Socioeconomic and Emotional Dimensions of Retiree Welfare in Kosovo: The Basic Pension and Implications for Sustainable Social Policy

$P K=a \sum_{t=1}^T w_t C_t$                      (1)

where,

• Pk is the amount of the contributory pension;
• α is benefit coefficient (percentage of income replaced by pension);
• W_t is annual salary in year t;
• C_t are the annual contributions paid in period t;
• T is the total number of years of work.

This equation shows that the contributory pension depends on earnings during the career and the level of contributions paid into the system. In Kosovo, the contributory scheme is challenged by the low level of contributions and informality in the labor market, affecting the future benefits of retirees.

b) Basic Pension (Social/State Pension Pillar I) - The basic pension in Kosovo, financed by taxes rather than individual contributions, could create long-term fiscal challenges, as Feldstein argues, by increasing the burden on the economy and becoming unsustainable if the growth in the number of retirees exceeds contributions from workers, while the lack of capitalization of savings does not encourage individuals to create sustainable retirement funds.

• Theoretical Aspect of Basic Pension Pllar I

The basic pension is a social benefit guaranteed by the state to the elderly, regardless of their previous contributions. It is part of the Welfare State system and is related to the theory of: i) Social Distribution Theory, ii) Beveridgian Pension Model, and iii) Welfare State Theory.

i) Social Redistribution Theory, as presented by Wilensky [40], emphasizes the state's central role in ensuring a fair and equitable distribution of resources, particularly for vulnerable groups such as low-income pensioners. The goal is to reduce social inequalities, provide financial support to those in need, minimize poverty, and foster social inclusion. Wilensky [40] explored how welfare states, including pension systems, act as redistributive mechanisms. He argues that social policies, particularly public pensions, are essential for reducing income inequality and ensuring economic security for older populations. His work contributes to Social Redistribution Theory by demonstrating how state intervention redistributes resources across social groups, mitigating disparities caused by market forces. A study by Titova and King [41] shows that the link between well-being and friendship is partially explained by the satisfaction of basic psychological needs. Additionally, the findings by Muñiz-Fernández and Cabanillas-Jiménez [42] reveal that pensioner-led households with dependents exhibit a 10.2 percent reduction in total spending compared to pensioner-led households without dependents. Johnson et al. [43] highlight that economic stability and social policies play a crucial role in shaping pension adequacy and the financial well-being of retirees. They investigate the growing importance of social support networks and advocate for policies aimed at poverty reduction as central to welfare state models. Furthermore, Kazuhiro [44] examines the impact of social identity on income redistribution policies and economic development, emphasizing that the scale of redistribution is often limited in many developing countries and how national identity influences these processes.

The equation of Social Redistribution Theory is: a) Initial income before state intervention [40]:

$Y=W+O I$                       (2)

where,

• Y is total income before taxes and transfers,
• W is income from work (wages);
• OI is other income (e.g., from capital, property, business).

ii) The Beveridge Model of Pensions [45] proposes that basic pensions should be financed by taxes and guarantee a minimum standard of living for all citizens, irrespective of their contributions. Applied in Kosovo, this model provides state-funded basic pensions aimed at reducing poverty among the elderly. The model is based on universalism, ensuring that all citizens receive a basic pension, regardless of their work history, to maintain a living above the poverty line. Beveridge [45] introduced this system to eliminate poverty and offer financial security for the elderly through a tax-funded mechanism.

A mathematical formulation of the Beveridge model can be presented as:

$P=P B+a\left(Y-Y_{\min }\right)$                      (3)

where,

• P is the total pension received by an individual;
• PB is the basic pension guaranteed by the state;
• Α is the coefficient of supplementary benefit based on previous earnings (in some variants of the Beveridge model, this value can be zero, making the system completely universal);
• Y is the earnings during working life;

• Y_min is a minimum income threshold that affects the increased pension benefit (if a system with additional elements on top of the basic pension is applied).

The universal pension system guarantees support for all citizens, regardless of their contributions, being financed by state tax revenues and ensuring a minimum standard of living for the elderly, while horizontal equality ensures that benefits do not depend on individuals' previous income or contributions.

iii) Welfare State Theory-Welfare State Theory emphasizes the state's responsibility to provide comprehensive social support systems, including pensions, healthcare, and unemployment benefits, to ensure social stability and equality. Esping-Andersen [46] classified welfare states into liberal, conservative, and social democratic models, highlighting how welfare provisions vary across countries based on their political and economic structures. He argued that the welfare state plays a crucial role in reducing inequalities, particularly for retirees. Catherine and Patrick [47] explored the shifting paradigms of welfare states, focusing on innovations needed to address aging populations. Lee [48] emphasized the importance of social support policies in reducing poverty and improving living conditions for vulnerable groups.

A more detailed approach can be formulated as follows:

b) Effect of taxes and social contributions:

$T=T_y+T_s$                       (4)

where,

• T_y are income taxes;
• T_s are social security contributions.

c) Social benefits and transfers:

$R=P+B+S$                       (5)

where,

• P are the pensions received from the state;
• B are other social benefits such as unemployment benefits or social assistance;
• S are subsidies or additional assistance from the government.

d) Income after state intervention (disposable income):

$Y_d=Y+T+R$                      (6)

$Y_d=(W+O I)-\left(T_y+T_s\right)+(P+B+S)$                      (7)

This formula shows the impact of fiscal and social policies on the incomes of individuals, including pensioners and vulnerable groups.

A general mathematical formulation of this theory is presented as:

$U=f(Y d, T, R)$                      (8)

where,

• U is social welfare (Social Welfare);
• Yd is disposable income after taxes and transfers (Disposable Income);
• T represents taxes and social contributions (Taxes and Social Contributions);
• R represents social benefits and transfers (Redistributive Transfers).

The basic pension supports the well-being of pensioners (W_τ), which can be described as a function of various socio-economic variables:

$\mathrm{U} \tau=f\left(\left(\mathrm{PB}, \mathrm{H}_\tau, \mathrm{M}_\tau, B_{\text {other }}\right.\right.$, Reg $\left.u_{\text {law}}\right)$                    (9)

where,

• Uτ  is pensioners' Welfare (dependent variable);
• PB  is basic pension;
• H_τ  is health expenses;
• M_τ is marital status;
• B_other  is other sources of income;
• Regu_law  is state support for pensioners (pension laws);
• Other varibles…

Kosovo's pension system includes two main pillars: the Old Age Pension (Basic Contribution), a state-funded social pension for all citizens over 65, and the Contributory Pension, based on individual savings for those who have made contributions [20]. The basic pension aligns with the Social Assistance Theory of Pensions, which advocates for minimum pensions to prevent poverty among the elderly [49-51]. Similarly, in Vietnam, a supplementary pension program exists for the elderly without contributory pensions, but it is limited and available to only a few [52].

2.2 The development path of Kosovo's economy in the context of pensions

Between 2021 and 2023, Kosovo's economy saw both progress and challenges. In 2023, GDP per capita reached 25% of the EU-27 average, indicating improved living standards, although growth was slow [53]. Real GDP growth was 4.3%, following a high of 10.7% in 2022, with a slowdown linked to reduced exports and trade contributions [20]. Nonetheless, household consumption and public investment continued to grow, and Kosovo experienced substantial wage and employment increases. Unemployment dropped from 28.7% in 2021 to 12.6% in 2023, while wage growth reached 7.6%. However, inflation rose to 11.6% in 2023, driven by high prices [20].

On the fiscal front, Kosovo's domestic debt decreased by 13%, with the Kosovo Pension Savings Fund [54] holding 48% of the debt. External debt increased by 8%, and Kosovo received a ‘BB-’ credit rating from Fitch Ratings in April 2024, potentially aiding in accessing international bond markets [55]. While there have been positive economic developments, challenges remain in managing inflation, the trade balance, and employment. The country has laws supporting pensions, social assistance, and vulnerable groups like people with disabilities and war victims:

In line with Annex 1 of the Law on the Status and Rights of Martyrs, Invalids, Veterans, and Civilian Victims, the Government of Kosovo approved a 20% increase in pensions and social benefits for war veterans and other eligible groups. This policy aims to enhance the social and economic well-being of individuals who contributed to or were affected by the war. The decision officially came into effect in January 2025 and reflects the government’s ongoing efforts to strengthen social protection mechanisms. Table 1 summarizes the main legislative frameworks governing social welfare and the rights of individuals in Kosovo [56].

Table 1. Laws related to social protection and the status of individuals in Kosovo

Source: [56]

Table 2. Summary of social protection schemes in Kosovo

Source: Kosovo Agency of Statistic (2025) Statistical Yearbook of the Republic of Kosovo 2025 [57]

Table 2 summarizes social protection schemes in Kosovo, including transfer schemes (social assistance, energy subsidies, and support for children with disabilities), pensions (contributory, non-contributory, and war-related pensions), mother & child schemes (maternity allowance and child benefits), and social services at the municipal level. The system is fragmented, and benefits are often considered insufficient. Based on Table 2, the current pension system, in Kosovo has undergone changes over the years, covering different categories of beneficiaries. If we compare the monthly payments for each category of pensioners in Kosovo in December 2024, they only cover basic living needs [57].

Table 3. Pension category and amount received

Source: Kosovo Agency of Statistic (2025) Statistical Yearbook of the Republic of Kosovo 2025 [57]

This study analyzes the well-being and income of basic scheme pensioners in Kosovo, through a descriptive analysis of data collected from the field. 100 pensioners belonging to the Basic scheme were surveyed. The questionnaire addressed issues related to the financial, legal, socio-emotional and health situation and addresses the challenges faced by Basic scheme pensioners in Kosovo.

4.1 Descriptive analysis

Analysis of study variables through mean and standard deviation provides a more complete analysis of the Welfare of Basic category pensioners in Kosovo.

The survey results show significant variability among pensioners (Table 5). The mean age is 1.62 with a high standard deviation (1.524), indicating diverse age ranges and financial needs. Marital status is relatively stable (mean 1.98, SD 0.714), while savings are low (mean 1.22, SD 0.54548). Pension schemes are uniform (0 on scale 0-0). Most pensioners believe their pension is adequate (mean 1.72, SD 0.45356), but spending patterns vary (SD 1.01840). Other income sources are limited (mean 0.52, SD 0.73512). Hobbies (mean 1.53, SD 1.43970) and state support (mean 1.1042, SD 0.80529) show significant diversity. Health issues are also varied (mean 1.60, SD 1.04978), with the greatest concern being financial survival (mean 2.32, SD 1.40611).

Table 5. Descriptive statistics

Source: author

The summary reveals significant variability in pensioners' living conditions, with some facing financial hardships while others have more stability. This highlights the need for targeted social and economic policies to address the diverse needs of pensioners.

4.2 The demographic statistic

In the basic category, pension income has a significant impact on the well-being of retirees, but many retirees also emphasize the importance of family support and socio-demographic factors. The state of pension schemes has been criticized for unfairness and financial inadequacy, while retirees who have never worked face great financial difficulties.

The research includes 100 retirees, all over the age of 65. Most of them are in the 65-70 age group, while the lowest percentage belongs to the ages of 74 and 77. R² = 0.681 and R² quadratic = 0.835 show a strong model for analyzing the well-being of retirees, explaining a large part of the variation in this well-being and suggesting opportunities for improvements in policies that support retirees (Figure 3).

The marital status of the surveyed pensioners (Figure 4), shows that the majority, 81% of them, are married. Meanwhile, 10 are widowed, 8 are divorced and only 1 is single. Most of the pensioners are aged 65, 66 and 67 years.

Pensioner savings (Figure 5): The majority of surveyed pensioners (70%) have no savings at all (ages 66, 70, 75 and 80), while 20% have saved a little (ages 66, 67, 69, 72 and 74), and only 10% of pensioners have sufficient savings (ages 65 and 72).

3.png

Figure 3. Age of the pensioners and profession of pensioners

4.png

Figure 4. Marital status and retirement age

5.png

Figure 5. Pensioners' savings and their age

4.3 The statistical tests

To analyze pensioners' well-being, several statistical tests are employed: a) Principal Component Analysis (PCA) [67]: Identifies key factors influencing welfare or Well-being. b) Cronbach’s Alpha [68]: Assesses the internal consistency of variables to ensure reliability. c) Variance Inflation Factor (VIF) [69]: Evaluates multicollinearity among independent variables to prevent complications in analyses. d) Auto-Clustering (AIC, AIC Change): Determines the optimal number of groups to identify distinct pensioner profiles.

These tests help understand the factors affecting pensioners' well-being and ensure the reliability of analytical results.

a) The Factor Analysis Tests, to find the dependent variable -Pensioners' Well-being (W) which is represented by 4 variables (Pensioners' Well-being = Financial Well-being, Health Well-being, Social and Legal Well-being and Emotional Well-being or Welfare’s).

Table 6 shows that all the extraction values are above 0.7, indicating that PCA provides a good explanation for these variables. However, "Health Welfare" has the highest explanatory percentage (85.5%), meaning that this factor has a strong influence on the analysis model. "Emotional Welfare" has the lowest value (77.6%), but it is still within the acceptable range.

The creation of a Common Index (Aggregation of all Four Components) is calculated as Total: 0.816 + 0.855 + 0.835 + 0.776 = 3.282. By aggregating all the components, you can create a Pensioners' Welfare Index (PWBI). This index represents an overall measure of welfare based on all dimensions (financial, health, social, and emotional).

Taking "Health Welfare" as the main component has the highest communality value (0.855), meaning it is most strongly linked to the total variance in the data. However, pensioners' health is not the main factor influencing overall welfare, so you cannot use this variable alone as the primary representative.

Option 1 (the Common Index of 3.282) is the best. The value 3.282, which we obtained from the sum of the communalities (0.816 + 0.855 + 0.835 + 0.776), shows the percentage of the total variance explained by the four welfare components.

The value 3.282 is the sum of the communalities and shows that these four factors together explain a significant portion of the total variance in your dataset. There are 4 factors with initial values of 1.000 each (because in PCA, each variable initially has a variance of 1.000), so a value of 3.282 out of 4.000 means that around 82% (3.282 / 4.000 × 100) of the total variance is explained by these factors.

The PCA analysis showed that the four main components (Financial Welfare, Health Welfare, Social and Legal Welfare, and Emotional Welfare) together explain 82% of the total variance in pensioners' welfare. This suggests that these factors are significant determinants in measuring the welfare of this age group.

Table 6. Communalities

Table 7. Total variance explained

Table 7 illustrates how the variance is distributed across the principal components in the Principal Component Analysis (PCA) of pensioners' well-being data. The first component accounts for 55.1% of the total variance, making it the most significant. The second component explains 26.96%, bringing the cumulative percentage to 82.05%. The third component accounts for 11.23%, increasing the cumulative percentage to 93.28%, while the fourth component explains only 6.72%, reaching a total of 100%. Ultimately, the first two components are the most important, as they explain over 82% of the variance and can be used to simplify the analysis by retaining the most critical information. This indicates that dimensionality reduction can be effectively achieved by using only the first and second components.

6.png

Figure 6. Scree Plot, eigenvalue and component number

The Scree Plot (Figure 6), assists in determining the number of components to retain for our analysis. The Eigenvalue Plot is utilized to visualize the distribution of variance among components. Components with eigenvalues greater than 1 are considered significant and should be retained to explain the variance. Eigenvalues or components less than 1 are typically not retained, as they do not explain sufficient variance and are deemed insignificant. In our case, the first two components are the most important, but they correspond to the first component above the "elbow" point in the Scree Plot.

Table 8 shows the loadings of each factor in the two main components extracted from the Principal Component Analysis (PCA). Here is the analysis of each relevant factor for each component. Component 1 is the most important and explains 55.098% of the total variance, indicating that it is the primary component helping to explain many of the changes in the pensioners' well-being data. Component 2 explains 26.95% of the variance and is also an important component, although less significant than component 1. It has a negative relationship with financial and health well-being but a positive relationship with social and emotional well-being. This suggests a divided model where social and emotional factors have an opposite impact from financial and health factors. Both components encompass important aspects of pensioners' well-being, helping to explain the diversity of well-being experiences across different groups.

Table 8. Component matrix^a

Table 9. Rotated component matrix^a

Table 9, component 1 (Financial Welfare (loading: .882) and Health Welfare (loading: .918), focuses on financial and health welfare, having high loadings for these two aspects. Component 2 (Social Welfare (loading: .912) and Emotional Welfare (loading: .838) is related to social, legal and emotional well-being, having high loadings for these two factors.

The rotation has helped to clearly separate these two groups of factors. This helps in understanding the impact of these factors on the overall well-being of retirees. Both components encompass important aspects of pensioners' well-being, helping to explain the diversity of well-being experiences across different groups. Taking "Health Welfare" as the main component has the highest communality value (0.855), meaning it is most strongly linked to the total variance in the data. However, pensioners' health is not the main factor influencing overall welfare, so you cannot use this variable alone as the primary representative.

Table 10. Component transformation matrix

Table 10 shows the following relationships: Component 1: 0.741 with the first rotated component, -0.671 with the second rotated component. Component 2: -0.671 with the first rotated component, 0.741 with the second rotated component. This indicates a clear rotation, with each component contributing to one of the new rotated factors.

Factor Analysis test for all variables

After finding the pensioners' well-being index (dependent variable), the index value was entered into SPSS along with the independent variables to eliminate variables that are insignificant in determining the well-being of basic category pensioners in Kosovo. To classify the 10 independent variables of the study (each independent variable has 4 answers, a total of 40 variables), the Factor Analysis test is used (Table 11).

In constructing analytical models on retirees (Table 11), variables with high communalities (> 0.7) should be considered key factors as they best explain the data structure, while variables with moderate communalities (0.5-0.7) may be important, but their impact remains more limited, and some of them could be eliminated to improve the model. Meanwhile, the variable "Adequate Pension," with the lowest communality (0.555), may not be a crucial component and should be reviewed for inclusion, confirming that well-being, state support, and health are the most important factors, while hobbies and the desired pension level have a more limited impact on the overall structure of the model.

In this case (Figure 7), 10 nodes for 10 independent variables, indicates a good separation of information between components, and the elimination of sub-questions helps simplify the analysis, focusing more on the variables that have a significant impact on the overall data structure.

7.png

Figure 7. Eigenvalue and component number

Table 11. Communalities

Figures 8 and 9 indicate whether the data has structural issues, but it is not the final test for the reliability of the data for GLM. For a stronger evaluation, specific statistical tests are used. The TwoStep Cluster test recommends conducting additional reliability tests before using GLM.

Figure 10 shows a moderate relationship between the independent variables (Dimensions 6 to 10) and the dependent variable (Well-being of pensioners), where the points are labeled with case numbers and form a moderately downward-curved line. This suggests that changes in the independent variables may impact the well-being of pensioners, but this impact is not direct or very strong.

10.png

Figure 10. Wr-Welfare of basic pension

Discrimination measure (Figure 11) shows the measures of discrimination for different dimensions, where the average of different variables ranges from 0.007 to 1.055. Pensioners are different in terms of factors such as age, marital status, hobbies, health, other sources of income, savings, pension spending, and state support and concerns of pensioners. The active total is higher for category 1-5.

11.png

Figure 11. Discrimination measure

12.png

Figure 12. The importance of the determinants of pensioner welfare

Figure 12 shows that the variables legal support, marital status, how the pension is spent, other resources from the son, daughter or investments contribute significantly to the well-being of pensioners, but the variables pension rate, savings, the desire to have a pension of 500 euros or the health of pensioners are not overlooked in terms of importance.

13.png

Figure 13. Answer research questions through descriptive analyses from the questionnaire

Answering Research Questions-Providing answers to research questions through descriptive statistics for Basic Category Pensioners (Figure 13).

Figure 13 provides insights into pensioners' welfare and challenges in Kosovo. Pension income's impact is mixed, with 53% finding it beneficial, but 44% seeing no effect. Income vs. socio-demographics is debated, as 41% disagree that income is more important. Family support, marital status, and children are key factors for welfare (49%). Pension schemes are widely seen as flawed, with 58% deeming categorization unfair. Legal challenges include low payments (35%) and lack of pension indexation (20%). Non-working pensioners struggle, with 62% facing severe financial difficulties. The findings highlight systemic issues in pension adequacy and implementation.

4.4 The testing of study hypotheses

In Hypothesis H1, the relationship between each variable and the Welfare of pensioners is examined based on the Pearson correlation result. Below, the potential economic and psychological effects of each variable are commented on:

Testing first Hypothesis (H1): "The variables of marital status, savings, adequate pension, pensioners' spending, other sources of income, pensioners' hobbies, legal support from the state, pensioners' health, and pensioners' concerns are positively and statistically significantly related to the Welfare of pensioners in Kosovo."

• Correlation Analysis and Interpretation of Variables

This section presents the results of the Pearson correlation analysis between the well-being of pensioners and a series of socio-demographic, financial and legal factors, according to hypothesis H1. This hypothesis predicts that these factors are positively and statistically significantly associated with well-being.

Table 15 shows the correlations and the level of statistical significance (p-value) for each factor:

In the Pearson correlation analysis examining the impact of socio-demographic, financial, and legal factors on pensioners' well-being, varying degrees of association were identified across variables. Personal savings showed a positive and statistically significant correlation with well-being (r = 0.373, p = 0.008), indicating that pensioners with greater savings report higher levels of welfare. Similarly, additional income from sources such as children or investments had an even stronger positive relationship (r = 0.471, p = 0.001).

Marital status demonstrated a positive but not statistically significant correlation at the 0.05 level (r = 0.237, p = 0.098). A sufficient pension of €500 (r = 0.168, p = 0.244) and hobbies (r = 0.093, p = 0.546) did not show statistically significant effects on welfare. Legal support from the state showed a very weak and negative correlation (r = -0.023, p = 0.877), suggesting no meaningful association.

On the other hand, health status exhibited a negative and statistically significant correlation with well-being (r = -0.352, p = 0.012), indicating that worsening health is associated with lower welfare levels among pensioners. The variable personal concerns, as well as the €120 pension amount, did not show significant associations (both r = -0.085, p = 0.557).

In summary, these results partially support Hypothesis 1 (H1) by demonstrating that financial and health-related factors are significantly associated with pensioners’ well-being, while legal factors and some personal characteristics do not have a statistically significant impact.

This indicates that hypothesis H1 is partially supported: some socio-demographic, financial, and health factors affect the well-being of retirees, but the influence of legal factors and some personal characteristics is not significant in this case.

Table 15. Correlation Pearson

**. Correlation is significant at the 0.01 level (2-tailed); *. Correlation is significant at the 0.05 level (2-tailed).

These findings directly address Research Questions RQ1, RQ2, and RQ3, highlighting the stronger influence of financial security and health status in determining pensioners’ welfare.

RQ1: How do pension income and socio-demographic characteristics affect pensioners' Welfare?

Financial factors such as personal savings (r = 0.373, p = 0.008) and additional income from family or investments (r = 0.471, p = 0.001) have a positive and statistically significant impact on pensioners’ welfare. Health status shows a significant negative relationship (r = -0.352, p = 0.012), indicating that poorer health is associated with lower welfare. Other socio-demographic factors, such as marital status (r = 0.237, p = 0.098), hobbies (r = 0.093, p = 0.546), and legal support (r = -0.023, p = 0.877), show weak or non-significant associations.

RQ2: Are income or personal/socio-demographic characteristics more important for predicting pensioners' Welfare?

Financial factors are stronger predictors of welfare than most socio-demographic characteristics.

Additional income (r = 0.471) and savings (r = 0.373) show higher correlations with welfare than marital status (r = 0.237) or hobbies (r = 0.093). However, health status remains an important exception among socio-demographic factors, with a significant negative impact (r = -0.352, p = 0.012).

RQ3: Which socio-demographic characteristics have the greatest impact on pensioners' Welfare?

Among the socio-demographic variables analyzed, health status has the strongest and statistically significant effect on pensioners' welfare (r = -0.352, p = 0.012). Other variables, such as marital status and hobbies, show no significant influence.

RQ4: How do pensioners evaluate the implementation of pension schemes and their shortcomings?

Responses to research question 4, which concerns pensioners’ assessment of the implementation and shortcomings of pension schemes in Kosovo, were obtained through descriptive analyses from the questionnaire (Figure 13). The results show that the majority of pensioners (58%) perceive that the categorization in the pension scheme is unfair, reflecting concerns about the way benefits are distributed and assessed. A further 12% complain about the lack of sufficient financial resources for pensions, while 30% state that their complaints are not taken into account by the responsible institutions.

These data provide a critical insight into pensioners’ perceptions and directly address research question 4, underlining the need for improvements in the transparency and fairness of the implementation of pension schemes.

In the future, it is recommended to develop further studies with mixed methodologies that include qualitative data to better understand aspects that were not fully addressed in this study, especially for question RQ4.

• Testing the Second Hypothesis (H2) Generalized Linear Models (GLM)

The second Hypothesis (H2): “The variables of marital status, savings, adequate pension, pensioners' spending, other sources of income, pensioners' hobbies, legal support from the state, pensioners' health, and pensioners' concerns have a positive and statistically significant impact on the Well-being (Welfare) of pensioners in Kosovo”.

This Multinomial Logistic Regression model analyzes the Welfare of pensioners based on several factors, using the Cumulative Logit as a link function, which predicts the probabilities of pensioners falling into different Welfare categories (e.g., “low”, “medium”, “high”). This model is useful for understanding how factors such as pension, health, state support, etc., influence the division of pensioners into these groups.

The Omnibus Test shows a Chi-Square value of 288.762 with 26 degrees of freedom and a significance value of 0.000. Since the p-value is less than 0.05, it indicates that the fitted model, which includes various predictors like marital status, savings, pension schemes, and others, significantly improves the prediction of pensioners' Welfare compared to the thresholds-only model. This confirms the importance of the variables in explaining pensioners' Welfare.

• Interpretation of Study Results Test of Model Effect

The high Chi-Square value (288.76) and p < 0.001 (in Table 16) indicate that the model is statistically significant. This means that including variables such as pension, health, and other income helps predict the well-being of retirees in a statistically significant way.

This result directly supports research question RQ1, which examines how income and socio-demographic characteristics affect well-being.

The relationship between the various factors that affect the Welfare of retirees can be examined using the results of the GLM model tests (Table 17). Next, the significance of each variable was tested by entering them one by one into the model. The p-values indicate whether the variable in question has a significant and significant effect in the model, or vice versa, if it does not.

Wald Chi-Square (Type I): This assesses the significance of each variable according to the order of entry into the model, assuming that the previous variables are already included. Wald Chi-Square (Type III): This tests the impact of each variable while holding constant the effects of all other variables, providing a more reliable and independent estimation from the order of the variables.

Marital Status: Significance: Statistically significant in both models (p = 0.000). Married pensioners tend to report higher levels of welfare, likely due to emotional and financial support from their spouses, contributing positively to their quality of life.

Savings: Significance: Not statistically significant (Type I: p = 0.215; Type III: p = 0.830). Savings do not have a noticeable impact on pensioners' welfare, implying that even with savings, pensioners may still face financial insecurity. Other factors like pension amounts and state support are more influential.

Adequate Pension: Significance: Highly significant in Type III (p = 0.000), not significant in Type I. Adequate pension levels are crucial for pensioners' welfare, as insufficient pensions contribute to insecurity and dissatisfaction. Many pensioners feel their pensions are inadequate to meet their needs.

Table 16. Omnibus test^a

Table 17. Model threshold (Tests of model effects)

Table 18. Summary of the results of Table 19 confirming H2 and the answers to the research questions

Table 19. Parameter Estimation GLM

Pension Spending: Significance: Significant in both models (Type I: p = 0.001; Type III: p = 0.000). How pensioners manage their pension spending significantly impacts their well-being. Effective management of finances contributes to higher satisfaction, with more control over finances leading to better welfare.

Other Sources of Income: Significance: Significant in both models (Type I: p = 0.000; Type III: p = 0.000). Additional sources of income, such as family support or investments, play a crucial role in enhancing pensioners' welfare. Pensioners who have other income sources beyond their pension tend to experience a better quality of life.

Hobbies: Significance: Not significant in Type I (p = 0.278), but significant in Type III (p = 0.014). While hobbies were not significant in the initial model, they became more significant when controlling for other factors. Hobbies can contribute to mental well-being and life satisfaction in retirement.

State Support (Pension Laws): Significance: Not significant in Type I (p = 0.352), nearly significant in Type III (p = 0.060). Legal support (pension laws) does not significantly impact pensioners' welfare. The results suggest that existing pension laws might not be sufficient to provide substantial support, indicating a need for reforms in pension policies.

Health: Highly significant in both models (p = 0.000). Interpretation: Health is a critical factor influencing pensioners' well-being. Poor health significantly affects their quality of life, emphasizing the need for accessible healthcare services and support for pensioners with health issues.

Major Concerns of Pensioners: Significance: Significant in both models (p = 0.015). Interpretation: Pensioners' concerns, such as financial insecurity or uncertainty about pensions, directly affect their mental well-being. Reducing concerns and providing emotional and financial security are essential for improving their quality of life.

In conclusion, the findings (Table 18) emphasize that multiple factors especially marital status, adequate pension, financial management, additional income, hobbies, health, and pension-related concerns play a crucial role in shaping pensioners' welfare. Addressing these areas can improve quality of life, while the lack of impact from savings and legal support suggests a need for targeted policy reforms.

Dependent Variable: Pensioners' Welfare (W) Constant’s 12.453

Model: (Threshold), 1. Pensioners' Age; 2. Pensioner Gender (Male); 3. Pensioners' Gender (Female); 4. Marital Status; 5. Adequate Pension: 500 euros; 6. Pension Spending; 7. Other Income: son, daughter, investments; 8. Pensioners' Hobby; 9. Pensioners' savings; 10. Pension concern decreases by 120 euros; 11. Proper support from the state; 12. Pensioners' health. Set to system missing due to overflow. b. Set to zero because this parameter is redundant. c. Computed based on the Pearson chi-square.

The equation for this study can be expressed as a logistic regression model (Table 19), where the dependent variable (Y) represents the Welfare of pensioners, and the independent variables (X) are the analyzed economic, legal, and psychological factors.

Pensioners' Welfare (W) = β0 + β1 (Pensioners' Age) + β2 (Pensioners'Gender Male) + β3 (Pensioners' Female Gender) + β4 (Marital Status) + β5 (Adequate Pension) + β6 (Pension Spending) + β7 (Other Income) + β8 (Pensioners'Hobby) + β9 (Pensioners'Savings) + β10 (Pensioners' Concern) + β11 (State Legal Support) + β12 (Pensioners'Health) + ε                   (20)

W = 12.453 + (-8.867) + β2(9.626) + β3(12.687) + β4(-7.850) + β5(-4.853) + β6(10.433) + β7(-12.115) + β8(-3.402) + β9(-7.460) + β10(12.214) + β11(29.120) + β12(-10.965) = 31.01%               (21)

The equation represents well-being (W), including a constant value (12.453) and the impact of 10 independent variables, each with a coefficient indicating its effect on well-being. The parts of the equation that represent the variables (β2, β3, etc.) correspond to economic, social, legal, and emotional factors and are multiplied by the respective values of these variables. In this case, the calculated value for well-being is 31.01%, while the remaining 68.99% may be influenced by variables not included in the study.

• The analysis of the table for each variable and comments on economic, legal and psychological aspect’s:

a) Pensioners' Age: Coefficient B = -8.867, Std. Error = 4.197, Wald Chi-Square = 4.465, Sig. = 0.035, Exp(B) = 0.000. Economic Impact, an increase in pensioners' age has a negative economic impact because, as people age, they often face challenges such as deteriorating health, decreasing income, and rising healthcare expenses. Older pensioners may find it harder to cover their daily living costs. The psychological Impact, Older pensioners often experience feelings of insecurity and fear about the future, which can negatively affect their emotional Welfare and contribute to anxiety and depression.

b) Pensioners' Gender (Male): Coefficient B = 9.626, Std. Error = 4.715, Wald Chi-Square = 4.169, Sig. = 0.041, Exp(B) = 15156.946. The Economic Impact, Male pensioners may have had greater opportunities for employment and career advancement during their working years, which can lead to higher income and pensions. This financial stability supports a better quality of life. The Psychological Impact, Male pensioners who had better economic opportunities may feel more secure and satisfied, reducing levels of anxiety and stress.

c) Pensioners' Gender (Female): Coefficient B = 12.687, Std. Error = 4.766, Wald Chi-Square = 7.086, Sig. = 0.008, Exp(B) = 323542.025. The Economic Impact, Female pensioners often face lower pensions due to shorter work histories and wage disparities. However, they may have stronger support from family and other social groups, which can help mitigate economic challenges. The Psychological Impact, despite having lower economic opportunities, female pensioners may feel more connected to their families and social networks, leading to positive emotions such as love and support, which contribute to emotional Welfare.

d) Marital Status: Coefficient B = -7.850, Std. Error = 1.743, Wald Chi-Square = 20.279, Sig. = 0.000, Exp(B) = 0.000. The Economic Impact, married pensioners may face economic pressure due to having to share financial resources among family members. This can create additional financial burdens and affect their ability to maintain a comfortable standard of living. The Psychological Impact, Increased responsibilities and expectations can lead to higher stress levels and anxiety, negatively affecting emotional Welfare.

e) Adequate Pension (500 euros): Coefficient B = -4.853, Std. Error = 1.286, Wald Chi-Square = 14.250, Sig. = 0.000, Exp(B) = 0.008. The Economic Impact, an insufficient pension (below 500 euros) directly affects pensioners' ability to maintain a decent standard of living and cover daily expenses. This can lead to poverty and financial insecurity. The Psychological Impact, Financial insecurity resulting from a low pension can lead to anxiety and stress, contributing to depression and feelings of helplessness.

f) Pension Spending: Coefficient B = 10.433, Std. Error = 2.487, Wald Chi-Square = 17.596, Sig. = 0.000, Exp(B) = 33945.136. The Economic Impact, Pensioners who manage their pension spending better can maintain a stable financial situation and are more likely to sustain a comfortable standard of living. Effective financial management contributes to economic Welfare. The Psychological Impact, feeling in control of their finances can reduce stress levels and improve emotional Welfare, leading to greater feelings of security and contentment.

g) Other Income (Son, Daughter, Investments): Coefficient B = -12.115, Std. Error = 2.699, Wald Chi-Square = 20.148, Sig. = 0.000, Exp(B) = 5.475E-6. The Economic Impact, Dependence on additional income sources (from children, investments) can have a negative economic impact as these sources are often unreliable or unpredictable. The Psychological Impact, the insecurity that comes from not having a stable income can negatively affect emotional Welfare, causing stress and anxiety.

h) Pensioners' Hobby: Coefficient B = -3.402, Std. Error = 1.441, Wald Chi-Square = 5.576, Sig. = 0.018, Exp(B) = 0.033. The Economic Impact, Pensioners engaged in hobbies may benefit from additional economic opportunities (e.g., selling art or crafts), which can contribute to financial support. The Psychological Impact, engaging in hobbies can provide enjoyment and fulfillment, which improves emotional Welfare and reduces depression and anxiety.

i) Pensioners' Saving: Coefficient B = -7.460, Std. Error = 2.239, Wald Chi-Square = 11.102, Sig. = 0.001, Exp(B) = 0.001. The Economic Impact, Pensioners without savings often struggle to maintain a decent standard of living and may face challenges covering unexpected expenses. The Psychological Impact, the lack of savings creates feelings of insecurity and anxiety, negatively affecting emotional Welfare.

j) Main Concern: Pension decrease (120 euros): Coefficient B = 12.214, Std. Error = 3.809, Wald Chi-Square = 10.281, Sig. = 0.001, Exp(B) = 201651.201. The Economic Impact, A reduction in pension amounts directly affects pensioners' ability to maintain a decent standard of living, making it difficult to cover expenses. The Psychological Impact, Financial insecurity due to pension cuts leads to anxiety and stress, increasing feelings of depression and fear about the future.

k) Adequate Support from the State: Coefficient B = 29.120, Std. Error = 5145.5635, Wald Chi-Square = .000, Sig. = 0.995, Exp(B) = 1.66495E+11. The Economic Impact, Lack of adequate support from the state leads to poverty and financial insecurity for pensioners, leaving them unprepared for economic challenges. The Psychological Impact, The absence of state support can cause feelings of isolation and dissatisfaction, negatively impacting social and emotional Welfare.

l) Health of Pensioners: Coefficient B = -10.965, Std. Error = 5145.5641, Wald Chi-Square = .000, Sig. = 0.998, Exp(B) = 1.730E-5. The Economic Impact, Health problems directly affect pensioners' expenses, adding financial pressure and making it difficult to maintain a comfortable living standard. The Psychological Impact, Health issues lead to insecurity and anxiety, decreasing emotional Welfare and contributing to depression.

Variables such as age, gender, inadequate pensions, and lack of state support have significant economic and psychological impacts on pensioners. These factors contribute to financial insecurity and stress, affecting overall Welfare. Effective pension management, social support, and engagement in hobbies can improve both economic and psychological outcomes for pensioners.

• Summary of Coefficients (Table 19)
  • Marital Status (-7.850, p = 0.000): Marital status has a strong effect, with married retirees having better well-being.
  • Adequate Pension (-4.853, p = 0.000): Adequate pensions are associated with higher well-being.
  • Expenses (10.433, p = 0.000): Effective management of expenses improves well-being.
  • Other Income (-12.115, p = 0.000): Other sources of income have a significant positive impact.
  • Hobbies (-3.402, p = 0.018): Activities and hobbies contribute to well-being.
  • Health (-10.965, p = 0.000): Health is one of the most important factors for well-being.
  • Savings (-7.460, p = 0.001): Contrary to expectations, savings do not show a clear positive impact.
  • Legal support (29.120, p = 0.995): No statistically significant impact, suggesting the need for legal improvements.

According to Eq. (21), the model's explanation percentage for well-being is 31.01%, which means that 68.99% of the influencing factors may be outside this model.

- Summary of Results H2 and Link to Research Questions

RQ1: Variables such as adequate pension, marital status, health and other sources of income positively affect the well-being of pensioners.

RQ2: Income (adequate pension and additional sources) has a greater impact than savings or legal support.

RQ3: The most influential socio-demographic factors are marital status and health of pensioners, while hobbies also have a positive impact.

RQ4: Pensioners assess that legal support from the state is insufficient and are concerned about the low level of pensions, seeking improvements.

Tables 19 and 20 confirm that marital status, adequate pension, pension spending, additional income, hobbies, and health significantly enhance pensioners' welfare. Savings and legal support show no significant effect, indicating areas for improvement. Concerns about low pensions negatively impact well-being, emphasizing the need for better financial security and policy reforms.

Table 20. Independent-Samples Kruskal Wallis test summary

The results of the Kruskal-Wallis test presented in Table 20 show a test statistic of 10.719 with 3 degrees of freedom and an asymptotic p-value of 0.013, indicating statistically significant differences between the compared groups at the 95% confidence level (p < 0.05). This leads to the rejection of the null hypothesis (H0), which assumes no significant differences among groups, and the acceptance of the alternative hypothesis (H1), which posits the existence of significant differences.

The analysis reveals that factors such as gender, marital status, savings, health status, adequate pension income, additional income sources, hobbies, and family support have a significant impact on the welfare of pensioners in Kosovo. Conversely, variables like pension spending, concerns about pension reduction, and state support did not show statistically significant effects in this study.

These findings highlight the importance of personal, economic, and social factors in improving the quality of life for pensioners, suggesting that social and economic policies should focus on strengthening these areas to enhance pensioners’ welfare. Furthermore, the lack of significant impact from state support and pension spending indicates a need for further investigation and targeted policy interventions in these domains.

14.png

Figure 14. Pensioner’s welfare

Retirement has become a pivotal phase in citizens' lives, reflecting significant changes in how income is organized after active employment. For most retirees, income is derived from multiple sources: primarily, earnings from investments which depend on market performance and can be variable and clearly defined social security benefits. In addition, this study has also examined the basic pension, analyzing its importance in providing a stable income level for retirees.