Healthcare Systems and Inequality in the European Union: A Comparative Analysis

Abstract

The third United Nations Sustainable Development Goal promotes health and well-being. Despite the existence of academic literature examining the relationship between health and income inequality, evidence on the role of healthcare systems in this inequality remains limited. This article aims to analyse the extent to which healthcare systems are associated with differences in economic inequality. To this end, a balanced panel of 27 European Union countries for the period 2005–2022 is used, applying t-tests for differences in means and linear regression models using S80/S20, Gini and Palma inequality measures. The main results show that countries with a Social Health Insurance System (SHIS) exhibit, on average, lower levels of income inequality, despite not being the highest spenders on healthcare. On the other hand, healthcare expenditure has a negative and statistically significant relationship with inequality, whereas in countries with a Mixed Healthcare System (MHS), this association is not statistically significant. A disaggregated analysis of public and private spending indicates that public expenditure is particularly relevant in SHIS countries being negatively associated with income inequality, whereas this relationship differs in countries with a National Health System (NHS). Thus, it is concluded that healthcare systems display significant differences in the relationship under study.

1. Introduction

Population health is a fundamental aspect of both social well-being and the economic development of countries. This is reflected in the third Sustainable Development Goal (SDG 3), which seeks universal access to healthcare services and, thereby, to ensure good health for all individuals. Additionally, the United Nations 2030 Agenda highlights the close relationship between achieving this goal and the attainment of others, such as ending poverty (SDG 1), reducing inequalities (SDG 10), and ensuring access to basic resources like clean water (SDG 6), among others [1].

This appears to be supported by the growing interest in this field within the academic literature. Numerous studies point to the existence of a relationship between economic growth and health, specifically concluding that the health status of the population affects productivity and, consequently, the productive capacity of economies [2,3,4]. Similarly, other authors indicate that health, as a component of human capital, can influence the quality of the labour force. Thus, the presence of health problems not only diminishes individuals’ financial capacity by limiting their access to the labour market, but also by increasing family spending on health-related needs [5,6]. Apart from its impact on economic growth, there is also evidence of a relationship between health and income inequality. Academic literature generally agrees on the existence of a bidirectional causal link between the two, meaning that health affects income inequality and, conversely, income inequality influences health outcomes [7,8,9]. Some studies find that unequal income distribution results in poorer health by limiting access to basic goods and essential services such as education and healthcare for segments of the population [3,4,10]. In contrast, other research shows that improvements in health contribute to reducing inequalities. In this context, the public sector plays a particularly important role: investment in health through increased public spending, combined with universal healthcare coverage, tends to reduce economic inequalities [7,9,11,12], while a greater reliance on private spending may lead to increased inequality [10].

In light of the above, it is worth examining the role of healthcare systems in these relationships, since they are the institutions responsible for shaping both population access to health services and the mechanisms through which these services are financed. While most academic literature focuses on analysing the efficiency of healthcare systems—noting that, although some countries are more efficient, this often depends more on the country’s economic factors than on the type of healthcare system in place [13,14,15,16]—there are studies that emphasize how the institutional characteristics of these systems (financing methods, level of expenditure, availability of healthcare resources, etc.) can influence both health outcomes and their economic and redistributive effects [3,4,8,17,18,19].

While previous studies suggest the existence of a relationship between healthcare resources and income inequality, this association may also be influenced by broader institutional, political, and socioeconomic factors. Therefore, the relationships analysed in the present study should be interpreted with caution.

In this context, the aim of this article is to contribute to the academic literature by analysing the relationship between health resources and income inequality among different healthcare systems, focusing on European Union member states and their institutional heterogeneity.

For this purpose, it is necessary to distinguish between the main models of health care systems that exist in the European Union. Following the classification proposed by previous literature [14], European Union countries can be broadly grouped into three healthcare system models: the National Health System (NHS), also known as the Beveridge model, which is primarily funded through taxes and provides universal coverage to its citizens; the Social Health Insurance System (SHIS), or Bismarck model, mainly financed through mandatory social security contributions from employees and employers; and the Mixed Healthcare System (MHS), which combines features of the previous two systems while also relying heavily on private funding through voluntary health insurance or direct payments. Healthcare systems can also be understood as complex institutional arrangements in which financing mechanisms, resource allocation, access conditions, and public–private interactions jointly shape both health and socioeconomic outcomes. From this perspective, analysing the relationship between healthcare resources and income inequality requires considering healthcare systems not only as financing models, but also as broader institutional frameworks embedded within specific social and economic contexts.

Taking this into account, the choice of inequality indicator becomes particularly important. Although wealth inequality also represents an important and growing dimension of socioeconomic disparities, it reflects the accumulation of assets over long periods and is often more strongly associated with intergenerational transmission mechanisms, property ownership, and financial asset concentration. By contrast, income inequality is more directly related to current living conditions and access to healthcare services, disposable household resources, and the redistributive effects of taxation and public expenditure policies. For this reason, income-based indicators are generally more commonly used in the studies examining the relationship between healthcare systems and inequality. In the present work, three indicators are used to capture different dimensions of income distribution, namely the S80/S20, the Gini index, and the Palma ratio.

2. Materials and Methods

2.1. Dataset

The final sample consists of an aggregated country-level balanced panel of 27 European Union countries covering the years 2005 to 2022, which allows for a comparative assessment of the relationship between healthcare systems and income inequality across European Union member states. The selected variables were chosen based on their relevance in the literature, their comparability across European Union countries, and the availability of consistent longitudinal data for the period analysed. In this regard and following the approach of Moreno-Enguix et al. [16] and the OECD [20], it is important for our study that the variables used are controllable by the authority responsible for healthcare management and, consequently, by the established healthcare system.

A total of ten variables were selected, grouped into two categories. Specifically, three variables are used as indicators of inequality to capture different dimensions of income distribution and to test the robustness of the results. The S80/S20 ratio focuses on disparities between the highest and lowest income population groups, the Gini index reflects overall income inequality, and the Palma ratio emphasizes the concentration of income between the richest and poorest segments of the population. The remaining seven variables relate to various measures of healthcare and economic resources.

Table 1. Summary of variables used and sources.

Variable	Description	Source
Income Inequality
S80/S20	Ratio of the average income earned by the 20% of the population with the highest income to the average income earned by the 20% of the population with the lowest income.	Eurostat [21]
Gini	Inequality ratio measured by the difference in income.	Eurostat & WIID [22,23]
Palma	Ratio obtained by calculating the relationship between the income of the richest 10% of the population and the income of the poorest 40%.	WIID [23]
Resources
HBeds	Number of hospital beds per 1000 inhabitants.	Eurostat [24]
Physicians	Number of physicians per 1000 inhabitants.	Eurostat [25,26]
HS	Type of Healthcare System, which may be National Health System (NHS), Social Health Insurance System (SHIS) or Mixed Healthcare System (MHS).
TotHe/Tpop	Total healthcare expenditure in euros among the country’s population.	Eurostat & OECD [27,28,29]
PubHe/Tpop	Public health expenditure in euros among the country’s population.	Eurostat & OECD [27,28,29]
PriHe/Tpop	Private healthcare expenditure in euros among the country’s population.	Eurostat & OECD [27,28,29]
GDPpc	Total value of the country’s Gross Domestic Product among its population.	Eurostat [30]

Source: Own elaboration.

provides a description of these variables, specifying their measurement and the source from which they were obtained.

Regarding the 27 European Union (EU) member states, these countries have attracted growing interest in recent years because, despite having similar socioeconomic indicators, they exhibit substantial institutional heterogeneity in the healthcare sector, with different healthcare system structures shaped by cultural and historical factors [16]. Therefore, and given the research objective, the classification of countries by healthcare system is based on the frameworks proposed by Busse et al. [10] and Gaeta et al. [11]:

• National Health System: Cyprus, Denmark, Finland, Ireland, Italy, Latvia, Malta, Portugal, Spain and Sweden.
• Social Health Insurance System: Belgium, Estonia, France, Germany, Lithuania, Luxembourg, Netherlands, Poland, Czech Republic, Romania, Slovakia, Slovenia and Hungary.
• Mixed Healthcare System: Austria, Bulgaria, Croatia and Greece.

2.2. Statistical Analysis

In order to examine which healthcare systems have higher health expenditures and the role their health resources play in reducing inequality, inferential techniques were employed. Additionally, a logarithmic transformation was applied to each variable in Table 1, except for HS, due to the variance they exhibited.

The empirical strategy adopted in this study is not designed to establish causal relationships, but rather to identify statistical associations between healthcare resources and income inequality. Hence, the results should be interpreted with caution, as they may be influenced by unobserved factors or reverse causality.

First, independent-sample t-tests were conducted to examine the differences between the means of two groups. Given the three healthcare systems, three pairwise t-tests were performed. Specifically, the Welch’s t-statistic was chosen for its robustness against heterogeneity in population variances.

On the other hand, a series of linear regressions were carried out. For each income inequality indicator, two regressions were performed, resulting in a total of six regressions. The first regression involves separating the resources, including total healthcare expenditure per capita:

$${lnY}_{it}={\beta }_0+{\beta }_{1}ln{GDPpc}_{it}+{\beta }_2{lnHBeds}_{it}+{\beta }_3{lnPhysicians}_{it}+{\beta }_4{lnTotHe/Tpop}_{it}+u_{it},$$

(1)

where $Y_{it}$ represents the corresponding income inequality indicator.

The second specification, considers public and private healthcare expenditure per capita separately:

$${lnY}_{it}={\beta }_0+{\beta }_{1}ln{GDPpc}_{it}+{\beta }_2{lnHBeds}_{it}+{\beta }_3{lnPhysicians}_{it}+{\beta }_4{lnPubHe/Tpop}_{it}+{\beta }_4{lnPriHe/Tpop}_{it}+u_{it}$$

(2)

This process is then applied to three subsamples grouping countries according to their healthcare systems, resulting in a total of 18 regressions. To ensure the robustness of the results, heteroskedasticity and autocorrelation-consistent (HAC) standard errors were applied. The results report the adjusted R value to allow for comparability in the quality of the model fits.

Version 4.5.0 of the R software was used to carry out this entire process.

3. Results

Looking at the countries by healthcare system type (Section 2.1), those with a National Health System (NHS) are primarily Nordic countries (Denmark, Finland, Latvia, and Sweden) and Mediterranean countries (Cyprus, Italy, Portugal, and Spain). Countries with a Social Health Insurance System (SHIS) are located in Central Europe (Belgium, France, Germany, Luxembourg, and the Netherlands) and Eastern Europe (Czech Republic, Estonia, Hungary, Lithuania, Poland, Romania, Slovakia, and Slovenia). As for countries with a Mixed Healthcare System (MHS), they have populations under 10 million and are situated in Central and Southeastern Europe (Austria, Croatia, Bulgaria, and Greece).

Table 2. t-tests performed on the SHIS (S) and the NHS (N).

Variable	HS	N	Mean	p-Value
S80S20	S	234	1.4854	0.0001
	N	180	1.5695
Gini	S	234	3.3467	0.0000
	N	180	3.4033
Palma	S	234	0.0292	0.0000
	N	180	0.1104
GDPpc	S	234	9.8863	0.0000
	N	180	10.1777
HBeds	S	234	1.7788	0.0000
	N	180	1.2897
Physicians	S	234	5.7741	0.0000
	N	180	5.9033
TotHe/Tpop	S	234	7.3947	0.0000
	N	180	7.7997
PubHe/Tpop	S	234	7.1304	0.0001
	N	180	7.4389
PriHe/Tpop	S	234	5.6003	0.0000
	N	180	6.2858

Source: Own elaboration.

presents the t-tests conducted on each transformed variable, comparing countries with a Social Health Insurance System (S) and those with a National Health System (N). Countries with a National Health System exhibit higher income inequality, as measured by all three indicators (S82S20, Gini, Palma). On the other hand, total healthcare expenditure per capita, as well as its breakdown into public and private spending, is higher in National Health System countries than in Social Health Insurance System countries. Regarding physical resources, a similar pattern is observed for the number of physicians, but not for hospital beds.

Therefore, this initial overview shows that countries with a National Health System have higher per capita wealth, spend more on healthcare (both public and private), have more physicians per capita, and exhibit higher levels of inequality.

Table 3. t-tests performed on the SHIS (S) and the MHS (M).

Variable	HS	N	Mean	p-Value
S80S20	S	234	1.4854	0.0000
	M	72	1.6824
Gini	S	234	3.3467	0.0000
	M	72	3.4527
Palma	S	234	0.0292	0.0000
	M	72	0.1849
GDPpc	S	234	9.8863	0.0011
	M	72	9.5749
HBeds	S	234	1.7787	0.4141
	M	72	1.8032
Physicians	S	234	5.7741	0.0000
	M	72	6.0814
TotHe/Tpop	S	234	7.3947	0.0295
	M	72	7.1419
PubHe/Tpop	S	234	7.1304	0.0040
	M	72	6.7677
PriHe/Tpop	S	234	5.6003	0.4675
	M	72	6.6797

Source: Own elaboration.

shows the t-test performed between countries in the Social Health Insurance System group and those in the Mixed Healthcare System (M) group. Countries with Mixed Healthcare System exhibit higher levels of income inequality (S80/S20, Gini, and Palma) than countries with Social Health Insurance System. They also have more physicians, lower total and public healthcare expenditure, and a lower per capita GDP.

Finally,

Table 4. t-tests performed on the NHS (N) and the MHS (M).

Variable	HS	N	Mean	p-Value
S80S20	N	180	1.5695	0.0003
	M	72	1.6824
Gini	N	180	3.4033	0.0050
	M	72	3.4527
Palma	N	180	0.1104	0.0100
	M	72	0.1849
GDPpc	N	180	10.1777	0.0000
	M	72	9.5749
HBeds	N	180	1.2897	0.0000
	M	72	1.8032
Physicians	N	180	5.9033	0.0000
	M	72	6.0814
TotHe/Tpop	N	180	7.7997	0.0000
	M	72	7.1419
PubHe/Tpop	N	180	7.4389	0.0000
	M	72	6.7677
PriHe/Tpop	N	180	6.2858	0.0000
	M	72	6.6797

Source: Own elaboration.

displays the t-test applied to countries belonging to the National Health System (NHS) and the Mixed Healthcare System (MHS) groups. The results show that NHS countries have, on average, lower income inequality across all three indicators compared to MHS countries. At the same time, NHS countries spend more on total, public, and private healthcare per capita, although they have fewer hospital beds and physicians, and a higher per capita GDP. This is noteworthy, as it suggests that despite higher healthcare spending, these countries have fewer human (physicians) and physical (beds) resources, which may indicate that spending is directed toward other healthcare resources or that the cost of these resources is higher in these countries.

Table 5. Comparative Summary of the Variables by Healthcare System.

	Social Health Insurance	National Health	Mixed Healthcare
Inequality	↑	↑↑	↑↑↑
GDPpc	↑↑	↑↑↑	↑
HBeds	↑↑	↑	↑↑↑
Physicians	↑	↑↑	↑↑↑
TotHe	↑↑	↑↑↑	↑
PubHe	↑↑	↑↑↑	↑
PriHe	↑	↑↑↑	↑↑

The greater the number of arrows, the higher the value of the variable. Source: Own elaboration.

provides a summary of the previous tables, showing that countries with a Mixed Healthcare System (MHS) exhibit the highest average income inequality and fewer healthcare resources per capita, with the exception of hospital beds. Countries with a Social Health Insurance System (SHIS) have the lowest levels of inequality and occupy an intermediate position regarding per capita healthcare resources (except for physicians, where they have fewer). Finally, countries with a National Health System (NHS) display intermediate inequality levels, but overall, they are the highest consumers of healthcare resources.

A more detailed analysis of the countries within each system reveals substantial intra-group heterogeneity, which is important for interpreting the aggregate results. For instance, within the MHS group, countries such as Bulgaria exhibit considerably higher levels of inequality compared to Austria.

Focusing on the NHS group, the Nordic countries show lower inequality compared to countries such as Italy, Latvia, and Spain, indicating that the same institutional model may operate differently depending on broader socioeconomic contexts.

In the case of SHIS countries, differences emerge between the former Soviet states and the rest: while Eastern European countries exhibit high inequality levels (Estonia, Lithuania, and Romania), the more westernized countries (Czechia and Slovenia) display lower inequality, even below that of Central European countries with a stronger capitalist tradition (Belgium, France, Germany, and Luxembourg). Some of the possible causes of these differences may be historical, political, institutional, or geocultural; however, these issues are beyond the scope of this study. This heterogeneity suggests that the estimated relationships should be interpreted as average patterns rather than uniform effects across all countries within each group.

While the t-test results provide an initial descriptive comparison across healthcare systems, they do not account for the joint influence of multiple factors. To address this limitation, the following section presents the results of the regression analysis separated by healthcare system.

3.1. Countries with a Social Health Insurance System

Table 6. Regression models for countries with a Social Health Insurance System.

	S80/S20 (1)	S80/S20 (2)	Gini (3)	Gini (4)	Palma (5)	Palma (6)
Constant	0.9922 (0.6887)	0.8455 (0.7579)	2.9278 *** (0.4203)	2.8063 *** (0.4565)	−0.6410 (0.6126)	−0.8369 (0.6683)
GDPpc	0.1038 (0.0938)	0.0982 (0.0982)	0.0586 (0.0572)	0.0645 (0.0591)	0.0755 (0.0834)	0.0875 (0.0866)
HBeds	0.1321 ** (0.0658)	0.1112 * (0.0656)	0.1070 *** (0.0401)	0.0909 ** (0.0395)	0.1589 *** (0.0585)	0.1375 ** (0.0579)
Physicians	0.1119 (0.0820)	0.1080 (0.0847)	0.0579 (0.0500)	0.0554 (0.0510)	0.1037 (0.0730)	0.1026 (0.0747)
TotHe/Tpop	−0.1919 *** (0.0738)		−0.0931 ** (0.0450)		−0.1302 ** (0.0656)
PubHe/Tpop		−0.2435 *** (0.0749)		−0.1407 *** (0.0451)		−0.1963 *** (0.0661)
PriHe/Tpop		0.1034 ** (0.0501)		0.0753 ** (0.0302)		0.1000 ** (0.0442)
R-squared	0.2129	0.2373	0.1540	0.1922	0.1614	0.1951
Adjusted R-squared	0.1986	0.2197	0.1387	0.1735	0.1462	0.1764

Significance level: <1% (***), <5% (**), <10% (*). Standard deviations are reported in parentheses. Source: Own elaboration.

summarizes the results of the regressions conducted using data from countries with a Social Health Insurance System (SHIS). The columns represent the dependent variable estimated in each regression, with the resulting coefficients (betas) shown below, and their corresponding standard deviations in parentheses. The significance level of the variables for estimating the dependent variable is indicated by asterisks (*), with their meaning provided in the table notes.

Additionally, each income inequality indicator appears in two models: the left-hand column (1, 3, and 5) considers total healthcare resources per capita, while the right-hand column (2, 4, and 6) separates resources into public and private spending.

Considering that explanatory variables are deemed statistically significant if their significance level is equal to or below 5%, the results in Table 6 indicate that all variables are significant in explaining income inequality across all measures used, except for GDP per capita and the number of physicians per thousand inhabitants. Additionally, it is worth noting that the only variables negatively associated with income inequality are total and public healthcare expenditure per capita; in other words, an increase in TotHe/Tpop or PubHe/Tpop is associated with lower income inequality, with the latter variable playing a major role on it.

By examining the Adjusted R-squared values, it is concluded that the best model is the one presented in column 2 of Table 6 (values in bold), which uses the S80/S20 ratio as the dependent variable and separates healthcare expenditure into public and private components as independent variables. This model shows that public healthcare spending per capita is negatively associated with inequality, whereas private healthcare spending increases it. Additionally, the R-squared indicates that approximately 23.7% of the sample variation in income inequality is explained by the estimated regression.

3.2. Countries with National Health System

Table 7. Regression models for countries with a National Health System.

	S80/S20 (1)	S80/S20 (2)	Gini (3)	Gini (4)	Palma (5)	Palma (6)
Constant	5.8364 *** (0.8933)	6.4746 *** (0.8161)	5.7244 *** (0.4802)	5.7483 *** (0.4566)	3.5925 *** (0.7341)	3.5841 *** (0.6969)
GDPpc	−0.3450 *** (0.0877)	−0.2941 *** (0.0686)	−0.1462 *** (0.0472)	−0.1082 *** (0.0384)	−0.2056 *** (0.0721)	−0.1455 ** (0.0586)
HBeds	−0.3000 *** (0.0536)	−0.2946 *** (0.0485)	−0.1745 *** (0.0288)	−0.1632 *** (0.0271)	−0.2390 *** (0.0440)	−0.2203 *** (0.0414)
Physicians	0.0461 (0.0955)	0.0679 (0.0875)	0.0174 (0.0513)	0.0499 (0.0489)	0.0146 (0.0785)	0.0693 (0.0747)
TotHe/Tpop	−0.0819 (0.0703)		−0.0912 ** (0.0378)		−0.1499 ** (0.0578)
PubHe/Tpop		0.0007 (0.0378)		−0.0495 *** (0.0211)		−0.0847 *** (0.0323)
PriHe/Tpop		−0.3076 *** (0.0469)		−0.1528 *** (0.0262)		−0.2366 *** (0.0400)
R-squared	0.6310	0.7089	0.6611	0.7116	0.6606	0.7127
Adjusted R-squared	0.6216	0.6995	0.6525	0.7023	0.6521	0.7034

Significance level: <1% (***), <5% (**). Standard deviations are reported in parentheses. Source: Own elaboration.

presents the results of the regressions conducted using data from countries with a National Health System (NHS). As in the previous table, the columns represent the dependent variable estimated in each regression, with the coefficients (betas) shown below and their standard deviations in parentheses. The significance level of the variables is indicated by asterisks.

In the case of these countries, all variables are significant in explaining income inequality across all measures used, except for the number of physicians per thousand inhabitants. Additionally, it should be noted that total healthcare expenditure per capita and public healthcare expenditure per capita in the regressions with the S80/S20 ratio as the dependent variable (columns 1 and 2, respectively) are also not significant. On the other hand, a closer look at the models shows that in those where healthcare expenditure is separated into public and private components (models 2, 4, and 6), private healthcare expenditure per capita has the strongest negative correlation with inequality, especially when measured by the S80/S20 ratio, followed by hospital beds per thousand inhabitants.

Additionally, by comparing the Adjusted R-squared values, it is concluded that the best model is the one with the Palma ratio as the dependent variable and public and private healthcare expenditure per capita as the independent variables (column 6 of Table 6, values in bold). It is also worth noting that this model has greater explanatory power than those estimated in Table 6, accounting for 71.3% of the sample variation in the dependent variable.

3.3. Countries with a Mixed Healthcare System

Finally,

Table 8. Regression models for countries with a Mixed Healthcare System.

	S80/S20 (1)	S80/S20 (2)	Gini (3)	Gini (4)	Palma (5)	Palma (6)
Constant	1.4478 (1.3285)	2.7586 ** (1.3361)	3.1842 *** (0.7821)	4.2498 *** (0.7879)	−0.7272 (1.3931)	1.2447 (1.4035)
GDPpc	−0.1922 (0.2104)	−0.3712 * (0.1887)	−0.1083 (0.1238)	−0.2301 ** (0.1113)	−0.1916 (0.2206)	−0.4134 ** (0.1982)
HBeds	−0.0145 (0.1192)	−0.1279 (0.1088)	0.0194 (0.0701)	−0.0638 (0.0642)	0.1322 (0.1249)	−0.0188 (0.1143)
Physicians	0.4711 *** (0.0965)	0.3815 ** (0.1517)	0.2797 *** (0.0568)	0.1870 ** (0.0895)	0.5234 *** (0.1012)	0.3505 ** (0.1594)
TotHe/Tpop	−0.1073 (0.1867)		−0.0605 (0.1099)		−0.0951 (0.1958)
PubHe/Tpop		0.0328 (0.1340)		0.0285 (0.0790)		0.0632 (0.1407)
PriHe/Tpop		0.0296 (0.0559)		0.0337 (0.0330)		0.0657 (0.0588)
R-squared	0.7835	0.7915	0.7629	0.7710	0.7349	0.7440
Adjusted R-squared	0.7693	0.7738	0.7473	0.7516	0.7176	0.7223

Significance level: <1% (***), <5% (**), <10% (*). Standard deviations are reported in parentheses. Source: Own elaboration.

presents the results of the regressions conducted using data from countries with a Mixed Healthcare System. As in the other tables, the columns indicate the dependent variable estimated in each regression, with the coefficients (betas) shown below and their standard deviations in parentheses. The significance level of the variables is indicated by asterisks.

The key difference for these countries, compared to those analysed in the previous sections, is that the variable physicians per thousand inhabitants is the only one statistically significant in explaining income inequality, regardless of the dependent variable used, whereas GDP per capita is only significant in models where healthcare expenditure per capita is separated into public and private components.

Regarding the coefficient values (betas), although the only healthcare-related variables that would have a negative association with income inequality are HBeds and TotHe/Tpop, these are not statistically significant. Therefore, in these countries, an increase in GDP per capita is the most negatively correlated with respect to income inequality (models in columns 2, 4, and 6 of Table 8), whereas physicians per thousand inhabitants appear to amplify inequality.

Finally, the best model, based on the Adjusted R-squared, is shown in the second column of the table, where public and private healthcare expenditure per capita are included as independent variables and the S80/S20 ratio as the dependent variable. Its Adjusted R-squared is slightly higher than that of the best model in Table 7 and clearly superior to that in Table 6. Additionally, the R-squared of this model explains 79% of the sample variation in income inequality based on the independent variables (values in bold).

The previous results highlight significant differences in the relationship between healthcare resources and income inequality across healthcare systems. The following section discusses the implications of these findings in the context of the existing literature.

4. Discussion

Table 6, Table 7 and Table 8 show that healthcare spending is negatively correlated with income inequality in countries with a Social Health Insurance System and to a lesser extent in those with a National Health System, in line with the findings of Anderson et al. [7] and Ferreiro-Pérez et al. [12]. This similarity may be explained by the redistributive role of healthcare spending, particularly when it improves access to essential services for lower-income groups. However, these authors did not analyse effects by healthcare system, which represents a notable finding regarding the differences in how public healthcare expenditure is associated with inequality depending on the system. In contrast, the results for countries with a Mixed Healthcare System indicate that total healthcare spending does not affect economic inequality, reaffirming the behavioural disparities observed according to healthcare system type.

Regarding the differentiation by type of expenditure, although Bhattacharjee et al. [10] and Ferreiro-Pérez et al. [12] argue that an increase in public healthcare spending reduces inequality, the results obtained here reveal differences depending on the country’s healthcare system (Table 6, Table 7, Table 8 and

Table 9. Summary Table of Regression Models (separating Public and Private Healthcare Expenditure) by Healthcare System.

	Social Health Insurance System			National Health System			Mixed Healthcare System
	S80/S20 (2)	Gini (4)	Palma (6)	S80/S20 (2)	Gini (4)	Palma (6)	S80/S20 (2)	Gini (4)	Palma (6)
GDPpc	0.098 (0.318)	0.065 (0.276)	0.088 (0.313)	−0.294 (0.000)	−0.108 (0.005)	−0.146 (0.014)	−0.372 (0.054)	−0.230 (0.043)	−0.413 (0.041)
HBeds	0.111 (0.092)	0.091 (0.022)	0.138 (0.018)	−0.295 (0.000)	−0.163 (0.000)	−0.220 (0.000)	−0.128 (0.245)	−0.064 (0.324)	−0.019 (0.870)
Physicians	0.108 (0.204)	0.055 (0.279)	0.103 (0.171)	0.068 (0.439)	0.050 (0.310)	0.069 (0.355)	0.382 (0.015)	0.187 (0.041)	0.350 (0.032)
PubHe/Tpop	−0.244 (0.001)	−0.141 (0.002)	−0.196 (0.003)	0.001 (0.986)	−0.049 (0.021)	−0.085 (0.010)	0.033 (0.808)	0.028 (0.720)	0.063 (0.655)
PriHe/Tpop	0.103 (0.04)	0.075 (0.013)	0.100 (0.025)	−0.308 (0.000)	−0.153 (0.000)	−0.237 (0.000)	0.030 (0.599)	0.034 (0.311)	0.066 (0.268)

Significance levels are reported in parentheses. Source: Own elaboration.

). Thus, in Mixed Healthcare System (MHS) countries, higher public health expenditure does not reduce inequality, whereas this association is observed in countries with the other two systems except in the case of the National Health System when the S80/S20 indicator is used. These findings suggest that the interaction between public and private sectors may play a more complex role than previously assumed.

As for private healthcare expenditure, our results (Table 6, Table 7, Table 8 and Table 9) show that it has a significant negative association with inequality in National Health System countries, while it increases inequality in those with a Social Health Insurance System, the latter being consistent with the findings of previous authors [10,12]. In the case of Mixed Healthcare System countries, this type of expenditure does not have a statistically significant relationship with income inequality.

However, the results obtained do not align with those proposed by Ataguba [17], who concluded that greater direct taxation would lead to higher levels of redistribution and, consequently, lower inequality. Table 2 shows that European countries with a National Health System, which is more heavily financed through taxation than other systems, exhibit higher levels of inequality than those with a Social Health Insurance System, despite having higher healthcare expenditure per capita. This discrepancy may be related to differences in the composition of public spending, the efficiency of resource allocation, or broader institutional factors that go beyond the financing mechanism itself.

Similarly, Table 7 shows that in NHS countries, private expenditure has a stronger negative relationship with income inequality than public expenditure, whereas in SHIS countries, public healthcare spending displays a negative correlation within equality, and private spending is positively associated (Table 6), even though this system is financed through social contributions. This finding points out that the association between healthcare resources and inequality is highly dependent on institutional context. This now disaggregated relationship would otherwise be hidden as in Bhattacharjee et al. [10] and Ferreiro-Pérez et al. [12].

Therefore, the conclusions of Siciliani and Cylus [4] also do not appear to hold, as the relationship between healthcare resources and income inequality does not seem to be explained solely by how the system is financed or by its redistributive mechanisms.

Turning to non-financial healthcare resources, an uneven association of hospital beds with inequality can be observed depending on the healthcare system, as shown in Table 6, Table 7 and Table 8 and summarized in Table 9. On the one hand, in the National Health System, an increase in the number of hospital beds is negatively associated with inequality, whereas in countries with a Social Health Insurance System, the relationship is positive. One possible explanation is that, in the former system, this increase benefits lower-income groups, while in SHIS countries, greater access restrictions linked to contribution-based financing do not help reduce inequality. As for human resources (physicians), they generally do not have a statistically significant relationship with inequality in the two main healthcare systems (Table 6, Table 7 and Table 9). Although the results do not allow for a clear explanation, a more in-depth analysis of the relationships of this variable may lead to future research.

The last variable analysed, GDP per capita, leads to several reflections. First, it is higher in countries with a National Health System than in those with a Social Health Insurance System (Table 9), which results in higher total, public, and private healthcare expenditure; however, this does not translate into lower inequality (Table 2). At the same time, across all six models for the NHS (Table 6) we can see a negative and significant relationship with inequality, whereas in SHIS countries the relationship is positive but not statistically significant (Table 7). This could suggest that in NHS countries, starting from higher levels of inequality, increases in GDP per capita may lead to higher healthcare spending which, combined with broader access to healthcare resources, helps reduce income inequality. Hence, the NHS capacity and incentives may be effective in reducing poverty. This result is in line with the findings of O’Donnell [3], although in that work such association was studied using ill-health factors like health-related loss of earnings, which reinforces a negative relationship even by using different proxies for inequality On the other hand, increases in GDP per capita in SHIS countries appear to have a neutral association with inequality due to their lower levels, which is in line with previous aggregate findings like those of Schneider et al. [18]. Overall, these findings highlight that similar levels of healthcare spending may be associated with different inequality outcomes depending on how healthcare systems are structured and financed. These results also suggest the existence of substantial heterogeneity within healthcare system categories, particularly among countries classified under Mixed Healthcare Systems, whose institutional arrangements are comparatively more diverse. From a theoretical perspective, the results can be interpreted within the framework of the redistribution function of the public sector. Healthcare is commonly considered a merit good, meaning that its provision is socially desirable regardless of individuals’ ability to pay. In this context, public healthcare spending may be associated with lower income inequality by improving access to essential services, especially for lower-income groups. This perspective may help to explain some of the differences observed across healthcare systems, as the redistribution capacity of healthcare is likely to depend not only on the level of spending, but also on how access to services is structured and financed.

In this regard, the present work focuses specifically on the relationship between healthcare systems and income inequality, without incorporating additional dimensions such as health outcomes, patient satisfaction, or perceived quality of care. Although these aspects are highly relevant from a systems perspective, they were beyond the scope of the present analysis due to data comparability constraints across countries and the objective of maintaining a focused institutional comparison.

At the same time, it is important to note that the relationship between healthcare spending and inequality may be bidirectional. Higher levels of inequality may generate greater demand for public intervention and redistribution through healthcare systems. This potential reverse relationship highlights the need to interpret the results as statistical associations rather than causal effects.

Additional analyses incorporating quadratic terms suggest the presence of some non-linear effects, particularly in countries with National Health Systems and, to a lesser extent, in Social Health Insurance Systems. However, these effects are not consistent across all specifications, indicating that non-linearity may play a complementary rather than central role in explaining the observed differences. Regarding the study’s limitations, these stem both from the availability of comparable data across European Union countries and from the scope of the analysis, which did not account for potential effects that other institutional or social factors might have on the relationship between healthcare systems and income inequality. Also, the findings should be interpreted as evidence of association rather than causation, given that the empirical strategy did not contemplate causality analysis. Another limitation of the study is the presence of substantial heterogeneity within each healthcare system group, which may hide country-specific dynamics that are not fully captured in the aggregated analysis.

5. Conclusions

Based on data from the 27 European Union member states, this study analysed the relationship between healthcare systems and income inequality. The results show that different healthcare models (Social Health Insurance System, National Health System, and Mixed Healthcare System) exhibit significant differences both in the resources available and in their association with inequality.

First, the results show that countries with a Social Health Insurance System (50% of the countries analysed) exhibit, on average, lower levels of inequality than countries with a National Health System (35%) and, especially, those with a Mixed Healthcare System (15%). However, NHS countries are those with the highest levels of healthcare expenditure per capita and the lowest number of hospital beds, while MHS countries display lower levels of spending but relatively higher endowments of physical and human resources. Finally, SHIS countries tend to occupy an intermediate position.

Second, the analysis also reveals that the relationship between healthcare expenditure and inequality varies depending on the structure of the healthcare system. This is particularly evident in countries with a Social Health Insurance System, where there is a negative and statistically significant relationship between total healthcare spending and income inequality, and to a lesser extent in countries with a National Health System. Additionally, different patterns are observed across systems regarding the impact of public and private healthcare expenditure, indicating that the configuration of the healthcare system shapes the redistributive effect of spending. Specifically, public healthcare expenditure is linked to lower inequality in SHIS countries, whereas in NHS countries it is private expenditure that has this behaviour. In contrast, in countries with a Mixed Healthcare System (MHS), its relationship is not statistically significant.

Regarding other healthcare resources, there is also no uniformity in the association of hospital beds and physicians per thousand inhabitants on income inequality across systems. Thus, while in countries with a NHS the number of hospital beds shows a negative and significant relationship, the opposite association is observed in countries with a SHIS. Meanwhile, in countries with a MHS, the only variable that shows a significant relationship with income inequality is the number of physicians, which has a positive association with income inequality.

Overall, the results obtained allow us to conclude that the institutional configuration of healthcare systems in Europe has a significant role on the relationship between healthcare resources and income inequality, as evidenced by the heterogeneity of the relationships displayed by different resources and inequality across systems.

From a public policy perspective, and with due caution, the results suggest that the relationship between healthcare resources and income inequality varies depending on the institutional design of the healthcare system, which has relevant implications for policy design.

Countries with a Social Health Insurance System display the lowest levels of healthcare spending while public healthcare expenditure is consistently associated with lower levels of inequality. Thus, policies aimed at strengthening public financing mechanisms and expanding coverage may be particularly effective in reducing disparities.

Regarding countries with a National Health System, the results indicate that private healthcare expenditure is more strongly associated with lower inequality levels. This suggests that a greater collaboration between the public and private sectors could enhance the system’s redistributive outcomes, although this should be approached cautiously to avoid undermining universal access.

Finally, in countries with a Mixed Healthcare System, the absence of statistically significant relationships between healthcare expenditure and inequality suggests that increasing spending alone may not be sufficient. In these contexts, structural reforms aimed at improving the efficiency, accessibility, and coordination of healthcare services may be more relevant than simply expanding financial resources.

As future research directions, it is suggested to expand the analysis by including additional socioeconomic and institutional variables that could help deepen the understanding of the relationship between healthcare resources and inequality. In this regard, the inclusion of additional inequality indicators, a more detailed breakdown of private healthcare expenditure, or the adoption of broader systemic and multidimensional approaches could represent interesting avenues for future research. It would also be valuable to examine whether the results would differ if countries were grouped based on geocultural similarities (Northern, Southern, Eastern, and Central Europe), economic factors, population size, levels of healthcare spending, or available resources. In addition, future research could build on this analysis by applying causal inference methods to analyse the direction and magnitude of these relationships, and non-linear models to assess whether the association between healthcare resources and inequality differs across the distribution of inequality levels. Another potential line of research is to investigate the role of human resources in healthcare systems, particularly physicians, in addressing income inequality, in order to better understand why this variable may be neutral or even positively associated with inequality depending on the type of healthcare system.