Salutogenesis for thriving nations: Sense of coherence and longevity across income strata

Abstract

Traditional salutogenesis literature and theories in health and development have emphasized the role of sense of coherence dimensions—comprehensibility, manageability, and meaningfulness—in shaping population health. However, their relevance at a macro scale, across varying contexts, has been overlooked. What is the relationship between these dimensions and population health at the national level? Do they exert a uniform effect across countries with varying income levels? Using the sense of coherence framework dimensions at the macro-level, we argue that comprehensibility, manageability, and meaningfulness are related to national life-expectancy. We examine their relationship across various economic contexts to refine the boundaries of their links with population health and provide a nuanced understanding of their interplay. Drawing on a panel dataset of 135 countries from 2017 to 2020 from the Quality of Government institute, International Telecommunications Union, and World Bank, we test our hypotheses using a high-dimensional fixed effects model. We find support for two dimensions, manageability, and meaningfulness, in improving life expectancy. We unpack these findings to understand whether they exert a uniform effect or vary across different levels of income. Our findings show that the effectiveness of each dimension on longevity varies by income level, suggesting their relevance is context specific. Through an ecological approach, this study extends Antonovsky's theory of salutogenesis to demonstrate how structural factors at the population level contribute to maintaining health.

Highlights

• •Antonovsky's salutogenesis framework is empirically applied via a cross-country study.
• •Sense of coherence dimensions shape global population health across income levels.
• •Manageability and meaningfulness are important, but differ in their impact.
• •Polices that address manageability are linked to longevity in high-income contexts.
• •Policies that address meaningfulness are linked to longevity across income levels.

1. Introduction

Research on salutogenesis suggests managing individual health involves the development of a strong sense of coherence (Antonovsky, 1987). One way in which individuals overcome the stressors and manage tension is through the degree to which they are able to comprehend, manage, and find meaning in chaos—order in disorder (Antonovsky, 1987). While existing research acknowledges that context is important to understand population health and the develop of a strong sense of coherence, less understood is how this process is shaped nationally across different economies and income levels (Mittelmark et al., 2017). Thus, in this paper, we focus on the mechanisms and boundaries of the salutogenesis model of health, and more specifically the components of sense of coherence, in which we argue are contingent upon economic context.

We begin by reviewing the literature on the salutogenesis model of health, its components, and mechanisms. We outline the components for sense of coherence and their importance as they relate to poverty and marginality. We turn our focus to the mechanisms of developing a sense of coherence to improve population health across multiple levels of income, and overcome the challenges posed by each stratum. We explore each component of sense of coherence, comprehensibility, manageability, and meaningfulness, while accounting for generalized resistant resources. We argue that the benefits that comprehensibility, manageability, and meaningfulness offer are useful to absorb stressors associated with health, and in the process, endogenize constraints that may seem exogenous, such as longevity.

We then extend this argument by suggesting that the relative effectiveness of the dimensions sense of coherence in improving population health depends on the extent to which these dimensions are useful across different income levels. While comprehensibility, manageability, and meaningfulness can emerge to be effective across all income strata, it may be that one is more specific to managing stressors at a certain income level, and therefore more beneficial to population health in this stratum. By exploring differences in income levels on population health and juxtaposing the links between the multiple components of sense of coherence and longevity, our paper provides a more complete picture of how stressors are absorbed.

To test our hypotheses, we use a panel of 150 countries, covering high-, upper-middle-, lower-middle-, and lower-income countries, from 2017 to 2020. Our aim is to answer the following research question: What is the relationship between national-level sense of coherence and population health? To what extent does a country's income shape the relationship between sense of coherence and population health? We find that the relationship between manageability and meaningfulness with life expectancy does in fact vary by income level, suggesting their relevance is context-specific. While manageability is positively related to longevity more in upper-middle income economies as compared to highest income economies, meaningfulness is positively associated with longevity more so in lower-income, and lower-middle-, and upper-middle- income economies than in high-income economies. We do not find evidence that comprehensibility is related to the effect of longevity in any context. We conduct various robustness tests, through multiple model specifications to support our findings.

While sense of coherence has been widely studied at the individual level, its application at the national level remains unexplored. Individual level studies have examined the effects on health outcomes, while more recent research has extended sense of coherence to the organizational level to explore shared norms and coping capacities. However, national-level sense of coherence remains less examined. This study addresses this gap by extending the discussion on the salutogenesis model of health through documenting its specific mechanisms at the macro level— endogenizing constraints that may appear exogenous to population health. In doing so, we demonstrate how the dimensions of sense of coherence operate at a higher scale and interact with the broader socioeconomic and institutional environment, to provide a more systematic understanding of how these constructs collectively influence population health, and answer calls to scale the application of the framework, not only to organizations and cities—but to nations (Mittelmark et al., 2017). By bridging these levels, we contribute to the theoretical evolution of the framework from an individual resource to a broader population-wide resilience framework.

In addressing this gap, we also contribute to the understanding of how income classification shapes the relative importance of these dimensions by establishing boundaries for their influence based on economic context. We find that manageability is more critical in higher-income settings, where infrastructure is well established and resources are abundant, while meaningfulness is more important in lower-income settings, where resources are more constrained, and it may be more necessary to find coping mechanisms for immediate survival. By defining these boundaries, we contribute to a finer grained and more nuanced perspective on how comprehensibility, manageability, and meaningfulness interact with contextual factors. This delineation challenges the assumptions that dimensions have a universal effect and highlights the importance of a tailored approached based on the underlying economic environment.

Increased worldwide interconnectedness means that global phenomena—such as climate change, pandemic, pollution, and resource depletion—transcend national boundaries and affect diverse populations in distinct ways (Alkharafi & Alsabah, 2025). They may expose vulnerable populations to shocks that wealthier populations can buffer against more easily, amplifying the need for a context-sensitive understanding. Globally, studies at the macro-level have demonstrated a relationship between the environment, resources, and population health. Recent studies show the relationship of environmental stressors, such as chronic exposure to particulate matter, with public health and longevity (Alkharafi, 2025; Nazarenko et al., 2020. Wilkinson and Pickett (2020) demonstrate that countries with greater income inequality experience higher rates of health related and social problems, including obesity, mental health illness, and reduced life expectancy. Barnes et al. (2023) argue that different types of political economies and social structures shape health outcomes, while Naik et al. (2019) suggest employment and improved working conditions reduce gender-based health inequalities. Even further, Ngepah & Mouteyica (2024) find that policies that encourage trade and ICT diffusion can impact regional disparities in health outcome. We extend research in global health by applying an ecological approach to examine the salutogenic health model at the national level, highlighting how macro-level factors influence population well-being.

2. Theoretical framework

2.1. Salutogenesis, sense of coherence, and generalized resistant resources

Salutogenesis, a term coined by medical sociologist Aaron Antonovsky, was created in the 1970's to capture a new conceptualization of health and well-being. The concept was created to capture the origins (genesis) of health (saluto) and focus specifically on the study health, rather than traditional mainstream medical literature which centers around pathogenesis, or the treatment of disease (Antonovsky, 1979, Antonovsky, 1987). Antonovsky questioned whether well-being can fully be understood and reach its potential without a solid foundation in a theory of health. Thus, rather than follow the standard medical literature which primarily focused on explaining pathology, Antonovsky developed a new paradigm and centering his concept on health to examine the factors that contribute to individuals' health and the strategies they employ to maintain it (Antonovsky, 1979).

To answer the main question underlying the salutogenic health model, what makes people healthy, Antonovsky introduces its core concept, the sense of coherence. Sense of coherence is defined as “a global orientation that expresses the extent to which one has a pervasive, enduring though dynamic feeling of confidence that (1) the stimuli deriving from one's internal and external environments in the course of living are structured, predictable, and explicable; (2) the resources are available to one to meet the demands posed by these stimuli; and (3) these demands are challenges, worthy of investment and engagement” (Antonovsky, 1987, p. 19).

The sense of coherence focuses on how individuals can create order out of chaos and enhance their adaptation capabilities to an environment in which stressors are inevitable. A strong sense of coherence allows individuals to organize resources productively in an effort to manage tension, cope with stressors, and ultimately re-orient themselves from pathogenic to salutogenic states. Stressors may not necessarily lead to stress, illness and decline—although they create immediate tension (Mittelmark et al., 2017). If resolved, the negative effects on heath can be avoided or minimized. Thus, coping and tension management are key intervening components between stressors and illness that can bring about balance to a disequilibrium in wellbeing.

Sense of coherence consists of three components: comprehensibility, manageability, and meaningfulness (Antonovsky, 1990). Comprehensibility is the extent to which an individual has a clear understanding of the problem and can make sense of their internal and external experiences. Access to education, reliable information, and communication systems enhance comprehensibility (Mittelmark et al., 2017). Manageability describes the extent to which individuals perceive that they have the adequate resources necessary to cope with the stressors and whether those resources are understood to be “in one's own hands or in the hands of legitimate others” (Antonovsky, 1990, p. 79). Meaningfulness refers to the extent to which individuals perceive life as worth living and maintain the will and motivation to channel stressors productively. Antonovsky (1990) refers to meaningfulness as the will to “seek to order the world and to transform resources from potential to actuality” (p. 79).

Antonovsky (1979) focused on the intersection of health, poverty, social class, and marginality, specifically examining the stressors of lower income and minority groups (Antonovsky, 1979, Antonovsky, 1990). He argued that socio-economic class impacts the number of resources individuals have to overcome stressors, claiming that lower income groups were “the class which clearly had the highest stress load … the constancy of imposed stressors in such life situations, the continuous emergencies life presents, make it immensely difficult to resolve tension” (Antonovsky, 1990, pp. 73-74). In a situation where higher socioeconomic groups are exposed to a similar stressor; the stressor may not necessarily lead to illness as a result of more resources for successful coping and adaptation. Thus, how serious stressors interacted with income and class, especially amongst the most vulnerable and marginalized became a focal point of his research. Table 1 presents conceptualizations and expressions of sense of coherence.

Table 1.

Sense of coherence: conceptualizations and expressions. The three core components of Antonovsky's sense of coherence framework are summarized along with operational expressions.

	Conceptualization	Expressions
Comprehensibility	Understanding and clarity of the environment and challenges within to cope	- Access to information
		- Literacy
Manageability	Internal or external resources that can be drawn upon to cope	- Economic conditions/policies
		- Infrastructure
Meaningfulness	Feelings of purpose, motivation, and willingness to engage	- Life satisfaction
		- Mental well-being

Source: Antonovsky (1990); Mittelmark et al. (2017).

Generalized resistant resources are those resources available to the individual, group, or community that facilitate individuals' ability to cope with stressors and contribute to developing or amplifying a strong sense of coherence (Antonovsky, 1987; Mittelmark et al., 2017). Antonovsky defines generalized resistant resources as “any characteristic of the person, group, or environment that can facilitate effective tension management” (Antonovsky, 1972, p. 99). These resources may include a variety of elements, including material resources such as income, social resources such as social support, social networks, and cohesion with one's roots, among a range of others (Mittelmark et al., 2017). Generalized resistant resources are enablers of a strong sense of coherence, and support individuals in perceiving life events as comprehensible, manageable, and meaningful. They can be utilized to overcome stress, enhance coping and maintain well-being (Lindström & Eriksson, 2005). Fig. 1 presents the conceptual framework depicting the salutogenesis model of health and the hypothesized relationships among its key components.

Fig. 1.

The Salutogenesis Model of Health.

A conceptual framework depicting the salutogenesis model of health and the hypothesized relationships between its key components, with income level moderating these relationships.

2.2. Extensions to group, firm, and city-level

In subsequent conceptualizations of the salutogenesis model of health (SMH), Antonovsky (1987) extends the model to a higher level and argues for a group sense of coherence, or a sense of group consciousness. He identifies a group maintaining a strong sense of coherence as “a group whose individual members tend to perceive the collectivity as one that views the world as comprehensible, manageable, and meaningful, and among whom there is a high degree of consensus in these perceptions” (Antonovsky, 1987, p.174). Antonovsky suggests these groups have a stronger ability to structure situations and over time can improve the overall sense of coherence of the other individual members within the group.

Jenny and Bauer (2013) apply the SMH to the context of organizations and examine sense of coherence in firms through the perceived comprehensibility, manageability, and meaningfulness of individuals work situation (Vogt et al., 2013). Through salutogenesis, they examine the promotion of health through a meaningful work-life balance in a professional working environment. They develop the organizational health development model (Bauer & Jenny, 2012; Jenny & Bauer, 2013) which operates at the individual and collective level to develop healthier, more resilient organizations. Mittelmark et al. (2017) call for the application of salutogenesis in cities and towns. They suggest that cities can facilitate well-being through urban planning, design, and collaboration across sectors. A growing body of research examines the relationship between developing a strong sense of coherence and context dependent resources (Bull et al., 2013; Maass et al., 2014), demonstrating the ways in which infrastructure, transportation, green spaces, recreational activity, and inclusive planning at the local level can lead to positive health outcomes (Mittelmark et al., 2017).

3. Hypotheses development

3.1. Sense of coherence and longevity

3.1.1. Comprehensibility

Comprehensibility refers to the ability to understand and make sense of internal and external experiences in the environment (Antonovsky, 1990). At the national level, access to internet serves as a structural enabler of comprehensibility in society through empowering populations with the tools to access information, improve communication, and engage in digital services in an instant. It reflects the extent to which populations have the infrastructural means to access and interpret information. Greater internet penetration increases exposure to health education and digital public services, contributing to a more intelligible environment. The internet serves as a key enabler for accessing health information, resources, and services that contribute to improving population health. This relationship is documented in the literature through empirical studies that link internet penetration with improved health metrics across countries (Liao & Luo, 2024; Liu et al., 2024; Soundararajan et al., 2023; Stoumpos et al., 2023; Yu & Meng, 2022).

Access to internet helps to raise comprehensibility by breaking down informational barriers. These tools empower society to acquire knowledge about healthcare, nutrition, and lifestyle modifications—enabling them to make informed decisions that promote well-being (Di Simone et al., 2024; Finlay-Jones et al., 2023; Jones et al., 2015). Having access to reliable information related to health management and intervention can promote healthier life choices. In economies with higher internet penetration, populations are more equipped to navigate complex healthcare systems and access services such as electronic health record and telemedicine, all of which contribute to improved health (Finlay-Jones et al., 2023; Lee et al., 2024).

Digital health innovations, such as mobile health applications and wearable devices play a key role in improving comprehensibility by unpacking and monitoring health outcomes (Finlay-Jones et al., 2023; Lee et al., 2024). Such tools provide feedback and recommendations aid in understanding ones physiological state by monitoring physical activity and breaking it down to the user. This process not only improves self-efficacy, but also reduces uncertainties about health, reinforcing the perception of life as manageable and understandable. Thus, access to the internet brings the gap between individuals and the complexities of health management, making health information more accessible, understandable, and actionable.

Moreover, internet access may improve population health by facilitating social connections. Social connections are fundamental to mental and emotional well-being. Positive social interactions contribute to higher self-esteem (Harris & Orth, 2020), greater empathy (Melloni et al., 2014), better cognitive functions (Kelly et al., 2017), and a longer life (Friedman et al., 2024; Holt-Lunstad et al., 2010). Internet access allows populations to connect with others more easily in shorter periods of time and through longer distances. It offers opportunities for engaging in virtual communities, reducing loneliness and stress, which have been shown to be linked to cardiovascular impairment (Holt-Lunstad et al., 2010). Therefore, comprehensibility through internet access not only provides access to information and improves cognitive functions, but also emotional and social resilience, which are key determinants of health and longevity. Thus, we hypothesize.

H1a

Countries with higher rates of internet access are associated with higher life expectancy.

3.1.2. Manageability

Manageability refers to the extent that individuals recognize they have adequate resources to cope (Antonovsky, 1990). Manageability is reflected through stable economic policies that address inequality, infrastructure that supports individuals access to services, and a robust industrial sector that contributes to economic growth, employment opportunities, innovation, and financial security. Industrial development and growth are a cornerstone for the structural foundation of an economy and its ability to generate wealth and provide resources that support its citizens (Facevicova and Kynclova, 2020; Hosono, 2022; Kuznets, 1968; Solow, 1956). By shaping the material and systematic resources available to populations, industrial growth can influence individuals’ resources and opportunities (Luca, 2002), and provide them with the tools necessary to obtain assistance to address challenges effectively (Antonovsky, 1990), ultimately influencing population health.

A strong industrial sector creates large-scale opportunities for employment and development in manufacturing, production, services, and other sectors (UNIDO, 2024). These offer individuals a stable source of income and a means to access basic necessities such as housing and healthcare, all of which are critical for managing life demands. Opportunities to earn a stable income reduce uncertainty and increase confidence in individuals to plan for the future, handle unexpected challenges, and invest in their long-term well-being (ILO, 2020). Opportunities generated by industry play a key role in lifting individuals out of poverty, particularly in developing economies where the reliance is primarily on agriculture (ILO, 2020; UNIDO, 2024). Industrial growth serves as a primary driver of economic development, not only to reduce poverty, but improve access to education and healthcare (Bloom & Canning, 2000). In this way, industrial opportunities provide individuals with the financial resources to invest in better living conditions and healthier lifestyles, reinforcing their sense of manageability.

Furthermore, industrial employment drives social mobility, enabling individuals to improve their social and economic standing (Falcon & Joye, 2021; Kerr et al., 1960; Treiman, 1970). In regions where industries underly economic development, workers may be eligible to receive training or educational opportunities that allow them to refine or develop new skills (Falcon & Joye, 2021). These initiatives enhance individuals’ expertise and ability, allowing them to be more adaptive to economic changes while reducing their vulnerability to unemployment (ILO, 2020). Additional training and expertise can facilitate access to improved opportunities and provide a sense of empowerment, reinforcing individuals sense of manageability.

Even further, secure and meaningful employment contribute to a sense of self-worth which is essential for well-being. In this way, industrial growth is likely to offer resources and the belief in one's ability to cope with stressors. Populations in economies experiencing industrial growth are more likely to experience economic stability and a sense of security (Rodrik, 2009; Ross, 2012). When the external environment offers resources and opportunities for growth, and supports communities in achieving stability and security, individuals are more likely to effectively manage stressors and attain higher standards of population health. Thus, we hypothesize.

H1b

Countries with higher industrial value are associated with higher life expectancy.

3.1.3. Meaningfulness

Meaningfulness refers to the degree that individuals perceive life as purposeful, valuable, and worth living (Antonovsky, 1990). Meaningfulness is reflected through national subjective well-being. Subjective well-being captures the collective value that individuals attach to their lives and society based on their perceptions of life-satisfaction and purpose, reflecting a populations sense of meaning and fulfillment. Through multiple elements, including life evaluation, affect, and eudaemonia, subjective well-being reflects the way in which populations perceive and experience their quality of life. Research shows subjective well-being is linked to longevity through psychological resilience, stress reduction, and health promoting behaviors (Diener et al., 2018; Ran et al., 2020; Steptoe et al., 2015).

Subjective well-being is essential for fostering psychological well-being (Ryff & Singer, 2008) and resilience in individuals, enabling them to better control stress responses, recover from adversity, and minimize its physiological impact (Friedman et al., 2007). Psychological resilience equips individuals with the ability to better regulate emotional responses, maintain optimism, and engage in constructive coping strategies when faced with challenges (Ryff & Singer, 2008). This includes seeking social support, problem solving, and reframing challenges in a positive light. Individuals with higher psychological resilience are less likely to develop anxiety or depression, both of which exacerbate health risks and reduce longevity (Mathew & Paulose, 2011; Poole et al., 2017; Ran et al., 2020; Steptoe et al., 2015).

When individuals experience positive emotions towards their life, such as joy, gratitude, and hope, they activate biological processes that counteract stress, such as decreased cortisol levels and improved immune function (Mathew & Paulose, 2011; Ong, 2010; Steptoe et al., 2015). These effects reduce the risk of developing illnesses such as cardiovascular disease (Boehm et al., 2020), which has been shown to be a major contributor to premature mortality (Read & Wild, 2020). In contrast, negative affect exacerbates stress responses and increases the risk of chronic illnesses (O'Connor et al., 2021). Prolonged experiences of negative emotions lead to higher level of stress hormones and can impair immune function and increase susceptibility to illness (O’Connor et al., 2021).

Individuals who perceive their life positively are more likely to feel a sense of purpose, value life, and engage in long-term behaviors that sustain health (Kushlev et al., 2020). When individuals are content with their lives, they experience less internal conflict and stress, reducing the likelihood of adopting harmful habits, such as smoking (Martín-María et al., 2020). Life satisfaction and positive affect have been shown to be predictors of healthy behavior (Kushlev et al., 2020) and life expectancy (Diener et al., 2018). This reflects a fundamental Salutogenic principle, that meaningfulness equips individuals to cope effectively with stressors (Antonovsky, 1990; Ryff & Singer, 2008) and is liked to their longevity (Diener & Chan, 2011; Diener et al., 2018). Thus, we hypothesize.

H1c

Countries with higher levels of subjective well-being are associated with higher life expectancy.

3.2. Generalized resistant resources

Generalized resistant resources refer to resources that facilitate individuals’ ability to cope with stressors and contribute to developing or amplifying a strong sense of coherence (Antonovsky, 1987; Mittelmark et al., 2017). The role of economies size in population health is foundational as it reflects the overall economic capacity and systemic strength to support comprehensibility, manageability, and meaningfulness and their effects on population health. Larger economies have more resource to support access to information, industry development, and subjective well-being (Deaton, 2008; Hamilton et al., 2006). They are more capable of reinforcing digital infrastructure, industrial growth and quality of life. Financial resources enable the expansion of industries and diversification, as well as the investments in social safety nets, cultural programs, mental health initiatives, and economic opportunities that foster a populations sense of satisfaction and well-being (Deaton, 2008; Hamilton et al., 2006). Thus, economy size serves as a systemic enabler and buffer for its population, by ensuring that these components operate effectively, and by providing the necessary resources for supporting population health (Lago-Peñas et al., 2013).

The role of regulatory quality in shaping population health is equally critical (Dunbar et al., 2023; Kruk, Gage, Arsenault, et al., 2018; Kruk, Gage, Joseph, et al., 2018; Wahl et al., 2020). Similarly, regulatory quality reflects the systemic strength to support comprehensibility, manageability, and meaningfulness and ultimately population health. Economies that maintain higher regulatory quality are more likely to have a supportive environment business for innovation and industry to thrive (Aghion et al., 2023). Regulatory quality underpins policies that promote digital literacy and ensure affordability and equity in digital infrastructure and services (OECD, 2023; Vassilakopoulou & Hustad, 2023). High regulatory quality enhances well-being by promoting more transparent and fair regulation, and trust in institutions. Regulations that protect the rights of individuals, ensure access to resources, and address social inequalities contribute to a greater sense of well-being (Dunbar et al., 2023; Kruk, Gage, Arsenault, et al., 2018; Kruk, Gage, Joseph, et al., 2018; Wahl et al., 2020). Thus, we hypothesize.

H2a

Larger economies (GDP) are associated with higher life expectancy.

H2b

Higher regulatory quality is associated with higher life expectancy.

4. Methodology

4.1. Data and sample

This study draws on country-level data from various sources. The dependent variable, life expectancy at birth, and a key independent variable, subjective well-being, were collected from the Quality of Government institute. The Quality of Government institute is an independent research institution in the Political Science Department at the University of Gothenburg aiming to promote research on the nature of quality governance that is impartial from government institutions. This was matched with data from the International Telecommunications Union (ITU) and World Bank national accounts for our other independent variables, namely access to internet, and industry value. Other variables, including regulatory quality and gross domestic product, were collected from the World Governance Indicators and the World Bank Development Indicators database, respectively. This resulted in a final model sample of 135 countries for four years from 2017 to 2020 with 501 observations across all regions and income levels.

4.2. Variables

Dependent variable. To measure health, this study uses life expectancy at birth (Antonovsky, 1979, Antonovsky, 1987; Eriksson & Lindström, 2007; Lindström & Eriksson, 2006), which represents the total number of years an individual is expected to live from birth based on mortality rates within a given population (Teorell et al., 2023).

Independent variables. To capture comprehensibility, this study utilizes individuals using the internet as a percentage of the population, reflecting the ability to obtain and engage with information to make sense of the environment (Antonovsky, 1987; 1990; Mittelmark et al., 2017). This measure represents individuals who have used the internet from any location in the last three months through a computer, mobile phone, or other devices (ITU, 2023). The internet provides access to a vast repository of information and serves as a tool for acquiring knowledge, enabling individuals to be more equipped to respond to stressors in the environment.

To capture manageability, this study employs industry value added, which represents the net output of various sector outputs—minus the immediate inputs for the respective sectors—as a percentage of national gross domestic product (World Bank, 2023). The reflects the capacity of industry to provide jobs and basic resources that support its citizens, such as electricity and water. Industrial growth shapes the material and systemic resources available to a population. Populations in economies experiencing industrial growth are more likely to have opportunities for work, access to critical basic services, as well as experience stability and security (Rodrik, 2009; Ross, 2012).

To capture meaningfulness, this study applies subjective well-being, which represents national averages for subjective well-being, measured by answers to the Cantril ladder question (Cantril, 1965) where individuals evaluate the quality of their life based on a scale of 0, indicating not satisfied at all, to 10, indicating completely satisfied (Teorell et al., 2023).

To capture generalized resistant resources, this study utilizes gross domestic product as a measure of economy size, which represents gross domestic product expressed in current international dollars converted by purchasing power parity. Additionally, this study uses regulatory quality, which refers to the ability of government to create and apply policies and regulations that advance development (Kaufmann et al., 2006). Generalized resistant resources are resources available to the individual, group, or community that facilitate individuals’ ability to cope with stressors (Antonovsky, 1987; Mittelmark et al., 2017), and may include a variety of elements, including opportunities for income, social structure and support, amongst others (Mittelmark et al., 2017). Economy size and governance quality are essential resources and characteristics of an environment that facilitate effective tension management.

To capture Antonovsky, 1979, Antonovsky, 1990 emphasis on the intersection of health, poverty, social class, and marginality, this study employs income level (Ignatow & Gutin, 2024), a classification that divides nations into four groups—low, lower-middle, upper-middle, and high income—according to gross national income per capita (World Bank, 2022). According to the World Bank analytic classifications in 2022, countries with GNI per capita in US dollar of $1135 or less are classified as low income, $1135-$4465 as lower-middle income, $4465-$13,845 as upper-middle income, and above $13,845 as high income. Fig. 2, Fig. 3 present life expectancy and income level around the world.

Fig. 2.

Life expectancy mean around the world (2017–2020).

Source: created by Stata thematic mapping using Quality of Governance Institute data.

Fig. 3.

Income classification around the world.

Source: created by Stata thematic mapping using World Bank data.

4.3. Empirical analysis

To estimate the potential relationship, this study uses a high-dimensional fixed effects model (Correia, 2017). Hypotheses are tested through a strongly balanced panel comprising of 135 countries across multiple levels of income for four years from 2017 to 2020. We apply a high-dimensional fixed effects regression model to control for multiple levels of fixed effects and account for unobserved heterogeneity across time and countries. In doing so, we reduce omitted variable bias, and ensure more efficient, accurate and robust estimates that isolate the unique contribution of the explanatory variables. The incorporation of both country and year fixed effects control for time-invariant characteristics of countries, such as geography and institutional structures, and year-specific shocks such as a pandemic or global financial crises, to isolate the effects. Standard errors are clustered to account for autocorrelation and heteroskedasticity. We estimate our model through multiple specifications, beginning with a simple ordinary least squares (OLS) regression. By comparing multiple model estimates, we demonstrate the robustness of our results under different specifications.

5. Results

We test the relationship in the first fixed-effects regression (Model 3) through the following equation:

$$Y_{it}={\beta }_0+{\beta }_1{X}_{1it}+{\beta }_2{X}_{2it}+{\beta }_3{X}_{3it}+{\beta }_4{X}_{4it}+{\beta }_5{X}_{5it}+{\alpha }_i+{\mu }_t+u_{it}$$ (1)

where the dependent variable Y_it is life expectancy for country i in year t. The first three independent variables of interest, X_1it, X_2it, X_3it denote access to internet, industry, and subjective well-being respectively, for country i in year t. The model also includes the intercept β₀, country fixed effects α_i, year fixed effects μ_t, and other variables including X_4it, and X_5it, to indicate market scale through gross domestic product and regulatory quality.

The second fixed-effects regression (Model 4) tests the relationship with an interaction term for industry:

$$Y_{it}={\beta }_0+{\beta }_1{X}_{1it}+{\beta }_2{X}_{2it}+{\beta }_3\left(X_{2it}\times D_i\right)+{\beta }_4{X}_{3it}+{\beta }_5{X}_{4it}+{\beta }_6{X}_{5it}+{\alpha }_i+{\mu }_t+u_{it}$$ (2)

where the term (X_2it × D_i) refers to the interaction between industry and the dummy variable D_i, which represents level of income—including high, upper-middle, lower-middle, and low income. The interaction is included to investigate the differential effects between industry and life expectancy based on the level of income.

The third fixed-effects regression (Model 5) tests the relationship with an interaction term for subjective well-being:

$$Y_{it}={\beta }_0+{\beta }_1{X}_{1it}+{\beta }_2{X}_{2it}+{\beta }_3{X}_{3it}+{\beta }_4\left(X_{3it}\times D_i\right)+{\beta }_5{X}_{4it}+{\beta }_6{X}_{5it}+{\alpha }_i+{\mu }_t+u_{it}$$ (3)

where the term (X_3it × D_i) refers to the interaction between subjective well-being and the dummy variable D_i, which represents level of income. Table 2, Table 3 present the summary statistics and regression estimates.

Table 2.

Summary statistics. The number of observations, mean, standard deviation, minimum, maximum and mean across the various income levels are reported providing an overview of the data distribution.

Variables						High	Upper-mid	Lower-mid	Low
	N	Mean	SD	Min	Max	Mean	Mean	Mean	Mean
Life expectancy a	724	72.07	7.626	52.30	84.62	79.87	73.18	67.94	62.11
Access to information b	951	63.09	27.11	1.309	100	87.69	68.74	45.85	19.69
Industry b	1085	25.95	11.45	2.759	80.31	26.14	27.34	27.05	20.52
Subjective well-being	654	5.569	1.105	2.180	7.890	6.684	5.527	4.845	4.301
Economy scale GDP c	1095	751.9	2639	0.1093	30,340	1090	986	442	52.5
Regulatory quality	1164	−0.0308	0.992	−2.387	2.227	1.038	−0.1591	−0.6360	−0.9794

Notes:d N indicates the number of observations.

^aDenoted in years.

^bDenoted in percent.

^cDenoted in billions of USD.

Table 3.

Regression estimates on the likelihood of life expectancy. Models reflect a progression in analytical rigor, from OLS, to country-level FE, to country and year FE.

	Variables	Model 1	Model 2	Model 3
		OLS	FE	HDFE
H1a	Access to information	0.168∗∗∗	−0.0151∗	−0.0122
		(0.0105)	(0.00818)	(0.0113)
H1b	Industry	−0.0142	0.113∗∗∗	0.0813∗∗∗
		(0.0205)	(0.0202)	(0.0181)
H1c	Subjective well-being	0.974∗∗∗	0.352∗∗∗	0.297∗∗∗
		(0.283)	(0.120)	(0.105)
H2a	Economy scale GDP	1.42e-13∗∗	3.30e-13∗∗∗	1.97e-13∗∗
		(6.74e-14)	(9.72e-14)	(7.90e-14)
H2b	Regulatory quality	1.563∗∗∗	0.698∗∗	0.435
		(0.331)	(0.299)	(0.277)
	Constant	57.54∗∗∗	68.74∗∗∗	69.99∗∗∗
		(1.399)	(0.898)	(0.979)
	Observations	510	510	501
	R-squared	0.760	0.157	0.997
	Number of countries		144	135
	Country FE	No	Yes	Yes
	Year FE	No	No	Yes

Robust standard errors in parentheses.

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

Notes: OLS, FE, and HDFE denote ordinary least squares, fixed effects, and high-dimensional fixed effects. Yes indicates FE are included. Adjustment variables include economy scale and regulatory quality.

Models 1, 2, and 3 presents the main relationships without the interaction effects. Model 1 presents an ordinary least squares regression with robust standard errors. Model 2 presents a one-way fixed effects regression, controlling for country-level unobserved heterogeneity. Model 3 presents a high-dimensional fixed effects regression, accounting for both time and country-level unobserved heterogeneity. As shown in Model 1, access to information (H1a: β = 0.168, p < 0.01), subjective well-being (H1c: β = 0.974, p < 0.01), and both economy scale (H2a: β = 1.42e-13, p < 0.05) and regulatory quality (H2b: β = 1.563, p < 0.01), indicating that they are a significant explanator of life expectancy. Once country-level fixed-effects are included in Model 2, access to information (H1a: β = −0.0151, p < 0.1) loses significance and reverses direction, while industry (H2a: β = 0.113, p < 0.01). Subjective well-being (H1c: β = 0.352, p < 0.01), economy scale (H2a: β = 3.30e-13, p < 0.01) and regulatory quality (H2b: β = 0.698, p < 0.05) remain significant but decrease in magnitude. When both time and country-level fixed effects are introduced in Model 3, only industry (H2a: β = 0.0813, p < 0.01), subjective well-being (H1c: β = 0.297, p < 0.01), and economy scale (H2a: β = 1.97e-13, p < 0.01) maintain their significance. This indicates that, when taken individually, with respect to sense of coherence—industry and subjective well-being, are significant explanators of life expectancy, confirming hypotheses H1b and H1c. However, access to internet, representing comprehensibility, did not exhibit a statistically significant association with life expectancy. The findings do not show support for H1a, suggesting that after adjusting for unobserved heterogeneity, variation in internet access may not consistently align with changes in population health. With respect to generalized resistant resources, economy scale is a significant explanator of life expectancy, whereas regulatory quality is not, confirming hypothesis H2a. To further understand the relationship, we unpack our main findings according to income level. Table 4 presents the relationship between industry and subjective well-being with life expectancy by income level.

Table 4.

Regression estimates on the likelihood of life expectancy with industry and subjective well-being by income classification. Models include the interaction terms to assess how the associations between life expectancy and SOC dimensions vary by income level.

Variables	Model 4	Model 5
	HDFE	HDFE
Access to information	−0.00956	−0.0113
	(0.0110)	(0.0109)
Industry	0.0202	0.0818∗∗∗
	(0.0416)	(0.0185)
Industry ∗ Upper-mid income	0.127∗∗
	(0.0639)
Industry ∗ Lower-mid income	0.0479
	(0.0490)
Industry ∗ Low income	0.0299
	(0.0498)
Subjective well-being	0.293∗∗∗	−0.119
	(0.103)	(0.139)
Subjective well-being ∗ Upper-mid income		0.688∗∗
		(0.342)
Subjective well-being ∗ Lower-mid income		0.373∗∗
		(0.170)
Subjective well-being ∗ Low income		0.517∗∗
		(0.220)
Regulatory quality	0.344	0.501∗
	(0.295)	(0.274)
Economy scale GDP	2.06e-13∗∗	1.72e-13∗∗
	(8.80e-14)	(7.18e-1)
Constant	70.04∗∗∗	70.50∗∗∗
	(0.999)	(1.016)
Observations	501	501
R-squared	0.997	0.997
Number of countries	135	135
Country FE	Yes	Yes
Year FE	Yes	Yes

Robust standard errors in parentheses.

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

Notes: HDFE denotes high-dimensional fixed effects. Yes indicates FE are included. Adjustment variables include economy scale and regulatory quality.

Model 4 tests the interactions between industry and income level, including high, upper-middle, lower-middle, and low income. High-income countries are specified as the default and the interaction terms represent comparisons relative to this baseline. As show in Model 4, only upper-middle income is significant (β = 0.127, p < 0.05), indicating that in upper-middle income countries, the effect of industry on life expectancy is greater than in high-income countries. Industry effects’ relationship with life expectancy is significantly larger for countries in this income category, as compared to high income countries, suggesting that income level modifies the relationship and strengthens its positive association. This demonstrates that a one-size-fits all explanation may not be the most appropriate in this case, as it underscores the context-dependent nature of the relationship between industry and life expectancy.

Model 5 tests the interaction between subjective well-being and the varying levels of income level. As shown in Model 5, upper-middle (β = 0.688, p < 0.05), lower-middle (β = 0.373, p < 0.05), and low income (β = 0.517, p < 0.05) are significant, indicating that the link between subjective well-being and life expectancy also varies based on level on income. This finding suggests that the effect of subjective well-being on life expectancy is greater in all income levels—upper-middle, lower-middle, and low income—when compared to high income. Thus, an increase in subjective well-being will result in higher positive improvement in life expectancy in all countries, as compared to high income countries. This effect is greatest in upper-middle income, followed by low-income countries. The heterogeneity in the effects highlights the complexity and nuance in the relationship between subjective well-being and life expectancy, ensuring that the findings are representative of diverse realities. This provides granular insights into the relationship that allow a deeper understanding of the variation in the relationships across groups. Fig. 4, Fig. 5 present coefficient plots to visualize how the interaction effects vary across income levels.

Fig. 4.

Coefficient plot illustrating the interaction between meaningfulness and income level on life expectancy.The upper-middle-, lower-middle-, and lower-income group lie near the zero line, reflecting their statistical significance.

Notes: Interaction terms compare each income group relative to the baseline category high income, with a 95 % confidence interval.

Fig. 5.

Coefficient plot illustrating the interaction between manageability and income level on life expectancy.

The upper-middle income group lies near the zero line, reflecting its marginal statistical significance (p = 0.048).

Notes: Interaction terms compare each income group relative to the baseline category high income, with a 95 % confidence interval.

6. Discussion

In this paper, we tested whether comprehensibility, manageability, and meaningfulness improve longevity. Our arguments were based on building a sense of coherence through access to information, industry, and subjective well-being, and understanding their boundaries in different income levels. While our main arguments were primarily supported through the main analysis and robust to multiple models and specifications, we had some counterintuitive findings which prompt a more nuanced interpretation of the role of manageability and meaningfulness in longevity. We expand on these findings and their implications for theory and practice below.

We hypothesized that comprehensibility, manageability, and meaningfulness would be positively related to life expectancy. While comprehensibility was expected to be positively associated with life expectancy, we find no evidence for this link. One possible explanation lies in the likelihood of structural barriers, especially in emerging economies, that can hinder the potential benefits of internet connectivity, such as limited digital literacy and unequal access to both internet services and quality digital resources (Anrijs et al., 2023; Zhang, 2013). These challenges may reduce the expected positive relationship between access to internet and population health. Another possible explanation is the saturation of internet access in higher-income economies (Poushter, 2016), where marginal increases no longer produce measurable differences in population health. In these countries, additional access to internet penetration may not yield measurable improvements in population health after a threshold of widespread connectivity is achieved.

We find that both manageability and meaningfulness are linked to life expectancy. When unpacking these links, we find a much more defined scenario, where manageability is only applicable under a specific level of income. We also find differences by income in the relationship with meaningfulness. Thus, one key contribution of our study is that when applying salutogenesis across economies, the relationship between manageability and meaningfulness with life expectancy varies by income level, suggesting their association is context-specific. In other words, there are boundaries to the reliance on manageability for improving population health. These boundaries stem from the initial level of income of a given economy. To improve population health, some nations, namely upper-middle income, would need to address manageability and meaningfulness, since both industrialization and subjective well-being have a positive relationship with life expectancy in these contexts. However, the benefits of industrialization are less pronounced in lower-middle and low-income nations, perhaps due to weaker institutional capacity that translate industrial growth into widespread health benefits.

Thus, to improve population health in these nations, namely lower-middle and low income, a focus on subjective well-being is especially important. Subjective well-being shows a stronger association with population health in lower-income contexts, as compared to high-income countries. We explain the amplified role of subjective well-being in lower income settings through the importance of psychosocial resilience. When communities face significant structural challenges such as poverty, limited healthcare access, and poor infrastructure, higher subjective well-being may be related to alleviating the adverse health effects of stress, poor living conditions, and foster stronger social cohesion and support networks—therefore serving as a critical buffer against adversity in resource-constrained environments (Ngamaba et al., 2017). This is especially relevant when compared to high-income countries where basic needs are largely met (Ignatow, G., & Gutin, 2024; McKay et al., 2023; Wilkie & Ho, 2024).

To improve manageability through upper-middle income contexts, it is essential to address barriers that may limit equitable access to industrial benefits. This may include regional inequalities and insufficient integration of marginalized populations into industrialized sectors (UNIDO, 2024). With deliberate policies, industrial growth can alleviate, rather than exacerbate socioeconomic disparities in these contexts. Balancing growth with health equity may require intersectoral strategies that integrate industrial policy with public health, education, and labor protection (ILO, 2020). These cross-sector efforts can enhance populations capacity to access resources that help them cope with life's challenges. This is another key contribution of our study.

Life expectancy in high-income countries is often driven by structural factors such as advanced healthcare systems, technological innovations, and economic stability (Lichtenberg, 2014; World Economic Forum, 2022). In contrast, in lower income contexts where structural determinants are weaker, psychosocial factors such as subjective well-being may play a disproportionately larger role in population health (Ngamaba et al., 2017). When basic needs are more precarious, improvements in subjective well-being may be tied to tangible health benefits. Thus, subjective well-being does not follow a diminishing return pattern, and instead the association appears stronger in lower-income contexts, as compared to high income. We extend theories of health, namely salutogenesis, by emphasizing the relative importance of meaningfulness in the context of structural inequality.

Overall, we find that both manageability and meaningfulness are non-linear determinants of life expectancy. These findings point to different relationships for the components of sense of coherence to improving longevity. Our theorizing and empirical finding contribute to the broader intersection of literature on health, industry, and income. We nuance the existing literature by challenge the simplistic narratives that industrialization and subjective well-being uniformly improve life expectancy, emphasizing contextual factors.

We provide granularity by uncovering the levels of income where industry and subjective well-being have stronger associations with life expectancy. In doing so, we complement the existing literature that examines not just which components of nations sense of coherence are effective in addressing population health, but where they are most effective (Antonovsky, 1990; Bronsan et al., 2023; Eriksson & Lindström, 2007; Esch et al., 2024; Kelly & Russo, 2018; Rohner et al., 2022; Roth and Valentinov, 2023; Schäfer et al., 2020). In particular, by identifying the factors that advance longevity, we add to the streams of literature that examine the intersection of population health and industry (Rind et al., 2014; Shen et al., 2022), as well as population health and subjective well-being (Chase, 2013; Johnston-Ataata et al., 2020), in upper- and lower-income economies.

6.1. Implications for policy and practice

For practitioners, our findings speak to the importance of fit between the policy design related to life expectancy and the socioeconomic environment (Mayfour & Hruschka, 2022; Mrejen et al., 2022). While both industry and subjective well-being are associated with life expectancy, they are context dependent. Upper-middle income countries may benefit by focusing on both subjective well-being and industry. On the other hand, enhancing life expectancy in lower-middle- and low-income countries requires primarily focusing on subjective well-being.

To improve subjective well-being through policy levers, investments in psychosocial resources such as expanding community-based programs that promote social connection and civic participation (Diener et al., 2018) and integrations of mental health services with other basic healthcare systems, especially in rural and underserved areas, may yield greater returns in improving population health in the specified contexts. These strategies align with cross-national research linking well-being to structural and social determinants (Diener et al., 2018; OECD, 2021). To improve industry in upper-middle income countries, we propose cross-sector strategies that balance economic growth with equity. This includes inclusive industrial policies that promote small businesses and support workforce upskilling and reduce spatial inequality in industrial development. This may also involve investing in social protections, such as health coverage and unemployment safety nets for informal workers, to ensure that the benefits of industrial growth and distributed across the population. These strategies are in line with research emphasizing the importance of managing equity with industrial expansion (Rodrik, 2014; UNIDO, 2020). These findings underscore the need for context specific targeted policies that address the unique challenges and opportunities in different income settings.

6.2. Limitations, future research, and conclusion

Our study is not without limitations. First, our explanatory variables, our measures of sense of coherence are crude for several reasons. While they may capture aggregates of comprehensibility, manageability, and meaningfulness, they may not necessarily translate to individual-level access to the equitable distribution of benefits. For example, gains at the macro-level may be accessible to a subset of the population, yet marginalized populations may lack access to the very same resource. Thus, capturing individuals' experience and capability in managing their stressors is not perfectly reflected in this sense. We encourage future research to refine measurements for sense of coherence. While Antonovsky's (1991) theory conceptualizes manageability as individuals internal or external resources that can be drawn upon to cope, our use of industry at the national level reflects a macro-level economic condition, rather than an individual-level resource. Although economic diversification can shape the availability of opportunities, such as employment, and strength of infrastructure—which can directly support manageability—this indicator may not capture the individual dimension originally intended in the salutogenic framework. The proxies used for sense of coherence are imperfect, given the macro-level focus on this study. While they are selected based on Antonovsky's (1991) conceptualization, they do not capture individual-level perceptions directly. As an ecological study using country-level data, the findings represent macro-level associations and do not imply causality at the individual level. Accordingly, we interpret our results as associations rather than causal effects.

Second, subjective well-being, can be an input and an outcome of better population health, raising a potential for reverse causality. Similarly, we recognize that individual-level confounders, such as age and gender, and time-varying confounders, such as social protection programs, changes in healthcare infrastructure, and political instability, are not explicitly included in the model but may influence both SOC proxies and population health outcomes. We use a high-dimensional fixed effects model to help control for time-invariant country characteristics and year-specific shocks, yet do not fully mitigate concerns related to endogeneity. The use of this model reduces unobserved heterogeneity across countries and over time, however, future studies could benefit from richer panel data to further isolate the influence of SOC dimensions from confounders not included in the model.

Third, the study period, 2017–2020, was selected based on the availability of data completeness of key indicators across countries in publicly accessible databases. This reflects a common reporting lag for internationally comparable macro-level indicators. While the inclusion of 2020 introduces the potential confounding effects of the COVID-19 pandemic, the model includes fixed effects to capture global shocks affecting all countries in a given year, thereby improving inference by isolating the effects.

Finally, regarding our dependent variable, using life expectancy indicates we only capture longevity, but not the quality of life or health status over time. Life expectancy reflects cumulative health outcomes and acts as an indicator of survival. However, life expectancy alone does not capture the process throughout, nor provide feedback about current health to demonstrate movement towards progress. Thus, we encourage future research to continue this discussion by examining the relationship between sense of coherence and multidimensional health metrics to provide a more holistic understanding of population health. Limitations notwithstanding, our study offers important implications for theory and practice. Theoretically, our study extends salutogenesis discourse by scaling it towards the national level, and nuances findings across diverse income settings. Practically, our study demonstrates the importance of fit and context-sensitive targeted health interventions.

Ethical statement

This study relies on publicly available secondary data. The authors did not interact directly with participants, and therefore is exempt from additional ethical approval processes.

Declaration of generative AI in the writing process

During the preparation of this work the authors used AI assisted technology. After using AI assisted technology, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Funding

This publication was made possible by the support of the Abdulla Al Salem University open access fund.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A.

Table A1.

Sequential regression estimates of SOC dimensions on the likelihood of life expectancy.

	Variables	Model 1	Model 2	Model 3	Model 4
		Partial Model	Partial Model	Partial Model	Full Model
H1a	Access to information				−0.0122
					(0.0113)
H1b	Industry			0.0474∗	0.0813∗∗∗
				(0.0254)	(0.0181)
H1c	Subjective well-being		0.287∗∗∗	0.287∗∗∗	0.297∗∗∗
			(0.106)	(0.101)	(0.105)
H2a	Regulatory quality	0.143	0.347	0.326	0.435
		(0.365)	(0.302)	(0.283)	(0.277)
H2b	Economy scale GDP	1.65e-13∗∗	1.58e-13∗∗	1.80e-13∗∗	1.97e-13∗∗
		(8.12e-14)	(6.98e-14)	(7.16e-14)	(7.90e-14)
	Constant	72.04∗∗∗	71.32∗∗∗	70.04∗∗∗	69.99∗∗∗
		(0.0598)	(0.608)	(0.968)	(0.979)
	Observations	695	520	519	501
	R-squared	0.996	0.997	0.997	0.997
	Country FE	Yes	Yes	Yes	Yes
	Year FE	Yes	Yes	Yes	Yes

Robust standard errors in parentheses.

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

Notes: Sequential high-dimensional fixed effects. Yes indicates FE are included. Adjustment variables include economy scale and regulatory quality.

Table A2.

Sensitivity analysis. Possible confounding variables are introduced to assess the stability of estimated SOC dimensions on life expectancy.

	Variables	Model 1	Model 2	Model 3	Model 4
		HDFE	HDFE	HDFE	HDFE
H1a	Access to information	−0.0122	−0.0184	−0.0124	−0.0123
		(0.0113)	(0.0140)	(0.0114)	(0.0114)
H1b	Industry	0.0813∗∗∗	0.0822∗∗∗	0.0834∗∗∗	0.0854∗∗∗
		(0.0181)	(0.0290)	(0.0184)	(0.0226)
H1c	Subjective well-being	0.297∗∗∗	0.291∗∗	0.292∗∗∗	0.285∗∗
		(0.105)	(0.114)	(0.103)	(0.126)
H2a	Regulatory quality	0.435	0.497	0.412	0.450
		(0.277)	(0.343)	(0.279)	(0.295)
H2b	Economy scale GDP	1.97e-13∗∗	1.95e-13∗∗	2.03e-13∗∗	1.98e-13∗∗
		(7.90e-14)	(8.77e-14)	(7.91e-14)	(8.08e-14)
	Public spending education		0.0329
			(0.107)
	Political stability			0.180
				(0.182)
	Undernourishment prevalence				−0.0369
					(0.0449)
	Constant	69.99∗∗∗	70.44∗∗∗	70.00∗∗∗	70.07∗∗∗
		(0.979)	(1.310)	(0.973)	(1.178)
	Observations	501	451	501	487
	R-squared	0.997	0.997	0.997	0.997
	Country FE	Yes	Yes	Yes	Yes
	Year FE	Yes	Yes	Yes	Yes

Robust standard errors in parentheses.

∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.

Notes: HDFE denotes high-dimensional fixed effects. Yes indicates FE are included.

Table A3.

List of countries by income level. The table provides an overview of all countries analyzed categorized by their income group.

Country	Income group
	High	Upper-Middle	Lower-Middle	Low
Australia	x
Austria	x
Bahrain	x
Belgium	x
Canada	x
Chile	x
Croatia	x
Cyprus	x
Czech Republic	x
Denmark	x
Estonia	x
Finland	x
France	x
Germany	x
Greece	x
Hungary	x
Iceland	x
Ireland	x
Israel	x
Italy	x
Japan	x
Korea, Rep.	x
Kuwait	x
Latvia	x
Lithuania	x
Luxembourg	x
Malta	x
Netherlands	x
New Zealand	x
Norway	x
Panama	x
Poland	x
Portugal	x
Romania	x
Saudi Arabia	x
Singapore	x
Slovak Republic	x
Slovenia	x
Spain	x
Sweden	x
Switzerland	x
United Arab Emirates	x
United Kingdom	x
United States	x
Uruguay	x
Albania		x
Argentina		x
Armenia		x
Azerbaijan		x
Belarus		x
Bosnia and Herzegovina		x
Botswana		x
Brazil		x
Bulgaria		x
China		x
Colombia		x
Costa Rica		x
Dominican Republic		x
Ecuador		x
El Salvador		x
Gabon		x
Georgia		x
Guatemala		x
Indonesia		x
Iraq		x
Jamaica		x
Kazakhstan		x
Malaysia		x
Mauritius		x
Mexico		x
Moldova		x
Montenegro		x
Namibia		x
North Macedonia		x
Paraguay		x
Peru		x
Russian Federation		x
Serbia		x
South Africa		x
Thailand		x
Turkey		x
Algeria			x
Bangladesh			x
Benin			x
Bolivia			x
Cambodia			x
Cameroon			x
Cote d'Ivoire			x
Egypt			x
Eswatini			x
Ghana			x
Guinea			x
Haiti			x
Honduras			x
India			x
Iran, Islamic Rep.			x
Jordan			x
Kenya			x
Kyrgyz Republic			x
Lao PDR			x
Lebanon			x
Lesotho			x
Mauritania			x
Mongolia			x
Morocco			x
Myanmar			x
Nepal			x
Nicaragua			x
Nigeria			x
Pakistan			x
Philippines			x
Senegal			x
Sri Lanka			x
Tanzania			x
Tunisia			x
Ukraine			x
Uzbekistan			x
Viet Nam			x
Zambia			x
Zimbabwe			x
Afghanistan				x
Burkina Faso				x
Chad				x
Ethiopia				x
Gambia				x
Liberia				x
Madagascar				x
Malawi				x
Mali				x
Mozambique				x
Niger				x
Rwanda				x
Sierra Leone				x
Togo				x
Uganda				x

Notes: Income classification –high, upper-middle, lower-middle, and low—are based on World Bank GNI per capita.

Data availability

The data that support the findings of this study are openly available through the specified databases with each variable clearly identified by its respective name.