Quantifying the influence of informal payments on self-rated health: evidence from 26 post-communist countries

Abstract

In contrast to previous studies that have focused on proximal outcomes such as access to and the utilization of healthcare, this study establishes and quantifies the influence of informal payments (IP) directly on population self-rated health, which can be considered the ultimate outcome. More specifically, we examine how making informal payments influences self-rated health by testing several theoretically grounded explanations of the influence of making IP. Using the quasi-experimental instrumental variable technique increases the likelihood that our findings are not the result of reverse causality, omitted variable problem and measurement error. Our main finding is that overall, making informal payments have a negative influence on self-rated health. However, this influence is higher for men, those who are poorer, live in rural areas, have a university education and have lower social capital. Theoretical approaches that have stood out in explanations regarding the effects of making IP on self-rated health are Public Choice Theory, Institutional Theory, and Sociological Theories of Differences in Life Opportunities, Social Determinants of Health and Social Capital.

Introduction

The aim of this paper is to examine the causal association between making unofficial out-of-pocket informal payments (IP) in healthcare and self-rated health. IP is defined as a cash or in-kind payment that is over and above any formally required payments, and that is made by users of the healthcare system to healthcare personnel (Cherecheş et al., 2013; Gaal, 2006). In addition, IP may be made instead of formal payments and even where there is no formal payment¹. IP is recognized as one of the main obstacles to achieving universal healthcare coverage (Vian, 2020). IP exists in many regions of the world, with between 2% and 80% of healthcare users having reported making IP (Khodamoradi et al., 2017). For instance, IP represents the greatest portion of out-of-pocket payments made in the healthcare sector in post-communist countries of the Balkans (Mejsner and Karlsson, 2017). The previous studies have found that the pressure to make IP reduces healthcare use among low-income users who (a) do not seek, or make delays in seeking care (Atanasova et al., 2012), (b) pursue opportunities for consultations with less specialized healthcare personnel, for instance, nurses instead of doctors (Habibov, 2009a,b) or (c) are forced to seek treatment in less specialized facilities (Habibov, 2010). To conserve journal’s valuable space, further discussion of IP across various social and demographic groups can be found in Supplementary Electronic Appendix A.

Surprisingly, to date, only cursory attention has been given to ways in which self-ranked health varies as a result of making IP. As far as we are aware, this is the first study that focuses on an in-depth investigation of the direct influence of making IP on self-ranked health. However, the single item ranking of self-ranked health (e.g. ‘How would you describe your overall health now?: Very Good, Good, Fair, Poor, or Very poor?’) is a strong predictor of morbidity and mortality (Idler and Benyamini, 1997; Idler et al., 2004; Knauper and Turner, 2003). Self-ranked health has been dependably demonstrated to be a reliable predictor of mortality and commonly exceeds the predictive power of other, more ‘objective’ medical factors (DeSalvo et al., 2006). Furthermore, the validity of this measure in predicting a variety of dimensions of health has been one of the most reliably reproduced findings in social epidemiology (Lindström, 2018). Not only is ‘poor’ health strongly associated with a higher risk of mortality than is ‘excellent’ health, but the chances of mortality display a gradual increase with each incremental worsening in self-rated health (Falconer and Quesnel-Vallee, 2017; Pinillos-Franco and Kawachi, 2018). As a result, this variable has been recommended to be one of the key indicators for international comparative health research and is being used more and more for cross-national comparative purposes (Kino et al., 2020).

With the above-described gap in the literature in mind, this study attempts to address the following questions:

1. Is making informal payments negatively or positively associated with self-rated health status?

2. Is this association similar in strength for women and men, and individuals with higher education and without it?

3. Is this association similar in strength for city dwellers and those who reside in rural areas, as well as for those with more and less social capital?

The innovative contribution of this study is fourfold. First, in contrast to previous studies, which have focused on proximal outcomes such as access and the use of healthcare, this study establishes and quantifies the negative influences on the ultimate outcome, population self-rated health. Second, from a more hypothetical standpoint, this study develops and tests several theoretically grounded explanations of the relationship between making IP and population self-rated health. It also tests theoretically grounded explanations for moderating the effects of gender, education, place of residence, the wealth gap and social capital. Fourth, from a research standpoint, this study uses a quasi-experimental instrumental variable technique to establish a causal association between making IP and self-rated health. Using this technique addresses the possible threat of reverse causality since health may also influence the likelihood of making IP.

Finally, this study focuses on post-communist countries since IP is widespread in these countries while, at the same time, health outcomes are worse (Vian, 2020). Supplementary Electronic Appendix B further discusses IP development and current rates of IP in post-communist countries.

Theoretical framework and hypotheses

One approach used to evaluate the relationship between making informal payments and self-rated health is through Public Choice Theory. According to this theory, bribe-seeking administrators target better off users because they believe that they have more money to pay towards bribes than do poorer users (Peiffer and Rose, 2018). This is particularly the case when so-called ‘grease-the-wheels’ payments (Habibov, 2016; Habibov et al., 2017a) are initiated by users to get faster service and a higher quality of care. In addition, even when IP is initiated by healthcare personnel, healthcare personnel may be more likely to solicit informal payments from better off users (Belli et al., 2004). Finally, some consider IP to be part of a customary tradition of indebtedness and reciprocity that leads to higher self-esteem and feelings of satisfaction for those who make IP (Morris and Polese, 2016; Habibov et al., 2021).

The competing view is that making IP has a negative effect on population self-rated health status for several reasons. First, individuals who make IP may consider themselves to be victims of crime, and thus believing that they had to make IP may result in feelings of guilt, loss of self-esteem, psychological trauma and the fear of future victimization (Shapland and Hall, 2007). Even in cases where users make IP to receive a faster and a higher quality of care, they may be dissatisfied with the services they receive and thus may experience negative emotions (Mejsner and Karlsson, 2017). Second, making IP can result in the impoverishment of households and lead to feelings of financial insecurity (Séne and Cissé, 2015). Finally, making IP can create deep feelings of injustice since with IP, individuals have to pay for services that they were entitled to receive for free (Cherecheş et al., 2013; Habibov et al., 2019a). Hence, two opposite hypotheses could be articulated:

H1: There is a positive causal association between making IP payments and self-rated health status.

H2: There is a negative causal association between making IP payments and self-rated health status.

Furthermore, Public Choice Theory posits that when public services are not part of a monopoly, users have the choice to consume public healthcare services and pay any associated informal payments or opt out of public healthcare in favour of private providers (Klitgaard, 1988). However, if private alternatives for public healthcare do not exist or are limited, then users have to continue with the public system and make IP. Individuals in rural areas may have less capacity to step out of public healthcare and find a private provider as compared with individuals residing in urban areas (Habibov, 2011; Justesen and Bjørnskov, 2014). Hence, we articulate the following complementary hypothesis:

H3: There is a higher magnitude of causal association between making IP and self-rated health status in rural areas than in urban areas.

The relationship between making IP and population self-rated health status may be moderated by educational attainment. On the one hand, Institutional Theory postulates that individuals with higher levels of education are less likely to be involved in making informal payments since they are more affluent and have a better understanding of how public services, such as healthcare, function (Peiffer and Rose, 2013). The reason is that corruption is often institutionalized (Zaloznaya, 2012). Institutional limitations, such as ambiguity about regulation (e.g. what are legitimate and non-legitimate IP), ambiguity about accountability (e.g. what should be the sanctions and specific penalties against making IP), ambiguity about enforcement (e.g. how to initiate investigations for extorted IP) and lack of citizen voice (e.g. how to advocate for changes in policies and regulations to eliminate or minimize IP) are important precursors of making IP which are likely to vary with the level of education (Morris and Polese, 2016; Zaloznaya, 2012; Derkyi-Kwarteng et al., 2021; Pourtaleb et al., 2020). Indeed, individuals with higher education are more likely to be able to differentiate between legitimate and non-legitimate IP and to understand accountability and enforcement issues and therefore less likely to make IP (Mokhtari and Ashtari, 2012; Pourtaleb et al., 2020). Individuals with higher education are also more affluent and advocate for themselves and for institutional changes more effectively so they can be viewed as high risk to extort IP (Khodamoradi et al., 2017; Mejsner and Karlsson, 2017). University education is also associated with a generally better understanding of the consequences of corruption and hence people with more education may be less tolerant of informal payments (Dabbous and Dimant, 2018).

In addition, individuals with university education are likely to be healthier than those without university education (Habibov and Cheung, 2018). For instance, individuals with higher education are more likely to understand and follow a healthier lifestyle (Lindström, 2018). Similarly, irrespective of their labour market income, individuals with higher education have higher likelihood to be enrolled to job-related health and social insurance programmes, which increase their access to healthcare and medications, provide sick days and maternity leaves, and improve their social protection (Habibov et al., 2019b). Consequently, complementary hypothesis that follows is:

H4: There is a higher magnitude of causal association between making IP and self-rated health status for individuals without university education than those with it.

Informal payments and self-rated health nexus may also vary according to gender. Overall, women have better health and a longer life expectancy since they have healthier dietary habits, a lower rate of smoking, alcohol consumption and a higher rate of physical activity than men (Habibov and Afandi, 2011; Lindström, 2018). At the same time, evidence in healthcare literature suggests that women are less likely to get involved in making IP (Mokhtari and Ashtari, 2012; Vian and Burak, 2006). Studies, which assess making IP in public service sectors, including public healthcare, concur with the conclusion that women are less likely to be involved in making IP as compared with men (Justesen and Bjørnskov, 2014; Swamy et al., 2001).

The opposite stream of the literature suggests that women may have higher propensity to make IP (Kankeu et al., 2016; Pourtaleb et al., 2020). Yet, another stream of the literature on making IP in healthcare and public service sectors suggest no significant association between gender and propensity to make IP (Habibov, 2016; Peiffer and Rose, 2018). Therefore, the next complementary hypotheses are as follows:

H5: There is a higher magnitude of causal association between making IP and self-rated health status for men than for women.

H6: There is a lower magnitude of causal association between making IP and self-rated health status for men than for women.

In addition, the influence of making IP on self-rated health likely differs between those who are poorer and those who are wealthier. Sociological theories of Differences in Life Opportunities suggest that poorer individuals are perceived to be relatively easier targets for extorting IP by public bureaucrats as compared with wealthier ones (Weber, 1948; Peiffer and Rose, 2014). Thus, the poorer are more likely to make IP because they are easier target to public bureaucrats. At the same time, the Social Determinants of Health perspective suggests that poorer individuals often have worse health than do wealthier ones (Kelley et al., 2007). Therefore, the poorer individuals are more likely to use healthcare because of their poor health and therefore more likely make IP. These two theories reinforce each other, suggesting the same outcome: the poorer are more likely to make IP either because the poorer are easier targets for extorting IP or because the poorer have to use healthcare more frequently. Therefore, both theories postulate a higher magnitude of causal association between making IP and self-rated health status for those who are poorer than for those who are wealthier.

The rival perspective is so-called ‘Robin Hood’ hypothesis that posits the poorer users are less likely to make IP than relatively wealthier since professionals are explicitly discriminate in favour of low-income users and therefore more likely ask the wealthier to make IP (Ensor and Savelyeva, 1998; Kankeu and Ventelou, 2016). Besides, several studies demonstrate equal distribution of IP incidents among the poorer and the wealthier (Gaal, 2006; Habibov, 2011). The results of these studies show that wealth and poverty factors are not determinants of making IP. Hence, we can articulate two rival hypotheses:

H7: There is a higher magnitude of causal association between making IP and self-rated health status for those who are poorer than for those who are wealthier.

H8: There is a higher magnitude of causal association between making IP and self-rated health status for those who are wealthier than for those who are poorer.

Finally, resources and the inequalities that they create in population self-rated health and in the likelihood of having to make IP cannot be reduced to a single dimension of education, gender, income and place of residence. Thus, Social Capital Theory postulates that higher levels of interpersonal trust are associated with more equitable access to public resources because of the restrictions that interpersonal trust puts on bureaucratic opportunism (Putnam, 1993). Higher levels of interpersonal trust also encourage the development of face-to-face contacts that serve to create stronger social bonds, which in turn support the provision of public services without the need for making IP (Mbate, 2018). Likewise, individuals with higher levels of social capital enjoy better health because interpersonal trust facilitates collective actions towards support for public healthcare (Kawachi, 2018). This positive influence of social trust is well-documented in the post-communist countries. Thus, individuals with higher levels of social trust are more likely to be willing to pay more taxes to support public healthcare than individuals with lower level of social trust (Habibov et al., 2018; 2017b). They are less likely to make IP in the healthcare sector because they utilize their reciprocal connections to get higher-quality service or prioritize access (Grødeland, 2013; Akçay, 2002). Conversely, individuals with lower social capital make IP more frequently than those with higher social capital (Vian et al., 2006; Camargo, 2017). Consequently, population health is better in the countries and communities with higher levels of social capital than in those with lower levels of social capital (d’Hombres et al., 2010; Goryakin et al., 2014; Kim et al., 2011). Similarly, individuals with higher levels of social capital enjoy better health because interpersonal trust reduces uncertainty, conflicts and stress (Kawachi, 2018).

The opposite perspective stresses the so-called dark side of the social capital by highlighting that the social capital may nurture and facilitate making IP. Borrowing funds to make IP from a trusted network of relatives, friends, neighbours and co-workers is ‘the most common response to “coping” with medical costs in low- and middle-income countries’ (McIntyre et al., 2006:862). Borrowing to make IP and cover other healthcare costs is widespread in post-communist countries, and healthcare personnel may especially target such users knowing that they have higher ability to make IP (Belli et al., 2004; Gotsadze et al., 2005). In addition, operating through a trusted network may reduce the risk that users will report IP incidents to the authorities.

Consequently, our final complementary hypotheses are as follows:

H9: There is a higher magnitude of causal association between making IP and self-rated health status for individuals with lower levels of social capital than for those with higher levels.

H10: There is a higher magnitude of causal association between making IP and self-rated health status forand self-rated health status for in individuals with higher levels of social capital than for those with lower levels.

Data and methods

Data

We test the hypotheses linking IP to health status on data from 26 post-communist countries using microdata from the 2016 Life-in-Transition Survey (henceforth the LITS). The European Bank for Reconstruction and Development and the World Bank provided financial and technical support for the Ipsos pollster company which conducted the LITS (EBRD, 2018). The LITS is a nationally representative survey that is administered by trained interviewers through face-to-face interviews with approximately 1000 respondents from among the non-institutionalized population of each participating country. Its overall response rate is 89%. Further discussion of the LITS design, sampling and implementation can be found in Supplementary Electronic Appendix C.

Outcome: self-ranked health status

In LITS, the respondents were asked to rate their health status according to a five-point scale ranging from ‘very bad’ = 1 to ‘very good’ = 6. Following previous studies, self-rated health status is treated as a continuous outcome variable (Kim et al., 2011; Habibov and Cheung, 2018).

Predictor: making IP in public healthcare

The LITS asked whether ‘an informal payment was made or gift was given to public healthcare personnel in the last 12 months for the services which should be delivered for free’. The wording of this question is identical to ones used in recent studies on informal payments in public healthcare (Habibov, 2016; Habibov and Cheung, 2017). A binomial response to this question (Yes = 1; No = 0) is used as the predictor of interest.

Covariates

We controlled for age, gender, education, social capital, political trust and rural–urban residence at the individual level, and gross domestic product (GDP) per capita, amount of public health expenditures as per the country’s share of GDP and inequality at the country level (see Supplementary Electronic Appendix D). For the purpose of this study, deciles of household expenditures were used instead of individual’s income to control for ability to pay IP (see Supplementary Electronic Appendix E). Descriptive statistics for all variables are reported in Table 1.

Analytic approach

The most straightforward approach to quantifying the influence of making IP on population self-rated health is to estimate single-stage Ordinary Least Square (OLS) in order to regress self-rated health status on making IP while controlling for the covariates at the individual and country levels. However, this approach is likely to suffer from reverse causality, omitted variables problems and measurement error. Reverse causality arises because health status may also influence the likelihood of making IP. The omitted variable problem arises because vital unobserved characteristics, such as previous medical history, may simultaneously influence both one’s self-assessment of health and their likelihood to make IP. Finally, given that health status is subjectively assessed, it may suffer from considerable amounts of measurement error. To address these problems, we use the quasi-experimental instrumental variable (IV) technique that is used to estimate causal relationships with cross-sectional data when randomization is not feasible (e.g. Kim et al., 2011; Habibov and Cheung, 2016). The IV model comprises two OLS equations. In the first-stage equation, making IP is regressed on the covariates and the instrument. In the second-stage equation, which is called the main equation, health status is regressed on the covariates and on the value of making IP that has already been estimated in the first equation. The challenge of estimating IV is that the instrument should be correlated with the predictor and should not have a direct effect on the outcome other than through the predictor (Cameron and Trivedi, 2010).

Our choice of instrument is guided by the previous literature in the field of social psychology of corruption, which has demonstrated that individuals who tolerate corruption and not willing to report it are more likely to be involved in making IP (see details in Supplementary Electronic Appendix F). In the spirit of the above-mentioned literature, we use willingness to report IP as the instrument. The LITS asked if a respondent agrees with the statement ‘If I would witness an act of corruption, I would feel personally obliged to report it’. Answers vary from Strongly disagree = 1 for those who are willing to tolerate corruption to Strongly agree = 5 for those who are not.

We use these answers as the instrument for making IP since it passes the required tests for relevance of the instrument (Cameron and Trivedi, 2010). The first test on the relevancy of the instruments is the first-stage F-statistics. If the value of the first-stage F-statistics is higher than a rule of thumb value of 10, then the instrument is relevant (Cameron and Trivedi, 2010, p. 193). In all our estimations, first-stage F-statistics is higher than 10, confirming that the instrument is relevant. The second test, the so-called test for a weak instrument, involves estimating Stock and Yogo’s critical values and comparing them with the first-stage F-statistics. If the value of the first-stage F-statistics is higher than value of the estimated Stock and Yogo’s critical values, then the instrument is relevant (Cameron and Trivedi, 2010, p. 193). In all our estimations, first-stage F-statistics is higher than Stock and Yogo’s critical values, confirming that the instrument is relevant, not weak, and is strongly associated with making IP. Similarly, the significant results of the Durbin and Wu-Hausman tests signal that the IV regression is more appropriate than is the single-stage OLS estimation.

Furthermore, there is no theoretical reason to believe that willingness to tolerate corruption influences health. Indeed, the correlation between health status and the selected instrument is very weak (r = 0.01) and provides suggestive evidence that the instrument is not likely to have any direct influence on the outcome.

Estimations

We commence with IV estimation for the whole sample in order to test Hypothesis 1 and Hypothesis 2 regarding the direction of influence between making IP and self-rated health. Next, we repeat the same estimation using single-stage OLS, which serves as a benchmark. If the direction in OLS is the same as in IV, then we can conclude that the direction of influence is not an artefact of IV. Testing Hypothesis 3 to Hypothesis 7 requires going beyond the estimation for the whole sample. As such, we conducted separate estimations for subsamples, e.g. for the urban and rural, and gender subsamples.

Results

Results of IV and OLS for the whole sample are reported in Table 2. The results of IV in Model 1 demonstrate that making IP leads to worsening self-rated health by −1.47. The results of single-stage OLS are reported in Model 2 and also demonstrate that making IP leads to worsening self-rated health, although the magnitude of the effect is lower. Therefore, the results of both regression models confirm Hypothesis 2, which postulates that making IP worsens self-rated health. As shown, the results of OLS considerably underestimate the influence of IP as compared with IV, which signals the presence of omitted variables and measurement error in OLS (Cameron and Trivedi, 2010; Habibov et al., 2019a).

Results for the urban and rural subsamples are reported in the first two columns of Table 3. In Model 3 for the urban subsample, making IP worsens self-rated health by −0.91. In comparison, Model 4 suggests that making IP in rural areas worsens self-rated health by −2.39. Hence, the results confirm Hypothesis 3, which posits a higher magnitude of causal association between making IP and self-rated health in rural areas as compared with urban areas.

The results for the subsamples of individuals with and without university education are reported in the next two columns of Table 3. The results of Model 5 suggest that making IP leads to a deterioration of self-rated health by −0.86 for individuals with university education. In contrast, the results of Model 6 indicate that making IP leads to a deterioration of self-rated health by −1.71 for individuals without university education. These results confirm Hypothesis 4, which posits a higher magnitude of causal association between making IP and health status for the individuals without university education as compared with those with it.

Models 7 and 8 show the results for the subsamples of men and women. Making IP weakens women’s self-rated health by −1.39. In contrast, making IP weakens men’s self-rated health by −1.58. These results therefore provide support for Hypothesis 5, which postulates a higher magnitude of causal association between making IP and self-rated health status for men than women.

The results for the subsamples of those who are wealthier and poorer are reported in the first two columns of Table 4. Those who are considered wealthier are those who fall in Deciles 6–10 of household expenditures. Those who are considered poorer are those who fall in Deciles 1–5. The results of Models 9 and 10 indicate that the influence of making IP on health is relatively more detrimental for those who are poorer, −2.49, than for those who are wealthier, −0.89. A comparison of the results for the wealthier and the poorer confirms Hypothesis 7 by signalling that the poorer are worse off as a result of making IP in comparison to those who are wealthier.

The results of the subsample of individuals with lower and higher levels of social capital are reported in the last two columns of Table 4. The subsample of those with lower social capital contains individuals who responded that they either have no trust at all, or some trust in strangers. The subsample of those with higher social capital contains individuals who responded that they have some or complete trust in strangers. The results of Model 11 suggest that making IP worsens self-rated health by −1.32 for individuals with higher levels of social capital. In comparison, the results of Model 12 for individuals with lower levels of social trust suggest that making IP worsens self-rated health by −1.56. The results above confirm Hypothesis 10, which posits a higher magnitude of causal association between making IP and self-rated health status for those with lower levels of social capital.

Limitations

Several limitations to this study need to be mentioned. First, although both hypothetical reasoning and empirical tests suggest that the instrument is not directly correlated with the outcome, such a direct influence cannot be completely ruled out. Second, insofar as the LITS was not specifically designed to analyse IP, the questionnaire does not allow us to discern between the types of health conditions, healthcare facilities and personnel most associated with IP, nor does it provide information about the amount of IP paid, and the frequency of making IP. Furthermore, the LITS does not collect information about IP in private healthcare. Finally, small country samples prevent us from doing a country-by-country analysis.

Discussion

In this study, we focus on addressing the critical question of the influence of making IP on health. The previous literature is inconsistent regarding its theories about the direction of the influence of IP on population self-rated health status and allows us to articulate two competing hypotheses: (1) making IP positively influences self-rated health and (2) making IP negatively influences self-rated health. We empirically test both hypotheses on a large sample of countries. The empirical results generated by this study allow us to reject the hypothesis that making IP positively influences population self-rated health. Instead, our results support the hypothesis that making IP has a negative influence on population self-rated health. Including the comprehensive set of covariates reduces the likelihood that our findings are the result of spurious correlations. The application of the quasi-experimental IV technique allows us to establish a causal association between making IP and population self-rated health status. Using IV provides us with more certainty that our findings are not the result of reverse causality, omitted variable problems and measurement error.

Our findings add value to the recent literature on unofficial out-of-pocket expenditures (Mejsner and Karlsson, 2017; Morris and Polese, 2016). In contrast to previous studies that focused on proximal outcomes such as access and utilization of healthcare, this study establishes and quantifies the influence of making IP on the ultimate outcome, population self-rated health. Our findings also contribute to the emerging theoretical debate about IP since the existing literature is not consistent regarding the outcome of IP (Habibov, 2016). This study provides a strong theoretical argument to explain why making IP has a negative influence on population self-rated health, and this hypothesis has been confirmed empirically. Finally, our findings also relate to the burgeoning literature on self-rated health (Kawachi, 2018; Kino et al., 2020). By showing that self-rated health varies by making IP, our study demonstrates the need to include making IP as one of the key predictors of self-rated health in future studies.

We also found that the combination of economic, social and institutional influences appears to moderate making IP with respect to the population health nexus. First, we did not find support for the economic argument that the burden of making IP is borne by those who are better off and who used IP payments to overcome the perceived administrative ineffectiveness of public healthcare, for instance, to get more personalized services or to jump the queue to promptly access treatments. On the contrary, the empirical evidence found in our study provides support for the Social Inequality argument. We found that those who are poorer are relatively worse off in terms of their health because of having made IP, when compared with those who are wealthier. Such results suggest a higher degree of vulnerability to exploitation through IP among those who are poorer. As Hunt succinctly explained, ‘Corruption hits people when they are down’ (Hunt, 2007, p. 574). In the milieu of healthcare reforms in post-communist countries, empowering the poor emerges as the priority goal. Fortunately, Cochrane’s systematic review, in addition to other related studies, suggests that the creation of independent complaint mechanisms, that could provide fast and reliable feedback, would be an effective method through which achieves this goal (Gaitonde et al., 2016; URC, 2017; Habibov and Cheung, 2017).

Second, we found support for the Public Choice argument that users who have a choice will more likely opt out of public healthcare in favour of using private providers. The importance of this finding should be taken seriously, as it signifies the development of a de facto two-tier healthcare system. The users who have the choice will choose to exit the public tier and move to the private tier to avoid having to make IP. Likewise, users who have no choice will have to continue to use the public healthcare tier and continue to pay IP. As a result, support for public healthcare provision, including the willingness to pay taxes to finance it, may considerably weaken, leaving it vulnerable to eventual collapse (Habibov et al., 2019b). In the context of healthcare reforms in post-communist countries, these findings suggest that funding for public healthcare should significantly increase and quality of healthcare administration should improve in order to prevent the development of a two tier-healthcare system which will create additional barriers for the poorer and disadvantaged groups of users.

Third, we found support for the Institutional Theory argument that suggests that education reduces the negative influence of making IP on health status. This finding highlights the role of a better understanding of how healthcare functions, for instance, which fees are legitimate and which are not, and suggests that lack of information is an important issue in the context of institutionalized corruption (Derkyi-Kwarteng et al., 2021; Pourtaleb et al., 2020; Mokhtari and Ashtari, 2012). Consequently, international initiatives aimed at increasing levels of community information and enhancing transparency in healthcare are important steps in the right direction (Vian, 2020). In addition, boosting prosecution of corruption and strengthening local mechanisms to counteract corruption through the development of a civil society movement and the dissemination of free anti-corruption media would provide disincentives for public administrators to participate in corrupt activities (Mbate, 2018). Thus, in the milieu of healthcare reforms in post-communist countries these findings indicate the needs for developing a systematic approach to improving good governance in public healthcare system. The first step towards improving transparency, responsibility and accountability could be joining international initiatives in healthcare reform led by the World Health Organization (Vian, 2008).

Fourth, we found that social capital reduces the negative influence of making IP on health status. This finding is in line with several strains of previous studies, namely, (1) social trust leads to more funding for public healthcare which reduces the need to make IP (Habibov et al., 2018, 2017b), (2) social trust creates stronger social bonds and individuals with higher social trust less likely make IP (Grødeland, 2013; Mbate, 2018; Vian and Burak, 2006), (3) social trust reduces uncertainty, conflicts and stress which lead to improvement in health (Kawachi, 2018), and (4) individuals in countries and communities with higher levels of social capital reported better health and overall well-being than those in countries and communists with lower levels of social capital (d’Hombres et al., 2010; Goryakin et al., 2014; Habibov et al., 2019c; Kim et al., 2011). Thus, increasing social capital provides one of the key pathways for mitigating the negative impact of making IP. In the context of healthcare reforms in post-communist countries, the interventions aimed at strengthening and maintaining a stock of social capital have the goal of increasing social interactions in the community and building up feelings of solidarity and mutual support (Habibov et al., 2019b). Importantly, examples of successful interventions that use social capital to enhance health include not only interventions in the physical locale, but also in the virtual space through the deployment of new technologies, for instance, virtual communities, internet chats and online networks (Guo et al., 2018). Hence, the healthcare reforms in post-communist countries should incorporate components explicitly aimed at developing and maintaining social capital since lack of social capital is associated with an increase in making IP.

Fifth, we found that theories of social inequality (i.e. Social Determinants of Health and Sociological theories of Difference in Life Opportunities) are particularly important in explaining the influence of making IP on health status. As such, most of the respondents in our study could be subject to a mixture of positive and negative influences simultaneously. As an illustration, one individual may be in a wealthier decile but reside in a rural area with no choice regarding their healthcare provider(s), while another may have university education but lack social capital. Given the lack of a developed statistical test, the influence of each of these characteristics vis-à-vis each other may not be significant in a strictly statistical sense. However, taken together, differences in education, place of living, gender, income and social trust are defining elements of social inequality. Social differences should therefore be seen as the foundation of inequality with respect to IP’s influence on the effect of health. In the setting of post-communist countries, these findings suggest that healthcare reforms should be guided not only by political, financial and administrative considerations. Such reforms should unequivocally incorporate social inequality perspectives. As shown above, the mixture of social factors influences the making of IP and health status nexus. It is plausible to believe that the exact mixture of such factors considerably varies across countries. Consequently, studying social factors which explain the influence of making IP on health status in every country should be the first step in designing effective and equitable public healthcare in post-communist countries.

Conclusion

Making informal payments has a negative influence on population self-rated health. This influence is higher for men, those who are poorer, and those who live in rural areas, have university education and have lower social capital. Theoretical approaches that stood out in explaining the effect of IP on self-rated health are Public Choice Theory, Institutional Theory and Sociological Theories of Differences in Life Opportunities, Social Determinants of Health and Social Capital.

Supplementary data

Supplementary data are available at Health Policy and Planning online.

Data availability statement

Data is available at www.ebrd.com.

Funding

No funding received for this work.

Ethical approval.

Ethical approval for this type of study is not required by our university.

Conflict of interest statement.

The authors declare that they have no conflict of interest.

Endnotes

References