Unbenanntes Dokument

1 What is social exclusion?

The rise and spread of the term social exclusion has been discussed and contested widely from different perspectives  [1] . In the existing research two complementary approaches are predominant: those focused on the lack of participation (processes) in inclusive social activities, and those stressing accumulated disadvantages (outputs).

According to the first approach, social exclusion is considered to be a situation where the individuals or groups involved do not participate in the sphere of socioeconomic interaction that constitutes inclusion, and therefore generates different forms of capability deprivation (Sen 2000). It is a ‘causal’ component of poverty (Lister, 2004), as it hinders on the ability of an individual to achieve the same results as his/her fellow citizens, when trying to reach the same goal with the same resources (Muffels & Tsakloglou, 2002).

In this vein social exclusion is considered to be a lack of participation in certain key social activities, a process by which "people are shut out, fully or partially, from any of the social, economic, political or cultural systems which determine the social integration of a person in society’ (Walker & Walker, 1997: 8). The lack of all or some social citizenship rights, as they are understood in developed economies, is therefore distinctive of social exclusion (Roche, 1997). Participation is thus regarded as a basic attribute that allows us to identify those in situations of exclusion, as opposed to those who are not (Barry, 2002:23).

Yet, exclusion can be also examined as a dynamic process of accumulated disadvantages. Burchardt et al. (2002:33) seek to identify levels of exclusion in Britain, by assessing with the British Household Panel Survey (BHPS) the extent to which exclusion is the result of a cumulative process of disconnections that overlap over time. Disconnections related to the sphere of ‘production’, which is the individuals’ position in the labour market; “consumption”, which is individuals’ capacity of acquiring services and products; ‘political engagement’, which is individuals’ involvement in processes of decision-making - principally voting, participation in unions, associations and public administration; and ‘social interaction’ or the extent to which individuals have stable networks of affective support. Likewise, Pantazis et al (2006), with the Poverty and Social Exclusion Survey (PSES) offer the measurement of social exclusion across four dimensions – namely, impoverishment, labour market exclusion, service exclusion and exclusion from social relations – and the relation between them. Whelan et al. (2001) suggest that individuals or households are at risk of social exclusion when they face important impediments in fulfilling three or more activities of everyday life.

Tsakloglou and Papadopoulos (2001) and Tsakloglou (2003), move in the direction of identifying exclusion as a series of social disadvantages (poverty, unemployment, illiteracy, housing, health and lack of sufficient incomes) as components of exclusion for four social groups: young adults, lone parents, sick or disabled, and retired. This approach is closer to a multidimensional concept of poverty, since it seeks to identify trends of cumulative disadvantages in key dimensions and how these correlate. Bransen et al. (2001) regard a person as vulnerable when he or she a) is not sufficiently capable of providing for their own necessities of life, such as shelter, food, etc. b) has several problems simultaneously (inadequate self-care, social isolation, lack of permanent or stable accommodation, mental health problems, etc.) c) does not, from the viewpoint of care professionals, receive the care and support they need to sustain themselves in society, and d) do not express care needs that readily fit into the mainstream care system, and therefore often experience unsolicited care or interference.

An approach that inspired us for designing and operationalising our variables, is suggested in Piachaud (2002). Drawing on the definition of exclusion as lack of participation described above, he argues that the likelihood of being socially excluded can be determined by the lack of four different types of capital, namely physical, financial, human and social capitals  [2] . Physical capital includes dwelling and properties; financial capital embraces income and financial assets; human capital refers to educational level and labour skills; and social capital represents the size and quality of family and social networks. This author argues that an individual with lower levels of capital is more likely to become excluded; and conversely, an individual with higher levels of capital is more likely to avoid exclusion. Piachaud, thus, links theoretically the exclusion from participation with the inequalities in ownership and access to different forms of capital.

2 Issues in social exclusion research

Researchers in the field of social exclusion face important limitations in terms of data availability as important target groups are generally excluded from socioeconomic databases (Subirats, et al., 2003). The homeless and other groups that are widely regarded as excluded – e.g. people living in precarious, unstable and/or institutionalized accommodation, substance abusers, migrants in irregular situation, nomadic populations, etc. – are left out of most socio-economic surveys that consider the general or “entire” population (Edgar & Meert, 2005; Atkinson et al, 2001). In spite of providing insights into vulnerable groups – generally below an income, or other dimension, thresholds  [3] , results from these studies fail to identify accurately those who are most excluded, their distinctive characteristics, and the socioeconomic processes underlying their situation. Some authors invoke the (numerically) residual importance of these groups as a disclaimer (Burchardt, et al, 2002)  [4] . Allegedly, the macro-sociological picture of the exclusionary processes within our societies might be unaffected by discarding these hidden groups from the analysis. Since reaching these groups is problematic and costly, it appears that social exclusion could be analyzed satisfactorily without empirical reference to those who suffer this phenomenon at its utmost intensity. These arguments that justify excluding hidden populations from analysis are problematic. Homelessness might be considered quantitatively a minor issue – certainly more widespread in a world scale outlook  [5] – but its individual and social effects are worthy of supplementary and more rigorous attention, especially within the social exclusion paradigm (Pleace, 1998).

A second problem is the tendency to split society between excluded and non-excluded, and the essentialization of both parties. When Room (1999) insists on the ‘catastrophic character’ of the ruptures leading to social exclusion, as distinct from poverty, he seems to denote two parallel scenarios, the one of inclusion and an almost irreversible one of exclusion. Yet if there seems to be a continuum instead of two parallel scenarios, as Richardson & Le Grand (2002) seem to suggest, such a thing as a catastrophic rupture becomes less evident. From our point of view, between the so-called ‘excluded’ and ‘non-excluded’, and amongst the excluded themselves, there are multiple and complex configurations of exclusion and inclusion that a dualistic approach simply cannot grasp. We should thus, consider whether it is appropriate to deal with social exclusion in dichotomous terms, or rather start considering degrees of exclusion, as research in fuzzy measurement of poverty suggests (Lemmi & Betti, 2006).

Yet our preference for processes rather than from static stances of exclusion is also challenging, as it is certainly difficult to apprehend how the dynamics exclusion take place. There seems to be some confusion between what constitutes a source of inclusion and what lead people to exclusion. The insertion of some dimensions is not always justified, or supported with adequate data. Lack of participation in processes of decision-making, elections, unions, political parties or NGO’s membership (Burchardt et al, 2002) can be a valuable sign of weak political involvement that reflects problems of political motivation, with the subsequent implications on political legitimacy. However, to consider this lack of participation as a dimension of exclusion in equal terms with production or consumption is to shift the focus away from socioeconomic welfare of individuals and groups, to wider considerations about social integration and cohesion. Unless non participation is due to legal restrictions or procedural impediments on voting or association, such as for immigrants in irregular situation, or citizens under an authoritarian regime, it is a priori unclear how abstention is not up to personal decision. Even when political participation were to be regarded as a key component of inclusion, we would miss at least two explanations: first, to consider whether there are significant impediments to the exercise of these rights caused by non participating in other dimensions of exclusion, mainly materials; and second, to find out, whether not exercising these rights has any implications on the participation on production, consumption and so on.

If we still agree with their emphasis on ‘social interaction’ as a dimension of exclusion, it is not clear whether household panel data has yet been designed to include adequate indicators of social support and social capital, controversial concepts per se in the sociological and economic literature. The indicators used in Burchardt et al (2002) capture the affective and community-based support dimension of social capital. However, this research misses the economic and logistic resources – irregular cash transfers, temporary accommodation, refereeing, employment opportunities, etc. provided by friends, relatives, or acquaintances – embedded in personal social networks (Lin, 2001) that could help the individual or group in avoiding exclusion or lifting out of it.

Finally we also have to consider the institutional context in which the dynamics of exclusion take place. Recent research has explicitly put welfare regimes and social exclusion in relation to one another (Bulpett, 2002; Ogg, 2005). The type of welfare regime is treated as an independent variable to explain variation in social exclusion amongst countries and the empirical connection between decommodification and de-familiarization with levels of social exclusion is identified. However the social exclusion concept is again taken for granted, and the task of building a theory on the relation between these two sociological constructs is not realized, leaving interesting results without theoretical guidance.

In order to do so, the constitutive elements of a theory of social exclusion should be embedded in the greater framework of welfare regime theory. In fact, Esping-Andersen (1999:36) advances a potential foundation for this development, when he warns against the hazards of assuming ‘functional equivalence’ between welfare agents. There is not such a thing as a perfectly coordinated frictionless interaction between the state, market, and households: when one of these institutions fails to provide resources, the compensation exerted by the others is never complete, so that the individual lays in a significantly more unprotected situation than before. We could sketch, tentatively, that these welfare-institutional dynamics systematically generate spaces of exclusion, i.e. mechanisms that accept or expel particular forms of risk and therefore particular types of person. But the extent and degree of this exclusion depends fundamentally on the structure of interaction existing between these welfare agents, that is, on the type of welfare regime.

From the standpoint of this investigation, social exclusion is not exclusively the alienation of individuals from the participation in the normal activities of our societies, but the “segregation” from the main institutionalised mechanisms of protection and safety-provision against risks. This process might subsequently lead to self-reinforced lack of participation that contributes to further deterioration and higher vulnerability. But these latter concerns refer more to the symptoms than the illness. If the main agents of welfare provision are not able to assist these individuals, or they seek to address their problems by means of generic policy, those excluded will likely resort to charities or third sector organizations with generally limited capabilities to create opportunities for inclusion. The latter, of course, would be a possible lawful alternative, but we do not exclude other forms of pooling resources, such as begging, informal economy or crime.

Addressing all these important issues at once is well beyond the scope and objective of this research. Our objectives are far less ambitious, and are concerned with providing some tentative insights into the following questions: (a) whether certain variables that are widely regarded as determinants of poverty and social exclusion perform differently in explaining variation within exclusion; and should these differences exists (b) whether our proposed understanding of social exclusion could provide them with reasonable theoretical support.

This research is, thus, oriented not to challenging existing research, but to contribute to it. Our aim is to add to the debate by suggesting that analyzing the exclusion of individuals normally outside the reach of socioeconomic databases is necessary to understand how the determinants of exclusion operate at its harshest level.

3 Data

Our research is based on a client registration database of the organization Arrels Fundació. This NGO, located in Barcelona’s downtown, has been helping excluded people for more than two decades, focusing on single homelessness as its major object of concern. Arrels Fundació manages a daytime open centre offering services, such as occupational workshops, basic health care, economic management, etc. to homeless clients, and also administers several flats and a larger accommodation centre to pursue ambitious projects of inclusion for targeted individuals. It has a long-standing commitment to participation in policy forums at both local and regional level, active membership in both Spanish and European policy networks, and has collaborated frequently with researchers working on the area of exclusion.

This dataset contains information given by the clients on the first time they arrive at the centre – before they start taking part in any program - and thus includes both usual clients and individuals who are “irregular” and use Arrels Fundació’s services sporadically. The information provided by these “irregular” users is rather biased and incomplete. We have not considered, thus, this group of sporadic users. Our sample includes those individuals who are frequent clients and are subject to the long-term inclusion strategies and programs implemented by the organization. The size of our sample, thus, is reduced down to 384 individuals

Due to its unconventional nature, the dataset suffers from some limitations as it was not conceived as a means for statistical analysis. The original design was unstructured and non-standardized, though we managed to correct it by means of recodification and categorization techniques. An additional issue to bear in mind is the scope for inference. The dataset is not a general survey of homeless individuals, but a collection of data from a specific organization in a particular city - Barcelona. Consequently, we do not seek to put forward general conclusions applicable to the whole Spanish or European population, or to different environments within Spain such as rural areas, where social exclusion might take different forms. Even if, as we discuss below, the profile of our observations seems to coincide with other existing research in Barcelona and elsewhere, we remain cautious about our findings. Despite these inconveniences, however, we consider that this dataset can still provide relevant insights into the problem of exclusion that could be overlooked dealing with conventional data.

The table below displays the frequencies in percentages of the sample for each variable  [6] . These results are contextualized within past and recent research of homelessness in Spain at local and national level.

Table 1: Frequencies for relevant variables

Variables

Percentage

Gender

 

Male

87.51

Female

12.49

Age

 

20-29

6.06

30-39

23.98

40-49

29.66

50-59

21.24

60-69

12.3

70-79

5.29

80-90

1.19

90-*

0.28

Income type

 

No income

36.48

Non-conditional

36.04

Conditional

21.98

Temporary Job

5.49

Educ. Level

 

Illiteracy

11.05

Primary

73..3

Secondary

13.44

Superior

2.21

Accomodation

 

Street

71.81

Under roof

28.19

Exclusion

 

Initial

55.29

Consolidated

44.71

Source: own elaboration with Arrels’ database.

Our data suggests that men are clearly overrepresented (87.5%) as opposed to women (12.49% of the sample). This has been already noticed by other studies. In Cabrera’s study (1999) male respondents represent 87% and female 13%. In the latest general survey of homeless people (INE, 2005), men make up to 82.7% of users of shelters and other services for homeless people.

Almost 75% of observations are within the age group from 30 to 59 years old, particularly the stratum 40-49. The mean age is 48 years old. The INE (Spanish National Institute of Statistics) survey (INE, 2005) identifies the mean age at 37.9 years old, in a more general survey, and Cabrera (1999) placed it around 42 years old. Hence our results are roughly in accordance with similar studies on the same topic.

The first category of income (“no income”), including those who earn no regular cash (they beg, receive money from friends), constitutes 36.5% of the sample. The second category represents those who are recipients of non-conditional benefits such as RMI (integration-related minimum income) or non-contributive pensions, and represent almost the same percentage than no-income category (36.04%). Those who receive pensions or unemployment benefits represent 21.98% while those individuals who are currently occupied in temporary jobs constitute a small group (5.49%). No individual in the sample holds permanent- fulltime jobs.

We have found differences with the INE (2005) report which suggests a fairly different picture: 19.9% of the sample has a regular wage and a 7.4% has some form or job-related income; those with no public provided income represent a 17.5%; those with no regular income (including those begging and those helped by families or friends) make a 55.2%

In terms of education almost ¾ of the sample has only primary educational skills. This, alongside illiterate individuals (11.05%) shows that around 85% of individuals are low-skilled. Only 13.44% had attended secondary school in the past, and the number of college graduates is significantly low (2.21%). Yet, the INE report suggests a completely different result, suggesting that up to a 64.8% of the sample has secondary education, and even a 13.2% has achieved higher education.

An overwhelming 92.37% of respondents live on their own (alone) – i.e. either single, divorced or widowed. Almost 72% of the respondents lived (slept) in the streets when asked, but there is a large group of individuals living in shelters, hostels, or shared flats in charge of organizations (28%). The INE (2005) report states that 45.6% is sleeping outside the margins of the existing safety net.

Summing up, being a man, with low educational and labour market attainments, having no regular income and problems of housing seem to constitute the profile of the average client of this organization. This matches quite accurately the profile of homelessness in Barcelona presented by Tejero & Torrabadella (2005:43) created with data from the city council. However, as we have seen in the light of the latest general homeless survey, our dataset is slightly biased to comprise those individuals in advanced forms of exclusion  [7] . The INE survey on homelessness was designed to collect, during a month in 2005, data from people that used services aimed at homeless people. This entails that the personal situations captured by their sample ranged from people who slept over one night in a shelter after sudden eviction and found accommodation quickly afterwards, to people who used services on a frequent basis from organizations like Arrels, or even people who were not homeless in a strict sense but used services anyway, such as free catering. In our study, initial exclusion is not meant to represent a person who has happened to sleep in a public shelter one night – although he or she would qualify as a homeless under the broader FEANTSA ETHOS typology  [8] and appear in the INE database – but a person who is suffering of habitual problems of accommodation yet not living chronically on the streets.

4 Logistic regression on the probability of consolidated exclusion

The dependent variable of the logistic analysis is a dichotomic variable labelled as “exclusion phase” that describes the overall state of the individual, being either in an “initial exclusion phase” or a “consolidated or chronic exclusion phase”. These two states of exclusion are adjudicated by social workers when the client first comes to the centre. Arrels Fundació has drawn the following working definition as a rule-of-thumb for standardising social workers’ tasks:

Initial phase of exclusion:

1- individuals living in unstable accommodation for less than 3 years;

2- still linked to, and supported, by relatives and friends;

3- working in sporadic jobs; and no substantial loss of working habits, motivation for inclusion, or self-care.

Consolidated phase of exclusion:

1- individuals living for at least 3 years in unstable accommodation;

2- very weak links, or none at all, with family or friends;

3- almost permanently unoccupied; and substantial or total loss of working habits, self-care or motivation for inclusion.

These categories are important operational tools for social workers, as they summarize the existing level of deterioration for preparing different courses of action and aid strategies accordingly. We are aware of the problems of accepting, as given, the working indicators of exclusion employed by this organization. Evidently, there is a fine line between introducing knowledgeable and nuanced judgement, and introducing biases in the generation of the data. Yet the existence of subjectivity does not entail lack of rigour or impartiality on behalf of the professionals working in that organization. If anything, this “bias” could provide details that extend well beyond the grasp of a structured questionnaire. Another reason in favour is that retrieving data from hidden populations is problematic not only in terms of access, but also in terms of the capacity of the interviewees to hand the information reliably, especially when sensorial and mental disorders are present. This adds to our rationale of accepting their categorization and to think of them as adequate qualitative indicators of a person's stage of exclusion, as they arise from an informed and thorough examination of the clients’ condition, that would otherwise be unattainable.

Our methodology is coherent with the theoretical framework proposed, insofar as it places the analysis in the universe of those people already in situations of exclusion, i.e.. primarily, but not exclusively, homeless individuals. Instead of basing our research upon a survey which considers the entire population we examine a primary dataset (N=384) exclusively focused on excluded people.

Our dependent variable is the likelihood of being in consolidated phases of exclusion, as opposed to being in situations of initial exclusion. This dummy variable allow us to find out similarities and differences among those groups, which are expected to provide some insights into those characteristics and conditions that favor the deterioration of those in preliminary exclusion. This dependent variable has been modelled as a function of the following predictors or independent variables: age, gender, housing type, educational level and income type. Empirical research in the field has dealt extensively with the role of these variables in explaining exclusion. These traits, amongst others, have been regarded as ‘risk factors’, that is, particular characteristics of individuals that increase their vulnerability in front of shocks or ‘trigger factors’ (FEANTSA, 2004) such as separation or social network breakdown, employment loss, eviction, etc.

There is evidence that ageing and retirement increases significantly the likelihood of being excluded (Tsakloglou, 2003; Ogg, 2005) and makes it difficult to improve individuals’ condition once in exclusion (Tsakloglou, 2003; Poggi, 2004). Our hypothesis is that time is a bad ally of exclusion, so that additional units of age from a certain age onwards (growing older) is likely to increase the risk of consolidation.

Some authors have pointed out that women tend to be more exposed to exclusion from the consumption/income dimension although are less likely to be excluded from the social relations dimension (Burchardt 2000; Gordon et al., 2000). As women’s risks have been traditionally less commodified – i.e. less dependent on the labour market – their vulnerability in front of ‘trigger factors’ linked to the labour market is lower than men’s. However, women are less likely to improve once excluded (Poggi, 2004). This might point out to the loss connections with sources of welfare being harsher on women than on men, since women find it more difficult to resort to the labour market.

Having a place to live seems to be a decisive variable for avoiding exclusion. However, as we have noted before, this has not been sufficiently considered in studies using panel data or general population samples. Homelessness has been thoroughly analyzed as a dependent variable  [9] , emphasizing demographic and socioeconomic individual characteristics or housing market conditions, but its conclusions are not connected to the wider context of social exclusion.

Education and training are regarded as key variables in ensuring participation in all social dimensions and in preventing exclusion (OECD, 1999; Poggi, 2004, Sparkes and Glennerster, 2002) by means of developing cognitive and non-cognitive skills. However, it is less clear whether these skills make an actual difference when individuals are excluded, and the access to the labour market is conditional on.

Although a priori income (whether wages and/or benefits) is a good predictor of exclusion (Burchardt et al, 2002), other authors consider that income and unemployment prove to be insignificant in explaining why individuals become homeless (Early, 2004). Indeed there is a vast literature on low wage dynamics suggesting that low-wage workers tend to be "trapped" in a vicious circle of employment in the low skilled sector, unemployment and periods out of the labour force that increase their risk of permanency at the bottom end of the occupational hierarchy (Bradley et al., 2001). Whether these people are more exposed to the risk of exclusion and poverty or not is not the goal of this research. We aspire to shed some light on why excluded people in initial phase of exclusion fall into consolidation. In this respects (the lack of) incomes may be a good predictor of the risk of moving from initial to consolidated phase of exclusion.

In the logistic regression model the effect of the proposed variables (gender, age, type of income and housing) on the risk of consolidated exclusion will be measured

Table2: Logistic Regression for the probability of being in initial (0) versus consolidated(1) exclusion level.

 

Odds R

S.E.

Gender

 

 

(Male ref)

 

 

Female

2.55*

1.1

Age

 

 

20-29

0.12 

0.27 

30 – 39

0.55

0.81

40 – 49

10.35

11.23

50 – 59

29.90**

32.59

60 – 69

47.16**

53.69

70 – 79

48.77**

57.25

80's – 89

8.1

11.6

90+

4.75

8.88

Educ. Level

 

 

(Illiteracy ref)

 

 

Primary

1.04

0.49

Secondary

0.41

0.23

Superior

1.12

1.12

Income Type

 

 

(no income ref)

 

 

Non-conditional

6.91***

3.05

conditional

7.33***

3.73

Temporary job

0.54

0.37

Accomodation

 

 

(Street ref)

 

 

Roof

4.20***

0.08

Source: own elaboration with Arrels’ database.

Significant: ***=P≤0.1 **=P≤0.05 *=P≤0.10

Specifications of the logistic model

Model Sensitivity

90.37%

Model Specificity

63.08%

Correctly Classified

80.17%

(cutoff point 0.5)

 

Pseudo-R2

0.32

Goodness of Fit test (Prob>chi2)

0.9982

 

 

N

384

The logistic model puts gender, age and housing as the most relevant variables in predicting the probability of consolidated exclusion With men as reference group, being female increases the risk of consolidation. This observation is not only suggested by the data, but also by the experience of Arrels’ workers who point out that although women are a minority within the homeless population, when they are in exclusion they tend to be in worse conditions than men, pushing them easily into consolidation. As noted above, the relatively higher decommodification experienced by women is covered not as much by the welfare state, but by the household - as abundant literature on the Spanish welfare state has stressed. A tentative conclusion would suggest that the social network breakdown – especially family-related – might be critical for homeless women, whose resources for inclusion are deeply embedded in their network of strong ties. The protection that traditionally secured a woman status in Spain was highly attached to the main breadwinner position, usually the husband. That generated a high degree of dependency from the household, that amongst other thing discouraged educational attainment and labour market participation for women. In the event of a household breakdown, the resources once available are no longer there, and an attempt to enter the labour market is hindered on by low skills and experience that could not be developed before.

Age is also strongly associated with consolidation. Being in the older age stratum, namely 50 to 59, 60 to 69 and 70 to 79, significantly increases the likelihood of being in consolidated exclusion. Two different issues could explain with these results: on one hand, elderly individuals are more difficult to include within the labour market than young individuals (notwithstanding the quality of employment). Their links to this sphere of protection is weaker, thus having less chances of inclusion. On the other hand, the elderly appear to be inherently more vulnerable, to have more complex needs, and to be more injured overall by the loss of their home than young people (Panell & Palmer, 2004).

Educational level appears statistically insignificant in explaining the risk of consolidation, which contradicts other results shown by the literature above. Consolidated exclusion seems to cut across individuals with different educational attainment, which suggests that education is not a good predictor of consolidated exclusion. In fact our analysis reveal no correlation or covariation between educational attainment and the independent variables included in the model. The effects of education on promoting cognitive and non-cognitive skills can be positive in preventing people from falling into exclusion. Yet, once in exclusion education does not seem to be fundamental in avoiding process of worsening-off. Human capital is allegedly crucial for establishing and maintaining links with the labour market, however it seems that for excluded individuals, who have lost most or all of these linkages, sustaining or regenerating them is not as much related to human capital as to logistic issues. The ‘No Home, No Job’ report (Singh, 2005), for instance, points out the problems of homeless individuals in the UK, perfectly willing to work, who due to lack of permanent accommodation face innumerable practical obstacles – from receiving personal correspondence, to opening a bank account – when searching jobs. Additionally, skills can be forgotten due to lack of activity, or damaged by disability, mental illness and addictions which are prominent amongst homeless individuals rendering educational attainment and work experience practically ineffective for inclusion in the labour market.,

Income type is statistically significant for the first two categories – getting a non-conditional income (non-contributive pensions, RMI) and a conditional income (contributive pensions, unemployment benefit). This should be interpreted as follows: regarding those not receiving an income, those receiving conditional and non-conditional income have a higher risk of consolidated exclusion. This result is challenging given its counter-intuitive nature. And something similar occurs with the variable “housing”. Those having any kind of housing, whether hostels, or shared accomodation, as oppose to those not having households (sleeping in temporary shelters or on the streets) show a higher risk of being in consolidated exclusion.

How can these results be interpreted? From our standpoint the fact that those who have relatively better material conditions (those having regular income and housing, regarding those not having either) exhibit higher risks of consolidation in exclusion, implies that private and public resources are mainly devoted to those in phases of consolidation. In other words, private and public resources are scarce and insufficient for those in situation of initial exclusion in Barcelona. To some extent these results accord with Early’s observation about the insufficient explanatory capacity of income and unemployment to predict homelessness.

5 Conclusions

As we have stressed, there are limitations in using general socioeconomic datasets to understand social exclusion. Insights on the nature of vulnerable individuals and groups do not appear to match with our observations about individuals separated from the mainstream welfare providers. Furthermore, those mechanisms affecting the lives of individuals in the “general population” seem to relate differently to the individual characteristics of excluded individuals. Panel data and general socioeconomic surveys fail to include these individual typologies, and hence, subsequent analyses could fall short in explain an important part of the problem. A “two-fold exclusion”, from mainstream mechanisms of protection and from the analytical framework of research on social exclusion, is an issue that needs to be addressed both with normative and social scientific concerns. It should thus be stressed the importance of collecting adequate data to analyze social exclusion, even if accessing people such as those in Arrels Fundació would arguably be more costly in terms of time and resources than those in the general public. Surveys would come up as an obvious choice, but anonymized client registration data from both public and private agencies is a possible option that should be considered since it could provide a constant source of data. This would require efforts to homogenise database structures, to establish strict ethical standards of data collection and handling, to make logistically and economically possible for organizations to create, maintain and share these databases, and for public agencies to be able to compile and release this information to researchers. Further research should discuss different approaches of obtaining data that allows analysts to come up with more accurate pictures of social exclusion. Both the understanding of the processes of exclusion and the characteristics of those excluded, and the possibility of informing policy against exclusion, justify this claim.

Some of our findings suggest counter-intuitive conclusions, which may have direct policy implications. Future policy designs should reinforce the preventive character of public intervention by providing additional material and non-material resources to those in initial phase of exclusion in Barcelona. Homeless households, for example, receive quicker assistance from authorities in an effort to sustain the family unit and most importantly to protect children. The single homeless, however, are less protected and additionally they do not have inner mechanisms of their own (as intra-household redistribution) to deal with their new situation. It cannot be expected that individuals who have gone recently through severe events such as eviction, economic failure, leaving prison, social network loss, come up with a solution without any support from the welfare providers. Far from being followed by straightforward recovery encouraged by need, deprivation is self-reinforcing (Paugam, 1995). The transition from stable accommodation and economic security to homelessness and exclusion is seldom frictionless. It is, thus, easier and less costly to address the problem in an early stage, when especially motivation and health are still relatively untouched, than when the individual has lost the last remaining bits of self-confidence and self-respect.

The service provision for homeless people in Barcelona is concentrated in Third Sector organizations. The current levels of funding and the organizational characteristics of these non-governmental organizations are in some cases inadequate to provide real opportunities for inclusion, especially for those individuals in consolidated stages. It can be argued that NGO’s have the ability to work with specificity and closeness to the problem they are addressing, in a way that neither the market nor the state can do (Goodin, 2003). In fact, the multiplicity of factors leading to homeless and exclusion, and their consequences in both physical and psychological conditions, requires both close individualized attention and integral intervention (Dean, 2003). Despite these advantages, in most cases, the lack of resources and, in some cases, of a long-term perspective makes intervention partially effective and inclusion, dramatically unlikely. If the welfare state is to delegate care for the excluded on NGO’s, it should make sure that these organizations are able to accomplish their task. Otherwise, the scope of these service providers is seriously restricted to mere patching and not to implement a fully inclusive program of intervention.

Finally, this study is focused exclusively on excluded groups of the population could be replicated in other cities for domestic and international comparisons. There is empirical evidence that stresses the fundamental role of the institutional context (welfare regime) in explaining social and economic cross-national differences. If social exclusion is directly related to the institutional framework in which individual interacts, it is reasonable to think that the spread and nature of social exclusion will vary from one regime to another. Therefore future research tracing connections between exclusion and institutional context are expected to be a very promising approach to fight effectively against exclusion. This study could act as a first step for such a purpose.



[1] See for instance: Levitas 2000, Dixon et al. 2005, Byrne 1999, Silver 1994, Littlewood & Herkommer 1999, Haan 1999 .

[2] In this paper he suggests the existence of a fifth component called “public infrastructure”. Nevertheless, he argues that this form is to be considered in terms of an analysis of nation-state exclusion, an issue that it is not especially well defined or explained in the article, and that is not relevant for our discussion.

[3] Definitions of poverty consider those at the bottom end of the income distribution classified according to different measures or poverty thresholds. Poverty is defined as a total net income below one half or 60% of the median. Those in extreme poverty are placed below 25% of the median.

[4] “The main limitation of this survey – common to all household surveys – is the omission of institutional and homeless populations. A high proportion of the non-private household population might be expected to be socially excluded. However they form a small proportion of the population as a whole’(Burchardt, Le Grand and Piachaud, 2002: )

[5] The UN estimated that there were 100 million homeless people worldwide, by referring to ´those who have no shelter at all, including those who sleep outside, or in public buildings, or in night shelters set up to provide homeless people with a bed’ (UN Centre for Human Settlements, 1996 quoted in Forrest, 1999)

[6] It has to be noted that in fact there are much more variables than those below, but some of them had very few respondents so they had to be discarded. Our aim is to keep working in collaboration with Arrels Fundació on the definition of their new database with a more theoretically-informed and a dynamic approach that might lead to further reports.

[7] INE’s survey was performed on a sample of services for homeless people in towns with more than 20,000 inhabitants in Spain. The range of services comprehended is greater than those offered by Arrels alone, hence the diversity of personal situations covered in this survey is broader and the results are more representative. We could argue that our sample includes only one sector, the worst-off, of the universe of homeless that the INE survey seeks to include.

[8] ETHOS typology includes: rooflessness (without a shelter of any kind, sleeping rough), houselessness (with a place to sleep but temporary in institutions or shelter), living in insecure housing (threatened with severe exclusion due to insecure tenancies, eviction, domestic violence), living in inadequate housing (in caravans on illegal campsites, in unfit housing, in extreme overcrowding). (See http://www.feantsa.org for more information).

[9] See Early 2003 2004, O’Flaherty 1996 2005, amongst others.

References

Atkinson, T., et al. 2001: Social indicators. The EU and Social Inclusion Oxford: Oxford University Press

Barry, B., ‘Social Exclusion, Social Isolation and the Distribution of Income’ in Hills, J., Le Grand, J. and Piachaud (eds.) 2002: Understanding Social Exclusion. Oxford: Oxford University Press.

Bulpett, C. 2002: ‘Regimes of Exclusion’. European Urban and Regional Studies, Vol. 9, No. 2, 137-149

Burchardt, T., Le Grand, J. and Piachaud, D. (1), ‘Introduction’ in Hills, J., Le Grand, J. and Piachaud, D.(eds.) 2002: Understanding Social Exclusion. Oxford: Oxford University Press.

Burchardt, T.; ‘Social exclusion: concepts and evidence’ in Gordon, D. and Townsend, P. (eds.) (2000), Breadline Europe: The Measurement of Poverty. Bristol: Policy Press.

Byrne, D. 1999: Social Exclusion. Buckingham: Open University Press.

Cabrera, P.J. 1998: Huéspedes del Aire: Sociología de las personas sin hogar en Madrid. Madrid: Publicaciones de la Universidad Pontificia de Comillas.

Dean, H. 2003: ‘Re-conceptualising Welfare-to-Work for People with Multiple Problems and Needs’ Journal of Social Policy, 32, 3, 441-459

De Haan, A. 1999: ‘Social Exclusion: Towards a Holistic Understanding of Deprivation’ Paper for the World Development Report 2001 Forum on ‘Inclusion, Justice and Poverty Reduction’.

Dixon, Carrier & Dogan 2005: ‘On investigating the underclass: Contending philosophical perspectives’. Social Policy & Society, 4:1, 21-30.

Early, D.W. 2004: ‘The determinants of homelessness and the targeting of housing assistance’ Journal of Urban Economics 55, 195-214

EARLY, D.W. (2005) ‘An empirical investigation of the determinants of street homelessness’ Journal of Housing Economics 14, 27-47

Edgar, B. & Meert, H. 2005: Fourth Review of Statistics on Homelessness in Europe. The ETHOS Definition of Homelessness. Brussels: FEANTSA

Esping-Andersen, G. 1999: Social Foundations of Postindustrial Economies. New York: Oxford University Press.

FEANTSA 2004: Prevention Strategies to Combat Homelessness. European Report. Brussels: FEANTSA

Goodin, R.E. 2003: ‘Democratic Accountability: the Distinctiveness of the Third Sector’ Archives européennes de sociologie, XLIV, 3, 359-396

Gordon, D. et al. 2000: Poverty and Social Exclusion in Britain, York: Joseph Rowntree Foundation.

Instituto Nacional de Estadística 2004: Encuesta sobre las personas sin hogar (centros). Madrid: INE.

Instituto Nacional de Estadística 2005: Encuesta sobre las personas sin hogar (personas). Madrid: INE

Lemni, A. & Betti, G. (eds.) 2006: Fuzzy Set Approach to Multidimensional Poverty Measurement, New York: Springer

Levitas, R. ‘What is social exclusion?’ in Gordon, D. and Townsend, P. (eds) 2000: Breadline Europe: The Measurement of Poverty. Bristol: Policy Press.

Lister, R. 2004: Poverty (Key Concepts). Bristol: The Polity Press

Littlewood, P. & Herkommer, S., ‘Identifying Social Exclusión: Some Problems of Meaning’, in Littlewood et al. 1999: Social Exclusion in Europe: Problems and Paradigms. Aldershot: Ashgate Publishing.

Muffels, R. & Tsakloglou, P. 2002: Introduction: Empirical Approaches to Analysing Social Exclusion in European Welfare States. In R. Muffels & P. Tsakloglou (Eds.), Social Exclusion in European Welfare States (pp. 1-20). Cheltenham: Edward Elgar.

O’Flaherty, B. 1996: Making Room: The Economics of Homelessness. Cambridge, Mass: Harvard University Press

O’Flaherty, B.; “Wrong person and wrong place: for homelessness, the conjunction is what matters”. Journal of Housing Economics 13(2004) 1-15

OECD 2000: Overcoming Exclusion through Adult Learning; Paris: OECD

Ogg, J. “Social exclusion and insecurity among older Europeans: the influence of welfare regimes”. Ageing And Society, 25,1 (2005).

Panell, J. & Palmer, G. 2004: Coming of age: opportunities for older homeless people under Supporting People. London: Homeless Link & UK Coalition on Older Homelessness (available online at www.olderhomelessness.org.uk/ documents/ coming_of_age.pdf)

Pantazis, C., Gordon, D. & Levitas, R. 2006: Poverty and Social Exclusion in Britain: The Millennium Survey. Bristol: The Policy Press

Paugam, S. 1995: ‘The Spiral of Precariousness: a multidimensional approach to the process of social disqualification in France’ in Room, G. (ed.) Beyond the Threshold: the measurement of analysis of social exclusion. Bristol: Policy Press.

Piachaud, D. 2002: ‘Capital and the Determinants of Social Exclusion’, CASEpaper 60, London: CASE-LSE.

Pleace, N. 1998: ‘Single Homelessness as Social Exclusion: The Unique and the Extreme’ Social Policy and Administration, vol. 32, No. 1 (March), 46-59

Poggi, A.; Social exclusion mobility in Spain, 1994-2000 ’, Working Paper wpdea0409, Department of Applied Economics, Universitat Autonoma of Barcelona

Poggi, A.; ’Does persistence of social exclusion exist in Spain?’ Working Paper wpdea0308, Department of Applied Economics, Universitat Autonoma of Barcelona.

Richardson, E. & Le Grand, J. 2002: 'Outsider and Insider Expertise: The Response of Residents of Deprived Neighbourhoods to an Academic Definition of Social Exclusion.' Social Policy and Administration 36, no. 5 (2002), pp. 496-515.

Roche, M & Van Berkel, R. 1997: European Citizenship and Social Exclusion, Aldershot: Ashgate

Room, G 1999: ‘Social exclusion, solidarity and the challenge of globalization’, International Journal of Social Welfare, 1999: 8: 166-174.

Sen, A. 2000, ‘Social Exclusion: Concept, Application, and Scrutiny’, Social Development Papers No. 1. Manila: Asian Development Bank.

Silver, H. 1994: ‘Social Exclusion and social solidarity: Three paradigms’, International Labour Review, vol. 133, n. 5 -6: 531-578.

Singh, P. 2005: No Home, No Job: moving on from transitional space. London: OSW. (available online at http://www.osw.org.uk/pdf/No_Home_No_Job.pdf )

Sparkes, J. & Glennester, H. (2002) ‘Preventing Social Exclusion: Education’s Contribution’ in Hills, J., Le Grand, J. and Piachaud, D. (eds.) Understanding Social Exclusion. Oxford: Oxford University Press.

Subirats et al. 2004: Pobreza y exclusión social. Un análisis de la realidad española. Barcelona: Fundación la Caixa.

Tejero, E. & Torrabadella, L. 2005: Vides al descobert. Els mons viscuts dels sense sostre. Barcelona: Editorial Mediterrania i Fundacio Jaume Bofill

Tsakloglou, P. & Papadopoulos, F. 2002: 'Poverty, material deprivation and multidimensional disadvantage during four life stages: evidence from the ECHP' in Barnes, M. et al; Poverty and Social Exclusion in Europe. Cheltenham: Edward Elgar

Tsakloglou, P. 2003: The risk of multi-dimensional disadvantage and social exclusion during four life stages in a dynamic perspective: Evidence from the ECHP", in Apospori, E. & Millar, J.; The Dynamics of Social Exclusion in Europe; Cheltenham: Edward Elgar

Walker, A. & Walker, C. 1997: Britain Divided: The Growth of Social Exclusion in the 1980's and 1990's, London: Child Poverty Action Group

Author´s Address:
Javier Ramos Diaz
Visiting Professor at Electronic Democracy Centre (e-DC)
Zentrum für Demokratie Aarau ZDA
Kuttigerstraße 21
CH – 5000 Aarau
Switzerland
Tel.: ++41 (0)62 836 94 44
Email: javier.ramos@upf.edu

Albert Varela
Doctoral Candidate
Department of Sociological Studies
University of Sheffield
Elmfield, Northumberland Rd
Sheffield
S10 2TU
United Kingdom
Email: a.varela@sheffield.ac.uk

urn:nbn:de:0009-11-27072