Caritas for the Goals of the Agenda 2030: A Study on the Services Provided in Campania

1. Introduction

In a world marked by constant change, characterized by increasingly complex socioeconomic challenges and growing disparities among populations, the need for transformative action has become ever more urgent.

In response to these pressing issues, the United Nations’ Agenda 2030 emerges as a guiding light towards a more sustainable and equitable future for all. This ambitious action plan, adopted in 2015 by all 193 UN member states, sets forth 17 Sustainable Development Goals (SDGs) that encompass a wide range of critical issues whose main objectives is the achievement of a sustainable and inclusive development [26].

The definition of sustainable and inclusive development is highly debated in literature. The classical definition of sustainable development was first given by the Bruntland Commission in 1987 which says that the sustainable development is

“the development that meets the needs of the present without compromising the ability of future generations to meet their own needs”.

Inclusive development, on the other hand, refers to the economic growth that ensures fair distribution across society and creates opportunities for all [21]. For an in-depth reading on “Sustainable Development”, see “The Sustainable Development Goals Report 2023 (United Nations)” [27]. In this rapidly evolving global landscape, humanitarian and social assistance organizations like Caritas play a fundamental role in translating the ideals and objectives of the Agenda 2030 into concrete and tangible actions at the community level. Founded on the principles of solidarity, compassion, and love for one’s neighbor, Caritas has been committed for decades to combating poverty, social injustice, and providing support and assistance to vulnerable individuals, regardless of their social, religious, or geographical conditions.

The 17 SDGs are categorized into three main areas: the "Social domain", focusing on poverty alleviation, reduction of social and economic inequalities, and improvement of access to education and healthcare; the "Economic domain", aimed at fostering innovative economic growth, generating employment opportunities, and sustain economic resilience; and the "Environmental domain", dedicated to biodiversity conservation, mitigation of climate change, and preservation of ecosystems.

In this context, Caritas primarily contributes to the Social domain by “promoting the testimony of charity, which is the concrete love for others. The dimension of charity permeates and enriches the life of communities” [5].

Through its global network, Caritas delivers a range of services from food support to healthcare, education, and vocational training, aimed at improving the lives of vulnerable populations. Beyond providing assistance, Caritas mobilizes resources and fosters a culture of solidarity and active participation in society through awareness-raising projects, advocacy, and training.

In this present work, we will focus on analyzing the important connection between the Agenda 2030 and Caritas, examining how the latter fits into the broader framework of global efforts to achieve the SDGs and how it contributes concretely to realizing a more just, fair, and sustainable world for all.

To explore in depth the impact and effectiveness of Caritas’ initiatives in implementing the Agenda 2030, with particular attention to the local context and social dynamics at play, we will employ Tandem Analysis tandem analysis [6,28], which considers a dimension-reduction technique, such as multiple factor analysis, and, then cluster analysis to identify similar groups of countries.

Through this deep dive, we hope to offer new perspectives and valuable insights that can inform and guide future efforts towards achieving the SDGs and improving the well-being and dignity of all people, everywhere in the world.

This paper is organized as follows: section 2 is a discussion of the theoretical framework; section 3 presents the data; section 4 introduces the data analysis methodology; section 5 is an analysis of the results; sections 6 and 7 provides the discussion and conclusion of the study.

2. Inequality and Poverty

In contemporary discourse, the intricate relationship between inequalities and poverty has garnered significant attention as researchers and policymakers seek to understand and measure these interconnected challenges [1].

In literature, inequalities have been studied from various perspectives. Smith [25] has defined them as “differences in the distribution of income and wealth, as well as economic and social opportunities, among different groups within a society”.

The linkages between inequalities and poverty are profound and multifaceted, shaping the lived experiences of individuals and communities.

At its core, inequalities serve as fundamental drivers of poverty, amplifying and perpetuating economic and social disparities. Socioeconomic inequalities, such as unequal access to education, employment, and healthcare opportunities, can significantly limit upward mobility and economic independence for disadvantaged individuals. This restricted access often correlates with higher levels of poverty, trapping individuals in cycles of deprivation and marginalization.

Conversely, poverty exacerbates and reinforces existing inequalities, creating barriers to social inclusion and equitable participation in society. Limited access to resources and opportunities due to poverty further entrenches disparities based on factors such as race, ethnicity, gender, and geographical location. As a result, those living in poverty face heightened vulnerabilities and diminished prospects for socioeconomic advancement.

The UN recognizes the importance of combating poverty by including this Goal in the Agenda 2030, specifically Goal 1, defining it as “Eradicating extreme poverty for all people everywhere by 2030” [26]. The Agenda 2030 is an international plan, whereas our study focuses on the Italian context, specifically in Campania. We will now proceed with a description of the phenomenon of poverty in Italy, followed by an analysis of how Caritas contributes to addressing this issue.

2.1. Poverty in Italy in the Context of the Agenda 2030

Focusing solely on Europe, it’s clear that we are still far from achieving Goal 1 of the Agenda 2030, which aims to reduce by 15 million the number of people at risk of poverty or social exclusion in Europe. Eight years after the adoption of the Sustainable Development Goals and three and a half years into the socio-health crisis caused by Covid, we have experienced significant setbacks. Regrettably, the pandemic, the energy crisis, and the war in Ukraine are having a highly negative impact on our progress toward these targets.

Currently in Europe, more than 95 million people, comprising 21.8% of the population, are living in conditions of poverty and/or social exclusion. This percentage remains relatively stable compared to 2021 when it was at 22% [9]. However, the impact of the Coronavirus is evident, as indicated by Figure 1, showing a reversal of the previously observed signs of improvement starting from 2020.

Through a comparison among European countries ([9]), we can notice that in Italy, people at risk of poverty and/or social exclusion account for 24.4% of the population, which is higher than the EU average [9].

As it is not the objective of this paper to define the risk of poverty or social exclusion,it suffices to know that individuals in this category include those who live in at least one of the following situations: in households at risk of poverty, defined as having an income below sixty percent of the national median income; in conditions of severe material and social deprivation; in households with low work intensity. For further details, please refer to the literature (they are, for example, defined in the annual report on poverty by ISTAT [14]. If we shift from a European context to a national context, the southern part of Italy, known as the Mezzogiorno, emerges as the most affected area by poverty according to ISTAT data [14].

3. A Look at Data

Caritas is making a relevant contribution to the cause by conducting research and promoting concrete actions. For a comprehensive exploration of this topic, refer to the Caritas Report published in 2023 [8]. In this context, we analyze a dataset concerning individuals who visited Caritas. The data are accessible on the Ospoweb platform, which is used by Listening Centers for entering data related to the beneficiaries of Caritas services.

The dataset consists of $n=1837$ observations representing individuals who sought support from Caritas. In particular, the information was collected by Caritas staff from the relevant provinces and subsequently aggregated to form the dataset used for analysis. Various types of information were collected for these individuals, amounting to a total of $p=37$ variables, as detailed in Table 1.

For each variable, the nature of the variable and its respective categories (in the case of qualitative variables) have been specified.

A problem encountered during data analysis concerns the presence of missing values. There are several ways to handle missing data [20], among these, one can choose, when feasible, to perform imputation using specific techniques. However, when the data do not allow for this, for instance due to an extremely high rate of missing values (in our case, exceeding 2/3 for certain variables), one may opt to remove that variable. This is the approach we have chosen.

Below, we report which variables have been removed:

RP; MIG ; NM ; SAF ; MD ; OIND ; EM ; UEM ; CI ; OS ; TN ; ACM ; COINV; LAV; SAN; CONP; ISTR; SSA; SE; SEAL.

After making the changes, the dataset consists of 17 variables. For the needs of the statistical techniques used which we will describe later, transformations were performed on the quantitative variables. Specifically, such variables - i.e. DRES; NC; CHD; FC were transformed into categorical variables, dividing them into non-empty and equidistant classes of values.

In the next paragraph, we will proceed with the description of the statistical techniques used.

4. Methodology: Tandem Clustering

For the multivariate analysis of our data, among the most widely techniques proposed in literature (such as multiple factor analysis for mixed data by Pagés [22], and non-linear principal component analysis by Gifi [11]) we consider Tandem Clustering [6,28], which can be viewed as a method of minimizing redundancies in the data.

Here, tandem clustering uses a dimensional reduction technique, specifically multiple correspondence analysis [2,11,12,19], to create new variables that are uncorrelated, and then applies cluster analysis to form classes using these new variables.

Starting from the factors extracted by multiple factor analysis (MCA), instead of using the original variables, the statistical units (the individuals who have sought assistance from Caritas, in our case) will be grouped by using a Hierarchical agglomerative clustering algorithm [24]. As a result, only the most important variables will lead to the identification of similar groups of individuals.

Indeed, the factors, being orthonormal, have the advantage of providing the same impact on the (dis)similarity index used to measure the distance between the groups of units. The results of this integrated analysis approach allow us to better specify the needs of individuals and find confirmation of the individuals’ profiles already identified in MCA.

It is common in the literature to use this type of technique for this kind of analysis. [7,16,23]

4.1. Multiple Correspondence Analysis

The Multiple Correspondence Analysis is utilized to explore the relationships among a group of categorical variables observed across a population of statistical individuals or units. By generating new variables (latent variables) and pinpointing an optimal lowe-dimensional space, MCA serves as statistical technique for assigning scores to units and to each category of variables. Analyzing survey data by using MCA can be made calculating a super-indicator matrix $\mathbf{X}=\left[\begin{array}{ccccc}{\mathbf{X}}_1|& \cdots & |{\mathbf{X}}_{\mathbf{k}}|& \cdots & |{\mathbf{X}}_{\mathbf{p}}\end{array}\right]$ of p ordered categorical variables observed on the same set of n individuals. Let D be the super-diagonal table of dimension J x J where the (k, k)th diagonal matrix contains the relative column marginal frequencies,

$$p_{.jk}=\frac{{\sum }_{i=1}^n{x}_{i{j}_k}}{n}$$

for the k-th variable. Observe that an indicator matrix implies coding the data in a complete disjunctive form [12,18]. For example, the matrix ${\mathbf{X}}_k$ consists of elements 0 and 1, where 1 represents that an individual/unit is classified into a category and a 0 indicates that it does not share that characteristic. Therefore, the total number of categories under consideration is

$$J=\sum _{k=1}^p{j}_k$$

where the generic variable k has ${\mathbf{j}}_k$ categories. There are many ways with which multiple correspondence analysis can be presented, one of those is to perform a (generalized) singular value decomposition of the super-indicator matrix

$$\frac{1}{p\sqrt{n}}X{D}^{-\frac{1}{2}}=U\mathsf{\Lambda }{V}^T$$

(1)

Where $\mathsf{\Lambda }$ is the diagonal matrix of the singular values, $\mathbf{U}$ and $\mathbf{V}$ are the right and left singular vector matrix, respectively, which allow the computation of the coordinates for units and variable categories. The coordinates of the categories allow to display also graphically the relationships existing among the variables. In particular, since each category is the center of gravity of the units (assisted by Caritas) that have chosen it, the proximity between two categories highlights those chosen by the same people or by very similar individuals: the proximity between two categories can therefore be interpreted in terms of association between them. Similarly, the proximity between two units allows to highlight the (dis)similarity among people.

4.2. Hierarchical Agglomerative Clustering

To identifying homogeneous groups of units, we perform a cluster analysis [13] on the unit coordinates obtained through MCA which contains a synthesis of all original variables information.

In the literature [13], the term cluster analysis indicates a set of statistical techniques used to group statistical units based on the similarity of their profile, described by a set of variables. The resulting units’ group should be characterized by a high degree of internal homogeneity and there should be a high degree of variability between the individuals’ groups. Not knowing a priori the suitable number of clusters to analyze, among the plethora of classification methods, we preferred to consider the hierarchical agglomerative clustering, whose logic can be summarized through the following steps.

In the initial stage, each individual forms a separate cluster. In the second step those two individuals, which have minimum distance, are merged. For the calculation of the distance, the WARD method has been used; it is based on the decomposition of the total deviance into deviance between groups and deviance within groups. At each step, then, those two groups that get the minimum within-group deviance are merged. The third step calculates the distance between the new cluster (group) and all the other units. Finally, steps two and three are repeated until a configuration is reached where there is only one group.

The clustering process can be graphically represented through a dendrogram, from which it is also possible to read the aggregation index appreciate how much a group is separated from the others. Of course, the aggregation index can be used in order to identify the suitable number of clusters: cutting the cluster tree after the fusion that correspond to low values of the aggregation index and before those corresponding to high values of the aggregation index [10].

5. Results

Let’s proceed with the presentation of the results. The quality of representation of each category can be assessed by examining the contributions of the categories in Table 2.

In general, the further categories are from the origin, the better the quality of the graphical representation. In Figure 2, the most characteristic categories are those farthest from the origin (which represents independence from the variables).

The categories that most characterize the survey participants are indicated by warm colors, while the less significant categories are indicated by cool colors. On the right side of the image along the horizontal axis, there are variables and categories characterizing the first dimension. From this graph, we can start forming an idea about the associations of the modalities. On the right side of the figure along the horizontal axis, there are those individuals who come from a foreign country (Georgia, Iran, Santo Domingo, etc.) and with immigration issues; they are mostly young (see Table 2). The second dimension, the vertical axis, is characterized by those individuals who mainly reside in the province of Caserta, are homeless and have turned to Caritas for economic issues (they received clothing and food parcels).

After conducting MCA, the research proceeds with the implementation of hierarchical clustering based on the coordinates of the individuals extracted by MCA. For selecting the optimal number of dimensions to consider, we chose to use the elbow method. Considering the Figure 3, it is evident that from the second dimension onwards, the eigenvalues stabilize, suggesting that we consider the number of dimensions preceding the flattening.

The implementation of hierarchical clustering based on the coordinates of the individuals extracted by the MCA produces the dendrogram as shown in Figure 3.

The dendrogram suggest us to consider three cluster. To confirm this choice, we used the silhouette index ([24]), which returned a value of 0.59, indicating that a structure in the data was found. Too see the characteristics of each cluster, refer to Table 3, Table 4 and Table 5. Below is a guide on how to interpret it. The first column of each table, titled CLA/MOD, represents the percentage of individuals with a specific category within a cluster relative to the total individuals who possess that specific category in the dataset. The second column, named MOD/CLA, expresses the percentage of individuals with a specific category relative to the total individuals in the cluster. The third column, indicated as Global, indicates the percentage of visitors with a specific category in the entire dataset. The fourth column displays the p-value, which represents the statistical significance of the category in each cluster. Finally, the last column provides the test value related to the considered categories.

5.1. Clusters Description

In this paragraph, we will describe the three clusters, highlighting the characteristics that represent them. For each cluster, one or more primary needs for which they sought help were defined.

The first cluster (Table 3) is mainly characterized by individuals with health problems (HP) and economic issues (POV). The services provided to them are material goods and services. It is evident that these individuals are strongly characterized by being Italians, residents of Marcianise, which could indicate a significant issue affecting this province.

Regarding the second cluster (Table 4), it shows that the defining variables and categories are those related to work. Indeed, the individuals in this cluster have been listened to by Caritas and have sought support concerning poverty and employment issues (EI). These individuals have reported living in houses with irregular contracts (Unregistred Rent, UR) and having temporary employment contracts (TC). We can say that this cluster refers to employment and poverty issues faced by individuals residing in the province of Caserta, mainly in Maddaloni, San Marco Evangelista, and other areas.

Finally, the third cluster (Table 5) reflects all those foreign individuals who have immigration issues. As we can see, in this cluster, most individuals are non-italian nationals (Santo Domingo, Georgia, Kyrgyzstan, etc.), characterized by having immigration-related primary and secondary needs. They do not have children (CHD) and are young (13-17 and 18-34). Finally, they are Irregularly Employed (IE). Given these characteristics, we can say that the individuals in this cluster are young immigrants looking to settle down by seeking employment.

6. Conclusion

The conclusions drawn from this study reflect a thorough analysis of the connection between the United Nations’ Agenda 2030 and the organization Caritas, highlighting the crucial role the latter plays in implementing the SDGs and contributing to realizing a more just, fair, and sustainable world for all. The Agenda 2030, with its 17 SDGs, provides a global framework for addressing complex socioeconomic challenges and promoting equitable and sustainable development worldwide.

Caritas, founded on the principles of solidarity, compassion, and love for one’s neighbor, has been committed for decades for combating poverty and social injustice, and providing support and assistance to vulnerable individuals, regardless of their social, religious, or geographical conditions. Through a wide range of programs and services, Caritas offers food assistance, shelter, medical care, education, and emotional support to those in need, thus contributing to the social domain of the Agenda 2030.

This study focuses on analyzing Caritas’ impact on society using Tandem Analysis, a technique aimed at identifying the primary needs of individuals seeking assistance from Caritas. The results show that Caritas plays an important role in meeting the needs of the most vulnerable people while contributing to the achievement of the SDGs in the local context.

The analysis conducted highlights the fundamental role of Caritas as a humanitarian and social assistance organization in translating the ideals and objectives of the Agenda 2030 into concrete and tangible actions at the community level. Through its ongoing commitment and dedication to solidarity and compassion, Caritas remains a beacon of hope and a positive agent of change in the fight against poverty and social injustice, thereby contributing to a fairer and more sustainable future for all. From a statistical perspective, we considered tandem analysis due to its mathematical properties and the abundance of useful, interpretative graphical displays it offers. However, recent advancements in tandem analysis alternatives warrant investigation. These alternatives integrate dimensionality reduction and classification techniques simultaneously rather than sequentially. For handling qualitative variables, several promising techniques merit consideration, among these MCA K-means [17], iterative Factorial Correspondence Biplot [15], Cluster Correspondence Analysis [28]. These methods, providing a powerful tool for simultaneous dimensionality reduction and classification, could enhance our analytical framework.