who have the item (%)

Respondents who did not have the item (%)

Reasons for not having the item Unaffordable (%)

Accessibility or time constraints (%)

80.6

77.34

22.66

44.19

12.79

8.72

34.3

77.2 64.2

74.64 64.88

25.36 35.12

23.08 26.02

21.03 16.73

28.72 30.86

27.18 26.39

94.3 91.7

93.79 91.93

6.21 8.07

31.91 40.98

23.4 27.87

23.4 22.95

21.28 8.2

69.4

65.17

34.83

67.3

12.93

6.46

13.31

78.1 50.3 89.8 60.4

73.97 53.95 90.84 72.58

26.03 46.05 9.16 27.42

71.43 25.14 32.86 17.14

9.69 13.71 15.71 21.9

6.63 53.14 34.29 50

12.24 8 17.14 10.95

85.8 86.0 69.2 95.0 83.1 96.1 90.3

90.25 86.41 69.19 94.26 81.90 94.99 86.05

9.75 13.59 30.81 5.74 18.1 5.01 13.95

29.33 40.38 27.54 40.91 48.2 65.79 66.98

30.67 19.23 20.34 9.09 9.35 2.63 5.66

28 28.85 34.32 36.36 11.51 13.16 8.49

12 11.54 17.8 13.64 30.94 18.42 18.87

94.5 97.0

92.69 95.70

7.31 4.3

53.57 63.64

8.93 6.06

7.14 24.24

30.36 6.06

97.7 95.2

96.75 94.28

3.25 5.72

56 34.09

20 27.27

24 25

0 13.64

References Alkire, S. (2007). Choosing dimensions: The capability approach and multidimensional poverty. In N. Kakwani, & J. Silber (Eds.), The many dimensions of poverty (pp. 89–119). New York: Palgrave Macmillan. Alkire, S., & Foster, J. (2011). Counting and multidimensional poverty measurement. Journal of Public Economics, 95(7–8), 476–487. Atkinson, A. B. (1998). Social exclusion, poverty and unemployment. In A. B. Atkinson, & J. Hills (Eds.), Exclusion, employment and opportunity (pp. 1–24). London School of Economics: Centre for Analysis of Social Exclusion. Barry, B. (2002). Social exclusion, social isolation and the distribution of income. In J. Hills, J. Le Grand, & D. Piachaud (Eds.), Understanding social exclusion (pp. 13–29). Oxford; New York: Oxford University Press. Batana, Y. M. (2013). Multidimensional measurement of poverty among women in SubSaharan Africa. Social Indicators Research, 112(2), 337–362. Battiston, D., Cruces, G., Lopez-Calva, L. F., Lugo, M. A., & Santos, M. E. (2013). Income and beyond: Multidimensional poverty in six Latin American countries. Social Indicators Research, 112(2), 291–314. Belhadj, B. (2011). A new fuzzy unidimensional poverty index from an information theory perspective. Empirical Economics, 40(3), 687–704. Belhadj, B. (2013). New fuzzy indices for multidimensional poverty. Journal of Intelligent & Fuzzy Systems, 24(3), 587–591. Betti, G., & Verma, V. (2008). Fuzzy measures of the incidence of relative poverty and deprivation: A multi-dimensional perspective. Statistical Methods and Applications, 17(2), 225–250. Betti, G., Cheli, B., Lemmi, A., & Verma, V. (2008). The fuzzy set approach to multidimensional poverty: The case of Italy in the 1990s. In N. Kakwani, & J. Silber (Eds.), Quantitative approaches to multidimensional poverty measurement (pp. 30–48). Basingstoke [England]; New York: Palgrave Macmillan. Bossert, W., D'Ambrosio, C., & Peragine, V. (2007). Deprivation and social exclusion. Economica, 74(296), 777–803. Bradshaw, J. (2001). Methodologies to measure poverty: More than one is best! Paper presented at the the International Symposium Poverty: Concepts and methodologies, Mexico City, March 28–29. Bradshaw, J., & Finch, N. (2003). Overlaps in dimensions of poverty. Journal of Social Policy, 32(4), 513–525. Bradshaw, J., Hoelscher, P., & Richardson, D. (2007). An index of child well-being in the European Union. Social Indicators Research, 80(1), 133–177. Breusch, T. S., & Pagan, A. R. (1980). The Lagrange multiplier test and its applications to model specification in econometrics. Review of Economic Studie, 47(1), 239–253. Burchardt, T. (2000). Social exclusion: Concepts and evidences. In D. Gordon, & P. Townsend (Eds.), Breadline Europe: The measurement of poverty (pp. 385–406). Bristo: Policy Press.

Unnecessary or unwanted (%)

Other (%)

Burchardt, T., Le Grand, J., & Piachaud, D. (1999). Social exclusion in Britain 1991–1995. Social Policy & Administration, 33(3), 227–244. Burchardt, T., Le Grand, J., & Piachaud, D. (2002a). Degrees of exclusion: Developing a dynamic, multi-dimensional measure. In J. Hills, J. Le Grand, & D. Piachaud (Eds.), Understanding social exclusion (pp. 30–43). Oxford; New York: Oxford University Press. Burchardt, T., Le Grand, J., & Piachaud, D. (2002b). Introduction. In J. Hills, J. Le Grand, & D. Piachaud (Eds.), Understanding social exclusion (pp. 1–12). Oxford; New York: Oxford University Press. Cerioli, A., & Zani, S. (1990). A fuzzy approach to the measurement of poverty. In C. Dagum, & M. Zenga (Eds.), Income and wealth distribution, inequality and poverty (pp. 272–284). Berlin; New York: Springer-Verlag. Costa, M., & De Angelis, L. (2008). The multidimensional measurement of poverty: A fuzzy set approach. Statistica, 68(3/4), 303–319. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. Dagum, C., & Costa, M. (2004). Analysis and measurement of poverty. Univariate and multivariate approaches and their policy implications. A case study: Italy. In C. Dagum, & G. Ferrari (Eds.), Household behaviour, equivalence scales, welfare and poverty (pp. 221–271). Heidelberg; New York: Physica-Verlag. D'Ambrosio, C., Deutsch, J., & Silber, J. (2011). Multidimensional approaches to poverty measurement: An empirical analysis of poverty in Belgium, France, Germany, Italy and Spain, based on the European panel. Applied Economics, 43(8), 951–961. Dewilde, C. (2004). The multidimensional measurement of poverty in Belgium and Britain: A categorical approach. Social Indicators Research, 68(3), 331–369. Diallo, F. L. (2012). Analysing multidimensional poverty in Guinea: A fuzzy set approach. Nairobi: African Economic Research Consortium. Filippone, A., Cheli, B., & D'Agostino, A. (2001). Addressing the interpretation and the aggregation problems in totally fuzzy and relative poverty measures. SER working papers no. 2001–22. Colchester: University of Essex. Giordani, P., & Giorgi, G. M. (2010). A fuzzy logic approach to poverty analysis based on the Gini and Bonferroni inequality indices. Statistical Methods and Applications, 19(4), 587–607. Gordon, D., & Pantazis, C. (1997). Measuring poverty: Breadline Britain in the 1990s. In D. Gordon, & C. Pantazis (Eds.), Breadline Britain in the 1990s (pp. 5–48). Aldershot: Ashgate. Gordon, D., Adelman, L., Ashworth, K., Bradshaw, J., Levitas, R., Middleton, S., et al. (2000a). Poverty and social exclusion in Britain. York: Joseph Rowntree Foundation. Gordon, D., Pantazis, C., & Townsend, P. (2000b). Absolute and overall poverty: A European history and proposal for measurement. In D. Gordon, & P. Townsend (Eds.), Breadline Europe: The measurement of poverty (pp. 79–105). Bristo: Policy Press. Grødem, A. S. (2008). Household poverty and deprivation among children: How strong are the links. Childhood, 15(1), 107–125.

44

C.-H. Leu et al. / Children and Youth Services Review 66 (2016) 35–44

Halleröd, B. (1994). A new approach to the direct measurement of consensual poverty. Sydney: University of New South Wales, Social Policy Research Centre. Halleröd, B. (2006). Sour grapes: Relative deprivation, adaptive preferences and the measurement of poverty. Journal of Social Policy, 35(3), 371–390. Halleröd, B., Bradshaw, J., & Holmes, H. (1997). Adapting the consensual definition of poverty. In D. Gordon, & C. Pantazis (Eds.), Breadline Britain in the 1990s (pp. 213–234). Aldershot: Ashgate. Hillyard, P., Kelly, G., McLaughlin, E., Patsios, D., & Tomlinson, M. (2003). Bare necessities: Poverty and social exclusion in Northern Ireland-key findings. Belfast: Democratic Dialogue. Ho, H. C. (2007). Effects of governmental income transfers on poverty reduction for elderly households: A case study of elderly poverty from 1990 to 2000 (published in Chinese). NTU Social Work Review, 15, 89–120. Ho, H. C., Wang, T. M., & Leu, C. H. (2003). The effects of poverty measurements on the composition of poverty population: The estimation and the modeling of budget standards. Journal of Population Studies, 27, 67–104. Hsueh, C. T. (2008). Child poverty trends in Taiwan from 1991 to 2005 (published in Chinese). Taiwanese Journal of Sociology, 40, 89–130. Jana, S., Nad'a, B., & Jana, T. (2012). Material deprivation in selected EU countries according to EU-SILC income statistics. Journal of Competitiveness, 4(2), 145–160. Kerrins, L., Greene, S., & Murphy, S. (2011). Going without: Measuring child poverty and social exclusion in Ireland. Dublin: Combat Poverty Agency. Le Grand, J. (2004). Individual choice and social exclusion. In K. Dowding, R. E. Goodin, & C. Pateman (Eds.), Justice and democracy (pp. 170–178). Cambridge: Cambridge University Press. Lee, S. L., & Wang, T. M. (2008). Multi-dimensional trends of child poverty in Taiwan: Application and decomposition of the SST index (published in Chinese). NTU Social Work Review, 18, 191–230. Lee, Y. J. (2007). Estimating the social-excluded population in Taiwan (published in Chinese). Journal of Population Studies, 35, 75–112. Lee, Y. J. (2011). The dynamics of social disadvantage cumulation: A case study of Taiwan. Social Policy & Social Work, 15(1), 39–90. Leu, C. H. (1996). Gender differences in poverty. Journal of Social Sciences and Philosophy, 8(2), 221–256. Leu, C. H. (2010a). Profiles and explanations of child poverty in Taiwan (published in Chinese). Taiwanese Journal of Social Welfare, 9(1), 97–137. Leu, C. H. (2010b). The proximate determinants of poverty in Taiwan: Growth, redistribution, and the population effects (published in Chinese). NTU Social Work Review, 22, 109–152. Leu, C. H. (2010c). The targeting performance of governmental social benefits: The economically disadvantaged as the target (published in Chinese). Social Policy & Social Work, 14(2), 49–90. Levitas, R. (2006). The concept and measurement of social exclusion. In C. Pantazis, D. Gordon, & R. Levitas (Eds.), Poverty and social exclusion in Britain: The millennium survey. Bristol: The Policy Press. Levitas, R., Pantazis, C., Fahmy, E., Gordon, D., Lloyd, E., & Patsios, D. (2007). The multidimensional analysis of social exclusion. Bristol: Department of Sociology and School for Social Policy, Townsend Centre for the International Study of Poverty and Bristol Institute for Public Affairs, University of Bristol. Mack, J., & Lansley, S. (1985). Poor Britain. London: George, Allen & Unwin. Main, G., & Bradshaw, J. (2012). A child material deprivation index. Child Indicators Research, 5(3), 503–521. Menchini, L., & Redmond, G. (2009). Poverty and deprivation among children in Eastern Europe and Central Asia. International Journal of Social Welfare, 18(3), 225–236. Miceli, D. (1998). Measuring poverty using fuzzy sets. Canberra: National Centre for Social and Economic Modelling, University of Canberra. Mussard, S., & Pi Alperin, M. (2005). Multidimensional decomposition of poverty: A fuzzy set approach. Sherbrooke, Quebec, Canada: GREDI/University of Sherbrooke. Neff, D. (2013). Fuzzy set theoretic applications in poverty research. Policy and Society, 32(4), 319–331. Nolan, B., & Whelan, C. T. (1996). Resources, deprivation, and poverty. Oxford: Clarendon Press. Nolan, B., & Whelan, C. T. (2010). Using non-monetary deprivation indicators to analyze poverty and social exclusion: Lessons from Europe? Journal of Policy Analysis and Management, 29(2), 305–325.

Pi Alperin, M. (2008). A comparison of multidimensional deprivation characteristics between natives and immigrants in Luxembourg. Luxembourg: IRISS at CEPS/INSTEAD. Pi Alperin, M. N., & Van Kerm, P. (2009). mdepriv - Synthetic indicators of multiple deprivation. v2.0 (revised March 2014). Luxembourg: CEPS/INSTEAD. Ravallion, M. (1996). Issues in measuring and modelling poverty. The Economic Journal, 106(438), 1328–1343. Ringen, S. (1988). Direct and indirect measures of poverty. Journal of Social Policy, 17(2), 351–365. Roelen, K. (2014). Multidimensional child poverty in Vietnam from a longitudinal perspective-Improved lives or impoverished conditions? Child Indicators Research, 7(3), 487–516. Saunders, P. (2008). Measuring wellbeing using non-monetary indicators: Deprivation and social exclusion. Family Matters, 78, 8–17. Saunders, P. (2010). Income support for families and the living standards of children. In S. B. Kamerman, S. A. Phipps, & A. Ben-Aryeh (Eds.), From child welfare to child wellbeing: An international perspective on knowledge in the service of policy making (pp. 275–292). Dordrecht; New York: Springer. Saunders, P. (2011). Down and out: Poverty and exclusion in Australia. Bristol, U.K.: Policy Press. Saunders, P., & Naidoo, Y. (2009). Poverty, deprivation and consistent poverty. Economic Record, 85(271), 417–432. Saunders, P., Naidoo, Y., & Griffths, M. (2008). Towards new indicators of disadvantage: Deprivation and social exclusion in Australia. Australian Journal of Social Issues, 43(2), 175–194. Sen, A. (1985). Commodities and capabilities. Amsterdam: Elsevier Science. Sen, A. (1999). Development as freedom. Oxford: Oxford University Press. Silver, H., & Miller, S. M. (2003). Social exclusion: The European approach to social disadvantage. Indicators, 2(2), 1–17. Spicker, P., Leguizamon, S. A., & Gordon, D. (2007). Poverty: An international glossary (2nd ed.). London: Zed Books Ltd. Swords, L., Greene, S., Boyd, E., & Kerrins, L. (2011). All you need is…measuring children's perceptions and experiences of deprivation. Dublin: Children's Research Centre, Trinity College Dublin. Szeles, M. (2004). Multidimensional poverty comparisons within Europe - evidence from the European community household panel. IRISS working paper 2004–05. Differdange, Luxembourg: CEPS/INSTEAD. Townsend, P. (1979). Poverty in the United Kingdom: A survey of household resources and standards of living. Berkeley: University Of California Press. Townsend, P. (1987). Deprivation. Journal of Social Policy, 16(2), 125–146. Van der Walt, S. (2004). A multidimensional analysis of poverty in the Eastern Cape province, South Africa. Stellenbosch working paper series no. WP03/2004. Stellenbosch: Stellenbosch University. Waglé, U. R. (2005). Multidimensional poverty measurement with economic well-being, capability, and social inclusion: A case from Kathmandu, Nepal. Journal of Human Development, 6(3), 301–328. Waglé, U. R. (2008). Multidimensional poverty: An alternative measurement approach for the United States? Social Science Research, 37(2), 559–580. Wang, T. M., Ho, H. C., & Liu, Y. L. (2008). The effects of income growth, economic inequality and changing family structure on poverty rates: A case study from 1990 to 2004 (published in Chinese). Taiwanese Journal of Social Welfare, 7(1), 29–63. Whelan, C. T., & Maître, B. (2007). Income, deprivation and economic stress in the enlarged European Union. Social Indicators Research, 83(2), 309–329. Whelan, C. T., Nolan, B., & Maître, B. (2014). Multidimensional poverty measurement in Europe: An application of the adjusted headcount approach. Journal of European Social Policy, 24(2), 183–197. Wright, G., & Noble, M. (2013). Does widespread lack undermine the socially perceived necessities approach to defining poverty? Evidence from South Africa. Journal of Social Policy, 42(1), 147–165. Yu, J. (2013). Multidimensional poverty in China: Findings based on the CHNS. Social Indicators Research, 112(2), 315–336. Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8(3), 338–353. Zellner, A. (1962). An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias. Journal of the American Statistical Association, 57(298), 348–368.