India is currently experiencing substantial transformations across multiple dimensions. From a political economy perspective, a central concept encapsulating this transition is “inclusive growth,” defined as economic growth that broadly enhances living standards across diverse segments of the population (Cerra et al., 2021). While inclusive growth has been a fundamental objective of public policies in India since independence (Kurian, 2008), this policy agenda gained renewed prominence following Narendra Modi’s rise to power in 2014, as his government articulated a development strategy explicitly aligned with his campaign slogan, “Sabka Saath, Sabka Vikas” (along with all, development for all).
An essential consideration in this context is that the effectiveness of inclusive growth hinges upon equitable political participation, broadly defined as citizens’ engagement in activities intended to influence government decision-making, policies, or leadership, including voting, attending political meetings, and engaging in public discussions (Lamprianou, 2012). Nevertheless, significant disparities in citizens’ political participation have persisted over time in India.
In this context, previous studies have examined potential factors contributing to unequal political participation among Indian voters (e.g., Mahanta & Nayak, 2013). However, they have not investigated whether disparities in voters’ health—conceived both as the general functioning of the body (Simon et al., 2005) and as a state of psychological and emotional well-being that enables individuals to manage life’s stresses, realize their abilities, and actively contribute to their communities (World Health Organization, 2022)—are systematically associated with disparities in political participation. Yet, from the perspectives of rational choice and political mobilization theories, both widely recognized frameworks for explaining political participation, a significant association between voters’ health and their level of political engagement can reasonably be anticipated.
Proponents of rational choice theory argue that individuals decide to engage in politics based on self-interest and a cost-benefit analysis, aiming to maximize personal gains while minimizing potential losses (Downs, 1957). Meanwhile, advocates of political mobilization theory emphasize that voters become politically engaged through external mobilization efforts, including formal campaign appeals as well as informal political conversations with friends and family. According to this perspective, alongside sociodemographic resources such as income and education, political-psychological resources—namely political information, interest, and efficacy acquired through interpersonal interactions within social networks—significantly shape individuals’ political participation (Rosenstone & Hansen, 1993; Schlozman et al., 2018). Despite their differing theoretical emphases, both perspectives commonly suggest that voters’ health can affect political participation by influencing their cost-benefit assessments and the extent of social interactions, ultimately contributing to variations in political engagement (Kim & Lee, 2023).[1]
Against this background, I aim to examine whether a correlation exists between Indian voters’ health and their levels of political participation, with a particular focus on electoral participation. This emphasis on voting, rather than other forms of political participation, is justified by the recognition that voting represents the most direct and foundational means through which citizens exert influence over government decisions.
Given the complexity of India’s healthcare system and the persistently low public expenditure on health (National Health System Resource Centre, 2023), disparities in health among Indian citizens are likely to remain deeply entrenched. Consequently, inequalities in political participation may also persist, potentially weakening democratic representativeness and responsiveness in India (Pacheco & Ojeda, 2020). Particularly in New Delhi, where logistical or infrastructural barriers persist, voters with poor health may face greater difficulties in reaching polling stations due to limited access to transportation and the lack of widespread vote-by-mail systems. As these inequalities provide critical insights into the broader trajectory of India’s ongoing political transition, examining the relationship between health and political participation in the context of India is both necessary and timely.
With this background, I seek to answer the following important question: Is there a relationship between Indian voters’ health and their electoral participation? The findings of this article can contribute to a deeper understanding of the interplay between health and political participation, offering valuable policy implications for Indian policymakers.
The remainder of this article is structured as follows: Section 2 elaborates on the two theoretical frameworks employed—rational choice theory and political mobilization theory—and discusses the association between voters’ health and electoral participation based on these perspectives. Section 3 describes the data and methodology, followed by the execution of empirical analyses and a robustness check using ordered logistic regressions (OLRs) in Section 4. Finally, Section 5 concludes with a summary of the results, implications, and limitations of this article.
Theoretical Framework
To establish a clear foundation for this article, it is essential to first define the concept of inclusive growth as it relates to the vision of “India for All.” While scholarly definitions vary, a common thread among them is the notion that inclusive growth entails economic expansion that enhances living standards across diverse population groups (Cerra et al., 2021). In the Indian context, the concept has evolved beyond purely economic considerations to encompass broader social dimensions, including advancement and empowerment. Specifically, inclusive growth in New Delhi is understood as fostering conditions in which all individuals—regardless of socioeconomic background, gender, ethnicity, or other identities—can actively participate in and benefit from the development process.
A critical consideration here is that achieving inclusive growth fundamentally depends on equality in political participation. Inclusive political participation practically facilitates inclusive growth by ensuring diverse voices influence policy making and resource allocation, thereby enabling equitable access to economic opportunities and social services. In this regard, Acemoglu and Robinson (2012) assert that inclusive growth requires the establishment of inclusive economic institutions. Further, they argue that the long-term sustainability of these economic institutions is unattainable without political inclusivity, which ensures broad-based citizen participation in governance. Nevertheless, a notable gap persists in political participation among citizens in India.
With this background, I employ rational choice theory and political mobilization theory to examine the interrelationship between health and electoral participation—areas that have received insufficient scholarly attention. These theoretical frameworks are selected not only for their established efficacy in explaining political participation but also for their complementary analytical strengths.
Rational choice theory emphasizes internal motivations, positing that political participation, such as voting, results from individuals’ self-interest-driven and cost-benefit evaluations (Downs, 1957; Markowski et al., 2024). In contrast, political mobilization theory underscores external factors—such as organized campaign strategies and social networks—that foster political engagement by enhancing individuals’ political information, interest, and perceived efficacy (Rosenstone & Hansen, 1993; Schlozman et al., 2018). Although these theories prioritize different dimensions—individual incentives versus social contexts—their integration provides a comprehensive framework for analyzing how health shapes political participation by influencing both internal decision-making and external mobilization opportunities.
Let us first examine the rational choice theory’s explanation regarding voters’ health and their electoral participation. This theoretical perspective assumes that voters are rational actors capable of assessing and comparing the costs and benefits associated with voting, subsequently making decisions that align with their personal interests. Accordingly, rational choice theory posits that voters’ decisions to participate or abstain from electoral activities are driven by their calculated assessments aimed at maximizing their individual interests.
Drawing from this theoretical framework, it is reasonable to expect that voters’ health directly influences how difficult or costly they perceive voting to be, thereby affecting their likelihood of participating (Kim & Lee, 2023). Specifically, less healthy voters may face greater physical challenges in activities such as traveling to polling stations or waiting in line, making voting more burdensome and less appealing (Engelman et al., 2022). Additionally, unhealthy voters may also face greater difficulties accessing political information or engaging with political issues due to the time and energy spent managing their health (Blais, 2000; Gagné et al., 2020; Kamens et al., 2019; Kim & Lee, 2023).
Likewise, political mobilization theory also provides further support for the relationship between health and electoral participation. Unlike rational choice theory, political mobilization theory emphasizes the critical role played by political actors, such as parties and candidates, in mobilizing electoral participation. Mobilization includes both direct interactions—such as personal meetings or electronic communications—and indirect engagement through social and political organizations (Rosenstone & Hansen, 1993). In line with this theory, voters with poor health are expected to participate less frequently in such interactions due to physical or mental constraints, thus limiting their opportunities for political mobilization and ultimately reducing their electoral participation.
A particularly significant issue highlighted by this theory is the tendency of political actors to exacerbate disparities in electoral participation by selectively mobilizing voters who are perceived to contribute more positively to electoral outcomes. Such targeted mobilization often leads to disparities along educational and income lines, disproportionately neglecting voters from low-education and low-income backgrounds (Wielhouwer, 2005). In this context, given that lower education and income levels correlate strongly with poorer health outcomes (Ettner, 1996; Ross & Wu, 1995), unhealthy voters are less likely to be targeted by political actors, further diminishing their electoral engagement.
Considering that these theoretical discussions emphasize the significant relationship between health and electoral participation among voters, they also have broader democratic implications, particularly concerning representativeness and responsiveness of democracy. The participatory mechanism of policy representation underscores the role of elections in conveying constituent preferences to policymakers, reinforcing the principle that unequal participation leads to unequal influence (Lijphart, 1997). Consequently, disparities in electoral participation based on health suggest varying levels of policy responsiveness among different groups. Empirical evidence further indicates that such disparities contribute to biases in both political representation (Gilens, 2013) and policy outcomes (Hill & Leighley, 1992). Moreover, voters generally receive greater representation than non-voters, as they actively express their preferences to elected officials (Griffin & Newman, 2005). Given that healthy individuals vote at higher rates than unhealthy individuals and that voters are typically better represented compared to non-voters, unhealthy individuals are consequently less represented than healthy people due to disparities in electoral participation (Pacheco & Ojeda, 2020). This dynamic raises critical concerns regarding the inclusiveness of democracy, the equitable distribution of political representation, and the extent to which policymaking adequately addresses the needs of marginalized populations.
Based on these theoretical arguments, I propose the following hypothesis.
Hypothesis 1 (H1): Indian voters with poorer health are significantly less likely to participate in elections compared to voters with better health.
Data and Methodology
To empirically test the aforementioned hypothesis, I utilize the seventh wave of the World Value Survey (WVS) dataset conducted in India from June to July 2023 (N = 1,692).[2] The initial sample size of this dataset is reduced to 1,049 in the actual analyses after excluding missing data, non-responses, “don’t know” responses, and responses irrelevant to the target variable. The WVS dataset is chosen for its reliability, extensive coverage, and representative sample size, making it a frequently employed resource in research. Moreover, its recent publication of Indian data distinguishes it from other similar comparative surveys.
Regarding the measurement of the dependent variable, I selected the item asking, “When elections take place at the national level, do you vote always, usually, or never?” This item assesses the degree of respondents’ electoral participation using a 4-point scale from 1 (always) to 4 (not allowed to vote). I recoded this item so that higher scores correspond to higher levels of electoral participation and “not allowed to vote” responses were discarded due to their irrelevance to the analysis of voter turnout.
Next, I operationalized individual health using an item that asks respondents, “All in all, how would you describe your state of health these days? …” The response options range from very good (1) to very poor (5). Similar to the dependent variable, I recoded this item so that higher scores indicate better health, simplifying interpretation.
However, the use of self-rated health as a measure of individual health is subject to several criticisms. Some scholars argue that subjective health assessments may not adequately account for variations in individuals’ health baselines and the nuanced interpretations underlying respondents’ self-evaluations (Borawski et al., 1996).
Despite these limitations, I contend that self-rated health remains a reliable and practical predictor for the following reasons: First, subjective health indicators have been shown to strongly correlate with objective health measures, demonstrating their reliability (Bjorner et al., 2005). Second, and more importantly, objective health measures often capture differences only at the extremes of the health spectrum, limiting their ability to provide sufficient variance for analyzing the relationship between health and electoral behavior. In contrast, self-rated health measures assess individual health on a continuum, encompassing a broader range of health conditions and offering a more nuanced perspective. This suggests that self-rated health measures may be more effective in capturing the link between health and electoral participation (Pacheco & Fletcher, 2015). Given these advantages, I adopted self-rated health as the primary measure of individual health.
To account for potential confounding factors, several control variables identified in previous studies as significant determinants of electoral participation are included (Ansolabehere & Hersh, 2013; Goldberg, 2014; Jensen & Jespersen, 2017; Kim & Lee, 2023). These factors encompass some sociodemographic variables, such as respondents’ age, gender, education, religious denomination, caste,[3] and income level, as well as political psychological variables, such as levels of partisanship, political ideology, political satisfaction, and political trust.
For sociodemographic variables, age is treated as a continuous variable in the main analyses, with a supplementary categorical measure dividing respondents into three age groups for robustness checks[4] (Appendix 1). Gender is measured based on respondents’ self-reported sex, coded as 0 for male and 1 for female. Education is classified into nine categories following the International Standard Classification of Education (ISCED) 2011. Religious affiliation is categorized as atheist, Christian, Jewish, Muslim, Hindu, Buddhist, or other. Caste is measured using four categories—Scheduled Caste (SC), Scheduled Tribe (ST), Other Backward Classes (OBC), and General[5]—numerically coded from 1 (SC) to 4 (General). Income level is assessed on a 10-point scale, where 1 represents the lowest income group and 10 represents the highest income group. Additionally, for robustness checks, income is also measured as a three-category variable (low, middle, high) (Appendix 2).
Turning to political psychological variables, partisanship is classified into three categories: 0 (not a member of a political party) to 2 (active member of a political party). Political ideology is measured on a 10-point scale, where 1 represents far left and 10 represents far right. Political satisfaction is assessed using a 10-point scale based on respondents’ evaluations of the political system in India, where 1 indicates not satisfied at all and 10 signifies completely satisfied. Political trust is measured as the average of three items assessing trust in the government, political parties, and parliament, following the approach of Kim and Lee (2023). Separate analyses are conducted using each trust variable individually to ensure robustness (Appendix 3).
Since the dependent variable is ordinal with more than two categories, an ordered logit model is employed for analysis. It is important to acknowledge that using an ordinal model may lead to some loss of information due to categorization. Additionally, the coefficients in an ordered logit model represent changes in the log odds of moving to a higher category rather than direct changes in probability, making interpretation more challenging compared to standard linear regression. Nonetheless, I employed the ordered logit model, as it is well-suited for analyzing ordinal dependent variables.
Table 1 presents the descriptive statistics. The mean level of electoral participation (M = 2.65) suggests that Indian respondents scored above the midpoint of the scale. Regarding health and gender differences, the data indicate that most Indian participants reported health above the midpoint of the scale (M = 3.89) and were proportionately distributed across gender (M = 0.48).
The mean values of certain numeric control variables, including income (M = 5.46), political satisfaction (M = 6.24), and political trust (M = 6.11), were above the midpoint of their respective scales. However, the mean levels of other factors, such as education (M = 3.28) and partisanship (M = 0.45), fell below the midpoint.
Results and Discussion
Table 2 presents the outcomes of estimations using the OLR model. A closer look at Table 2 reveals a statistically significant positive coefficient for individuals’ health. Specifically, the finding suggests that healthier voters are more likely to participate in voting, while those who are less healthy are less likely to vote (*p < .*001). In terms of percent change, a one-unit increase in the health variable results in a 28 percentage point increase in the odds of electoral participation when accounting for other predictors. This outcome robustly supports H1.
Beyond the main independent variable, I found that several control variables—including age, gender, education, religious affiliation (Islam, Hinduism, and Buddhism), caste (particularly the OBC category), income, partisanship, and political satisfaction—have significant effects on electoral participation. Specifically, age, education, and political satisfaction exhibit statistically significant positive correlations with electoral participation. In terms of magnitude, each one-unit increase in age is associated with a 3 percentage point increase in the odds of electoral participation, while education corresponds to a substantial increase of 17 percentage points. Additionally, political satisfaction is linked to an 9 percentage point increase in electoral participation. These findings suggest that older, more educated, and politically satisfied individuals in India are more likely to engage in the electoral process.
In contrast, individuals’ gender, caste, income, and partisanship are negatively associated with electoral participation. After controlling for other factors, each unit increase in these variables corresponds to a decrease in the odds of electoral participation by 33 percentage points for females, 82 percentage points for Muslims, 71 percentage points for Hindus, 82 percentage points for Buddhists, 39 percentage points for OBCs, 7 percentage points for income, and 35 percentage points for partisanship. These findings suggest that Indian voters who identify either as female, Muslim, Hindu, Buddhist, or OBC, as well as those with higher incomes or stronger party affiliations, are less likely to participate in elections.
Subsequently, I re-evaluated the model reported in Table 2 by employing an alternative indicator for the dependent variable as a robustness check. Specifically, I used an indicator that measures the extent of voters’ intention to participate in local elections. For this analysis, the dependent variable is recoded so that higher scores indicate higher levels of electoral participation at the local level. The findings from this robustness check are presented in Table 3.
The robustness check confirms that the earlier findings remain unaffected by the specific operationalization of the dependent variable, indicating the robustness of the main analysis outcomes. The health variable remains significant in this robustness check, with its coefficient showing consistent patterns as observed in the previous regression analysis, thereby supporting H1. Controls that were initially significant in the main analysis maintain their significance and exhibit consistent patterns. However, the effect of the gender variable, which was initially significant, loses its statistical significance in the robustness check.
In summary, the findings from both the main analysis and robustness check confirm the significant effect of individuals’ health on their electoral participation. This indicates that disparities in electoral participation due to disparities in health do indeed exist in the Indian context, demonstrating consistency with previous literature and robustly supporting H1. These results align with prior research that has identified health as a significant determinant of voter turnout (Gollust & Rahn, 2015; McGuire et al., 2021) and reinforce the broader understanding that marginalized groups, particularly those with poorer health, face barriers to electoral participation.
In addition, a number of control factors were identified as relevant to Indian voters’ participation in the election as well. These variables include individuals’ age, education, religious denominations (i.e., Islam, Hinduism, and Buddhism), caste categorizations (i.e., OBC), income, partisanship, and political satisfaction.
Given their substantial influence on Indian politics, I specifically examined the effects of religious affiliation, particularly among Muslim and Hindu voters, income, and partisanship. The lower electoral participation among Muslim voters may be attributed to the declining number of Muslim candidates in general elections, which dropped from 115 in 2019 to 78 in the 2024 general elections. Muslim voters often exhibit a preference for candidates from their own community due to shared social and religious identities (Heath et al., 2015). From a rational choice theory perspective, the reduction in Muslim candidates may have constrained political options, thereby reducing the perceived benefits of voting and lowering voter motivation. Similarly, from a political mobilization theory perspective, the absence of strong Muslim candidates may have diminished mobilization efforts, further discouraging electoral participation among Muslim voters.
Regarding Hindu voter behavior, I posit that the perception of an assured BJP victory likely contributed to reduced voter turnout. Opinion polls consistently projected a landslide victory for the BJP and its allies in the recent 2019 and 2024 general elections. This expectation may have fostered complacency among party workers, leading to decreased mobilization efforts and, consequently, lower Hindu electoral participation (Kumar & Ahmed, 2024).[6] From a rational choice theory perspective, Hindu voters may have calculated that the costs of voting outweighed the benefits, particularly given the widespread belief that the election outcome was already determined. Similarly, from a political mobilization theory perspective, the expectation of a BJP landslide may have led to diminished campaign outreach and voter engagement efforts, as party workers and political organizations prioritized constituencies perceived as more competitive. With fewer mobilization initiatives, Hindu voters may have been less mobilized to participate, further contributing to the decline in turnout.
With respect to the results concerning income and partisanship, it is well-established—particularly in advanced democracies such as the United States—that stronger party affiliation and higher income levels are typically associated with higher voter turnout. However, the findings in this article may suggest a different pattern in the Indian context, which may reflect unique sociopolitical dynamics, such as disillusionment among wealthier voters or the complex role of party identification in a multiparty system with regional and identity-based parties. Rather than viewing this as a contradiction, I believe these unexpected findings highlight the need for further empirical investigation. Future research should thus explore the underlying mechanisms that may account for these associations in India, potentially considering factors such as political cynicism, strategic disengagement, or the role of informal mobilization structures that might influence lower-income or less-affiliated voters differently.
Conclusion
Drawing from rational choice theory and political mobilization theory, I examined the hypothesis that disparities in health among Indian voters lead to disparities in electoral participation (H1). Findings from both the main analysis and robustness checks confirm that individuals’ health has a significant effect on electoral participation. This indicates that disparities in health contribute to unequal electoral participation in the Indian context, providing strong empirical support for H1.
Additionally, I examined the impact of religious affiliation on electoral participation in India, focusing on Muslim and Hindu voters. The decline in Muslim candidates may have limited political choices for Muslim voters, reducing both motivation and mobilization efforts. Meanwhile, lower Hindu voter turnout may be attributed to the expectation of a decisive BJP victory, which likely led to complacency among party workers and a reduced perception of voting benefits. These patterns align with rational choice and political mobilization theories, though further empirical validation is needed.
Lastly, I argued that the unexpected findings related to income and partisanship point to a divergent pattern in the Indian context, likely shaped by distinctive sociopolitical factors such as disillusionment among wealthier voters or the complexities of party identification within a multiparty and identity-driven system. Rather than interpreting these results as contradictory, they highlight the need for further research into underlying mechanisms, including political cynicism, strategic disengagement, and informal modes of political mobilization.
In light of the findings, I propose several policy recommendations to address health-related barriers to voting in India. First, policymakers should establish institutional frameworks that enhance accessibility for voters with health-related challenges while maintaining electoral integrity. Addressing barriers for persons with disabilities is particularly crucial to ensuring equal voting rights. Currently, many persons with disabilities face significant challenges in reaching polling booths, with some even having to crawl on the floor to cast their votes (The Hindu, 2024). While the Election Commission of India (ECI) recently introduced home voting for voters above 85 years of age and persons with disabilities with at least 40% disability[7] in the 2024 general elections, concerns regarding the transparency and security of home voting persist.
To balance accessibility with electoral integrity, India could implement several alternative solutions. For instance, deploying mobile polling stations in communities with a high concentration of elderly and disabled voters would allow them to vote without traveling long distances. Additionally, providing free or subsidized transport services—such as wheelchair-accessible vans and special rickshaws—on election day could significantly reduce barriers to electoral participation. Establishing designated “priority pick-up points” at community centers for voters needing assistance and partnering with ride-hailing companies (e.g., Uber, Ola) to offer free rides for voters with disabilities could further improve accessibility. Furthermore, developing a secure electronic voting system for voters with severe disabilities and piloting an online voting program for authenticated persons with disabilities could enhance participation. Introducing “priority voting hours” (e.g., early morning or late evening slots) would also help reduce crowding and waiting times for vulnerable voters.
While this article presents significant implications, it also acknowledges certain limitations. For instance, like all survey-based datasets, the WVS dataset may have inherent reliability concerns. To strengthen the validity of the findings, future research should consider supplementing this dataset with in-depth interviews or qualitative studies to provide a more comprehensive assessment of the relationship between health and electoral participation in the Indian context.
Funding
This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2017S1A6A3A02079749).