“Changes in the system of country’s population health care depending on the level of providing affordable housing”

This paper is devoted to providing affordable housing as a significant factor in public health management, inclusive growth, and SDG. The purpose is to empirically prove changes in the system of country’s population health care depending on the level of providing affordable housing. The empirical base includes time series and panel data for 27 EU countries during 2011–2019. Due to correlation analysis (Shapiro-Wilk test- ing, Spearman or Pearson correlation, lags in time), regression analysis, and building a dynamic panel estimation model with Sargan testing in STATA, the study empirically confirmed and formalized the impact of affordable housing funding on changes in the system of country’s population health care. In particular, the study found the depen- dence between overcrowding level and the share of homeowners with mortgages (a decrease of overcrowding level by an average of 0.61% with a time lag of 2 years due to an increase by 1%); the share of tenants on concessional terms/free (0.41% with a time lag of 3 years); and the share of public spending on housing development (0.25% with a time lag of 3 years). The direct relationship between the overcrowding and mortality from dangerous diseases (tuberculosis, AIDS, viral hepatitis, mental and behavioral disorders, diabetes, pneumonia) was also revealed. Public spending on housing under social protection programs (subsidies, etc.) proved to be the least effective. Preference should be given to the development of affordable mortgage lending (faster and greater effect). Generally, it impacts public management decisions in the health care system, social, and housing spheres.


INTRODUCTION
Today, improving public health is one of the priorities of public policy. However, despite understanding the importance and current unsatisfactory situation in the national public health model, the financial resources allocated to support this area and solve existing problems are limited and objectively insufficient. In this regard, it is important to reform the health care sector with an emphasis on social partnership and cooperation of crucial government actors, economic agents, and other stakeholders. This is vital to find adequate, timely, and effective mechanisms for implementing and improving the public health model, forms, methods, and tools for its practical implementation.
During the COVID-19 pandemic, governments and local authorities sought to strengthen existing networks and partnerships with communities to best meet people's needs (Mańka-Szulik & Krawczyk, 2022) while strengthening cross-sectoral cooperation (Sienkiewicz-

LITERATURE REVIEW
The interconnection between housing conditions and civil health has been repeatedly raised by civil leaders, scientists, politicians, etc. As a result, certain aspects are reflected in the WHO recommendations (WHO, 2011(WHO, , 2018(WHO, , 2021a(WHO, , 2021c. Moreover, this issue is one of the present directions of public policy and one of the potential possibilities to improve public health management in close relation to the social and financial policy of each country. The bibliometric analysis based on the Scopus database tools shows that there are only 83 documents for 52 years from 1970 to 2021 according to the search request "public health management and social housing financing" (the search in titles, abstracts, and keywords in the publications indexed by the Scopus database) ( Figure 1). Besides, more than 50% of the above publications were indexed in the last 15 years out of the 52 studied. Generally, the dynamics of scientific and publishing activity on the researched issue is positive, as evidenced by the trend shown in Figure 1.
The largest number of documents is affiliated with the USA, the United Kingdom, Canada, the Netherlands, Hong Kong, and Belgium. At the same time, the publications of authors from the USA, the United Kingdom, Canada, Australia, the Netherlands, Hong Kong, Singapore, and Sweden are the most cited as a result of the bibliometric analysis based on VosViewer software.
By conducting a bibliometric analysis using the VosViewer program, clusters of scientific research that are the most closely related to the key question of "public health management and social housing financing" were identified. The minimum number of keyword usages was set at 5, so 125 keywords were selected out of 1024 from the sample of publications. Then, for each of the 125 keywords, the total strength of the co-occurrence links with other keywords was calculated, and the keywords with the greatest total link strength were selected.
The four clusters of scientific research that are the most closely related to the key request of "public health management and social housing financing" are shown in Figure 2:  • the 1 st cluster -management, organization, and administration; public policy; social policy and security; public health service; health care planning and quality; legislation; demographic factors; sociology; migration; geographic factors; • the 2 nd cluster -financial management; government financing; program evaluation; economics; housing; public housing; public assistance; health care system and policy; health insurance and promotion; health status; public health; humans; poverty; employment; social behavior; social responsibility, support, and welfare; • the 3 rd cluster -government; local government; government programs; socioeconomics; socioeconomic factors; • the 4 th cluster -education; social psychology.
In the context of public health management and social housing financing, it is advisable to pay attention to separate areas of scientific development on this issue. Therefore, Habeeb (2021)  However, the impact of overcrowding on civil health, on the one hand, and the impact of financing social and affordable housing and its mechanisms on the level of overcrowding in the context decision-making in public health management, on the other, have not been investigated enough, empirically confirmed, and formalized, taking into account modern situation and time evolution. This determines the timeliness and relevance of this study and forms its aim.
The study aims to empirically prove changes in the system of country's population health care depending on the level of providing affordable housing.

METHODS
The bibliometric analysis was conducted based on the Scopus database tools and VosViewer software program. Statistical and graphical analysis was used for trends visualization and comparison in the EU and Ukraine, connected with overcrowding and public health particularly mortality from dangerous diseases. The evaluation of statistical significance, strength, and nature of the influence of changes in overcrowding level on changes in mortality from a number of dangerous diseases and the effect of changes in affordable housing financing on changes in overcrowding level was performed on the basis of correlation analysis. Before direct calculating correlation coefficients, the Shapiro-Wilk test for the subordination of factor variable to the normal distribution law was made. As a result, certain correlation coefficient (Pearson or Spearman) was chosen. Furthermore, possible time lags from 0 to 3 years were taken into account to increase obtained values adequacy. These impacts were also formalized by regression analysis and building a dynamic panel-data estimation model (

RESULTS
To conduct the study, a sample of 27 EU countries was formed, the data for which are analyzed for the period from 2011 to 2019 (time constraints are due to the availability of published statistics on the open portal of the EU statistical office for all studied indicators). The first analyzed indicator was the dynamics of housing overcrowding in the EU for 2011-2019 (Eurostat, 2022d). The lowest levels of overcrowding are in Cyprus, Ireland, Malta, the Netherlands, Belgium, Spain, and Luxembourg. Instead, the highest -in Romania, Latvia, Bulgaria, Croatia, and others. The gap between countries is from about 2% to 46%.
Then TB mortality in the EU for 2011-2019 (Eurostat, 2022a) was studied for further comparison with the trend in overcrowding in the EU.
A comparative analysis of the EU countries, as well as Ukraine, in terms of housing overcrowding and tuberculosis mortality in 2019 is presented in Figure 3.
In Ukraine, the death rate from tuberculosis (3 418 deaths in 2019) is not only significantly higher than the average level of EU countries but also hundreds and even thousands of times higher than in some EU countries. Disappointing statistics on mortality in Ukraine relate to other socially dangerous diseases too.    The highest rates are observed in Italy, France, and Germany.
It was also the systematization of data on mortality from mental and behavioral disorders, diabetes mortality, and mortality from pneumonia in the EU for 2011-2019 for further stage of research (Eurostat, 2022a). A comparative analysis of Ukraine and EU countries in terms of mortality rates from the above deadly diseases in 2019 is presented in Figure 6.   Mortality from mental and behavioral disorders is very high in Germany, Spain, France, and Italy compared to other EU countries. In the same countries, there are significant mortality rates from diabetes. Pneumonia is also a common cause of death in many countries (as in Ukraine).
A correlation analysis was performed based on the above data to test the positive impact of reducing overcrowding on the level of mortality from a number of deadly diseases. Before the direct calculation of the correlation coefficients, the Shapiro-Wilk test for the subordination of the factor variable to the law of normal distribution (Shapiro & Wilk, 1965) was made in the STATA software package. According to the results, the method of calculating the correlation coefficient -Pearson (Prob > z more than 0.05) (Pearson, 1896) or Spearman Prob > z less than 0.05) (Spearman, 1904)  One of the priority measures to reduce the overcrowding of housing is the financing of social housing. So, for the next research stage, the data on the following indicators were generalized: the share of homeowners with outstanding mortgages/housing loans, the share of tenants on preferential terms or free of charge, the share of public spending on housing development, the share of public spending on housing as a direction of social protection in the EU countries in 2011-2019 (Eurostat, 2022b(Eurostat, , 2022c.
To confirm or reject the hypothesis about the positive impact of the growth of the above indicators of social housing financing on the level of overcrowding, a correlation analysis was performed with preliminary verification of factor variables for compliance with the law of normal distribution by the Shapiro-Wilk test. Summary results of the assessment of statistical significance, nature, and strength of the impact of changes in social housing financing on changes in housing overcrowding based on a sample of 27 EU countries for 2011-2019 are shown in Appendix A, Table A2.
The results of the correlation analysis confirm the dependence between an increase in the financing of social housing and a decrease in the level of overcrowding, as evidenced by the predominant inverse nature of the impact of indicators (negative correlation coefficient). In particular, the following will help to reduce the level of overcrowding: • increase in the share of homeowners with mortgage/housing loan (reverse impact is in 21 of the 27 countries; high or very high interconnectedness is in 23 of the 27 sample countries, most often with a time lag of 1 year); • increase in the share of tenants on preferential terms/free of charge (reverse impact is in 14 out of 26 countries; high or very high interconnectedness is in 23 out of 26 sample countries (in Slovenia, this impact is not statistically significant), which is most often manifested in time lag 3 years); • increase in the share of public spending on housing development (reverse impact is in 15 of 26 countries; high or very high interconnectedness is in 21 of 26 sample countries (in Croatia, this impact is not statistically significant), which is most often manifested with a time lag 3 years); • increase in the share of public spending on housing as a direction of social protection (reverse impact is in 11 out of 22 countries; high or very high interconnectedness is in 19 out of 22 sample countries (in 5 sample countries, the impact of this indicator is not detected due to consistency in the dynamics of the study period), which is most often manifested with a time lag of 1 or 3 years).
It is empirically confirmed that the level of overcrowding can decrease by an average of 0.61% with a time lag of 2 years due to an increase in the share of homeowners with mortgages/housing loans by 1%.  The results of the assessment of the impact of changes in the share of tenants on preferential terms/free of charge on changes in the level of overcrowding are presented in Table 2.
The coefficients LnO (with a time lag of 1 year) and LnR (with a time lag of 3 years) are statistically significant (P > | z | does not exceed 0.05). Prob > chi2 = 0.00 does not exceed 0.05 and indicates the adequacy of the constructed model, which is also confirmed by the results of the Sargan test of overidentifying restrictions. The negative LnR coefficient characterizes the inverse relationship between the studied indicators. The constructed dynamic model for estimating the impact of changes in the share of tenants on preferential terms / free of charge on changes in the level of overcrowding is as follows: ,-1 ,-3 1.14 0.5 -0.41 .
It is empirically confirmed that the level of overcrowding can decrease by an average of 0.41% with a time lag of 3 years due to an increase in the share of tenants on preferential terms/free of charge by 1%.
The assessment results of the impact of changes in the share of government spending on housing development and housing as a direction of social protection on the change in the level of housing overcrowding are presented in Table 3.

DISCUSSION
Maqboo et al. (2015) made similar conclusions about understanding housing as a determinant of health. However, conclusions were not based on empirical evidence of cross-country analysis as in the case of this study. Garland et al. (2013) investigated the impact of social housing, especially on asthma, using the methods of questionnaires, observations, and testing, and this study is based on economic-mathematical modeling. However, the study did not pay attention to the problem of financing affordable housing and the according possibilities in decision-making in public management.
This paper strongly agrees that quite often, health professionals are not involved in the direct public decision-making. This can lead to a loss of opportunities to ensure and improve public health (Robert Wood Johnson Foundation & The Pew Charitable Trusts, 2016) and, in particular, taking into account housing affordability as its important determinant.
Rolfe et al. (2020) used a realist research approach, longitudinal study of tenants, and staff interviews to determine connections between tenants' experience of housing and their health. Meltzer and Schwartz (2016) proved the relationship between health and housing affordability due to building multivariate regression models. But the authors used the data only for New York, and the proposed models include some dummy indicators. Finally, Rahman (2007) proposed a case of the impact of housing finance on socio-economic uplift, but only for Bangladesh.
That is why this study, based on empirical evidence of 27 countries for 9 years and economic-mathematic modeling, objectively proves the dependence between an increasing the direction and volume of financing social housing and reducing the overcrowding of both own and rented housing and, accordingly, and an influence on the quality of living conditions (sanitary conditions, environmental factors, etc.). Moreover, it can have a positive impact not only on physical but also mental health and behavior of citizens (reduction of mental and behavioral disorders due to unbearable financial burden because of housing problems) and help free up financial resources of households to improve healthy eating, pay for medical supplies and medical services in the field of health care, which will eventually reduce mortality from a number of dangerous diseases: tuberculosis, AIDS, viral hepatitis, mental and behavioral disorders, pneumonia, diabetes.
Of course, the proved effect is not instantaneous and quantitatively significant, delayed in time, as evidenced by the defined time lags and low coefficient values in the regression equations. However, public management focused on improving health  care system and provision of affordable housing is justified in the medium and long term. The above results have a great potential both in future research on investigated issues and decision-making in public management in health care system, social and financial fields on cross-countries, nation-al and local levels of decisions making. Further research can be aimed at evaluating the effectiveness of the implementation of various mechanisms for the financial provision of social housing and detailing the forms and methods of its organization and management of this process.

CONCLUSIONS
The study empirically confirmed and formalized changes in the system of country's population health care depending on the level of providing affordable housing based on correlation-regression analysis and modeling the impact assessment of social housing financing on housing overcrowding and, at the same time, the impact of housing overcrowding on mortality from dangerous diseases (sample from 27 EU countries for 2011-2019).
It was found the dependence between the level of overcrowding and financing affordable housing. The level of overcrowding can decrease by an average of 0.61% with a time lag of 2 years due to an increase in the share of homeowners with mortgages/housing loans by 1%; it can decrease by an average of 0.41% with a time lag of 3 years due to an increase in the share of tenants on concessional terms/free by 1%; it can decrease by an average of 0.25% with a time lag of 3 years due to an increase in the share of public spending on housing development by 1%. Accordingly, it has been confirmed the direct dependence between reducing overcrowding and reducing mortality from tuberculosis (mostly without time lag or with a lag of 3 years), AIDS (mostly with a time lag of 3 years), viral hepatitis (mostly without time lag), mental and behavioral disorders (mostly without a time lag), diabetes (mostly with a time lag of 2 years), and pneumonia (mostly with a time lag of 3 years).
These issues affect public management decisions in the system of country's population health care, social, and housing spheres. The governments of the EU countries and especially in Ukraine (where the overcrowding rate is more than 50%) should pay particular attention to the impact of affordable housing providing as an essential factor in public health. When choosing directions/mechanisms of financing affordable housing, preference should be given to the development of affordable mortgage lending (faster and greater effect) compared to providing rental housing on concessional terms or free and public financing of housing development costs. At the same time, public spending on housing under social protection programs (subsidies, etc.) proved to be the least effective in addressing the problem of overcrowding and, consequently, improving public health. Note: * means the value of the correlation coefficient is not statistically significant for the studied period; r is the correlation coefficient; t is the time lag (from 0 to 3 years), at which the value of the correlation coefficient is the maximum for the studied period;  Note: * means the value of the correlation coefficient is not statistically significant for the studied period; -means the connection is not established, based on constant data in the dynamics of the studied indicator; r is the correlation coefficient; t -time lag, at which the value of the correlation coefficient is the maximum for the studied period