Insurance plays an essential role in managing global and regional investment flows and mitigating personal and business risks (OECD, 2021). It has also evolved alongside other major financial sectors in responding to economic uncertainties (Low and Fekete-Farkas, 2021). As a financial instrument, insurance contributes not only to individual welfare but also to national economic development by facilitating risk management, supporting capital market growth, and promoting long-term economic stability (Outreville, 1996; Beck and Webb, 2003; Outreville, 2014). The relationship between insurance development and economic growth has been well documented (Ward and Zurbruegg, 2000; Arena, 2008; Haiss and Sumegi, 2008; Han et al., 2010), and scholars have long examined the determinants of insurance demand across markets (Outreville, 1990; Brown et al., 2000; Park and Lemaire, 2012). While extensive evidence exists for countries with high insurance penetration, growing attention has been given to emerging markets where insurance systems remain less developed (Arena, 2008; Ching et al., 2010; Pradhan et al., 2016). Nevertheless, empirical research on the determinants of insurance demand in ASEAN markets remains limited.
The ASEAN insurance market continues to develop as income levels, financial inclusion, and life expectancy rise. Regional insurance statistics highlight persistent heterogeneity within ASEAN. While markets such as Singapore, Malaysia, and Thailand record relatively high penetration levels, emerging economies like Indonesia, Vietnam, and Myanmar show modest penetration but gradually rising density, indicating greater engagement with insurance products over time.
Specifically, Vietnam’s insurance market has demonstrated consistent expansion, with insurance density rising significantly from USD 32 in 2015 to USD 100 in 2023, driven by strong income growth and financial sector development. Indonesia exhibits a similar pattern; although penetration remains low, insurance density has increased steadily, supported by rising middle-class income and the development of both conventional and takaful (Islamic) sectors. In contrast, Myanmar remains in the early stages of market formation with extremely low penetration, yet its demographic growth signals considerable long-term potential if market conditions improve.
Taken together, Myanmar, Vietnam, and Indonesia represent three emerging ASEAN economies with distinct institutional, demographic, and cultural environments, each at a different stage of insurance-sector development. Despite their differences, all three countries share meaningful long-term growth potential. Yet empirical research on how consumers in these markets form preferences for insurers (social vs. private) and insurance products (life vs. non-life) remains scarce. Existing studies tend to focus on single-country analyses or on macro-level insurance demand, leaving a gap in understanding the behavioral and socioeconomic determinants of consumer insurance choices within ASEAN.
The purpose of this study is to address this gap by examining the determinants of insurer preferences and insurance product preferences among consumers in Myanmar, Vietnam, and Indonesia. We administered an online survey targeting adults aged 20 and above in the three countries and obtained 242 valid responses from Myanmar, 401 from Vietnam, and 488 from Indonesia after excluding incomplete responses and students. Using these data, we employ multinomial logistic regression models to analyze how socioeconomic, demographic, and behavioral factors influence individuals’ choices between social and private insurance and between life and non-life insurance.
This study contributes to the literature in two ways. First, it provides comparative evidence on early-stage insurance markets in ASEAN, focusing on three countries that have received limited academic attention despite their substantial growth potential. Second, by incorporating socioeconomic and behavioral determinants into a unified analytical framework, the study offers practical insights for insurers and policymakers seeking to design market entry strategies, expand coverage, and improve consumer engagement in emerging ASEAN markets.
Both social and private insurance serve as risk transfer mechanisms within the broader framework of risk management and share fundamental insurance principles. However, while social insurance prioritizes the social welfare purpose of providing the minimum guarantee for all citizens, private insurance functions as a collective mechanism to pool risks among individuals with similar exposure. Myanmar, Vietnam, and Indonesia have all established social insurance frameworks covering areas such as industrial accident insurance, public pension, health insurance, and unemployment insurance. In Myanmar, the Social Security Board (SSB), established in 1956 in accordance with the Social Security Act enacted in October 1954, operates social insurance. The social insurance system in Vietnam is largely divided into compulsory social insurance, voluntary social insurance, unemployment insurance, and health insurance. Indonesia has established a framework for public social security systems since the late 1990s, and the National Social Security System Act was enacted in 2004. Following fundamental reforms mandated by this Act, Indonesia’s restructured social security system has been fully operational since January 2014.
Existing literature on social insurance demand, particularly in emerging markets like Chile and Africa, identifies income, age, and residence as key determinants. For instance, Sapelli and Torche (2001) and Pardo and Schott (2014) found that individuals with lower income, older age, and those living in suburban areas tend to prefer social insurance. Conversely, demand for private insurance is often positively associated with income, education, and employment status (Liu and Chen, 2002).
Recently, behavioral factors such as trust and religion have gained attention. Scheve and Stasavage (2006) argued that religious individuals are less likely to prefer social insurance, while Guiso (2012) and Courbage and Nicolas (2020) emphasized that trust in insurance companies significantly boosts private insurance demand. In Asian markets, researchers have also noted the role of religiosity and financial literacy in shaping private insurance consumption (Gill et al., 2018).
Insurance can be broadly divided into life insurance and non-life insurance. Table 1 shows the proportion of life insurance, non-life insurance premiums, and total premiums in emerging Asian markets as of 2023. When looking at premiums and proportions, these markets are estimated to be in the early stages of insurance development, and the non-life insurance market is much larger than the life insurance market. Historically, in the early stages of the insurance market in developing countries, non-life insurance accounted for a large proportion, and as the market matures, the share of life insurance tends to increase.
Insurance Premiums in Emerging Asian Markets
| Ranking | Country | Life Insurance premium (%) | Non-life Insurance premium (%) | Total premium |
|---|---|---|---|---|
| 2 | China | 3390,400 (54%) | 333,264 (46%) | 723,664 |
| 10 | India | 100,185 (74%) | 35,773 (26%) | 135,958 |
| 27 | Thailand | 14,566 (55%) | 12,030 (46%) | 26,380 |
| 32 | Malaysia | 14,349 (71%) | 7,778 (38%) | 20,237 |
| 34 | Indonesia | 10,576 (58%) | 5,671 (31%) | 18,353 |
| 43 | Vietnam | 6,551 (69%) | 2,896 (31%) | 9,447 |
| 45 | Philippines | 5,277 (68%) | 2,480 (32%) | 7,758 |
| Others | 647 (75%) | 217 (25%) | 864 |
Note: Data are for 2023; units are millions of USD. Source: Sigma (2024) Authors’ reorganization.
The determinants of life insurance demand have been extensively studied, with economic factors such as income, inflation, and dependency ratios being consistently significant (Browne and Kim, 1993; Beck and Webb, 2003). Demographic and social factors also play a critical role; education and trust generally exhibit a positive relationship with life insurance consumption (Millo and Carmeci, 2015; Luciano et al., 2016), while the effect of religion, particularly Islam, often shows a negative correlation (Outreville, 2018).
Similarly, non-life insurance demand is influenced by income and urbanization (Esho et al., 2004). However, recent studies in emerging markets highlight the importance of behavioral factors. Trust and financial literacy have been identified as strong positive drivers for non-life insurance products, including crop and rainfall insurance (Gine et al., 2009; Cole et al., 2013). Trinh et al. (2016) further suggested that cultural dimensions, such as long-term orientation and uncertainty avoidance, differentially impact non-life insurance demand in developed versus developing countries.
We collected data using an online survey targeting adults 20 years of age or older in Myanmar, Vietnam, and Indonesia. The survey was conducted by distributing online questionnaires to residents of Yangon and its suburbs (Myanmar), Ho Chi Minh City and its suburbs (Vietnam), and Jakarta and its suburbs (Indonesia). After excluding invalid responses and students, a total of 242 responses from Myanmar, 401 from Vietnam, and 488 from Indonesia were used for statistical analysis. Tables 2 shows the structure of the survey questionnaire.
Structure of Survey Questionnaire
| Variables | No. | Item Nos. | Measures |
|---|---|---|---|
| Demographic | 7 | 1–7 | Multiple choice/short |
| Income | 1 | 8 | Short answer |
| Religiosity | 3 | 9 | Likert 7 scales |
| Financial literacy | 7 | 11–14 | OX, Multiple choice |
| Financial trust | 1 | 15 | Likert 7 scales |
| Government trust | 4 | 16 | Likert 7 scales |
| Social insurance trust | 6 | 17 | Likert 7 scales |
| Private insurance trust | 6 | 18 | Likert 7 scales |
| Insurance preference | 1 | 21 | Multiple choice |
| Insurer preference | 1 | 22 | Multiple choice |
Table 3 provides and the definition of variables. The independent variables of this study were income, religiosity, financial literacy, financial trust, government trust, social insurance trust, private insurance trust, education, age, and gender. Independent variables such as religiosity (REL), government trust (GOVT), social insurance trust (SIT), and private insurance Trust (PIT) were measured using a 7-point Likert scale (1 = Strongly Disagree/Distrust, 7 = Strongly Agree/Trust). To derive composite variables, we calculated the sum of the scores of the individual items belonging to each factor. Consequently, higher summed scores indicate higher levels of trust or religiosity. The specific survey items used for each variable are listed in Appendix A.
Definition and Classification of Variables
| Variables | Definition |
|---|---|
| lnINC (Income) | Log annual income [ISD] |
| REL (Religiosity) | Sum of likert scales |
| FLI (Financial literacy) | Sum of correct answers |
| FINT (Financial trust) | Sum of likert scales |
| GOVT (Government trust) | Sum of likert scales |
| SIT (Social insurance trust) | Sum of likert scales |
| PIT (Private insurance trust) | Sum of likert scales |
| EDU (Education) | 1= > high school |
| 0= ≤ high school | |
| AGE (Age) | Age in years |
| GEN (Gender) | 1=male |
| 0=female | |
| SPP (Insurer preference) | 1=prefer social ins. |
| 2=prefer private ins. | |
| 3= neutral | |
| LNP (Insurance preference) | 1=prefer life ins. |
| 2=prefer non-life ins. | |
| 3=neutral |
Respondents were asked to indicate their relative preference between two options (e.g., Social vs. Private Insurance) by selecting one of five ratios: 100:0, 75:25, 50:50, 25:75, or 0:100. For the multinomial logistic regression analysis, we categorized these responses into three distinct preference groups:
- a)
Preference for Option A (e.g., Social Insurance): Respondents who selected ratios of 100:0 or 75:25.
- b)
Preference for Option B (e.g., Private Insurance): Respondents who selected ratios of 0:100 or 25:75.
- c)
Neutral (Reference Group): Respondents who selected the ratio of 50:50.
This classification method was applied identically to both ‘Insurer Preference (Social vs. Private)’ and ‘Product Preference (Life vs. Non-life)’. Multinomial logistic regression analysis was employed to test the hypotheses, as it is appropriate for dependent variables with three or more unordered categories.
We formulated hypotheses based on previous literature on the determinants of insurance demand. In particular, we incorporated trust and religiosity as independent variables, which have received limited attention in previous studies. Religiosity influences one’s attitude toward the utility of insurance (Sihem, 2019); for instance, religious individuals tend to have higher subjective life satisfaction (Ellison, 1991) and are significantly less likely to purchase insurance than their non-religious counterparts (Burnett and Palmer, 1984). The insurance industry is built on trust more heavily than any other financial sector. Insurance products are inherently trust-based because consumers purchase them relying on the promise that they will receive payouts in the future in return for paying premiums. Therefore, trust plays a more pivotal role in insurance demand decisions than in other financial industries (Courbage and Nicolas, 2020). In developing countries, which are in the early stages of insurance market development, trust is paramount, and a lack thereof is a major barrier (Guiso et al., 2008; Yeshiwas et al., 2018).
H1. There is a negative correlation between income and social insurance preference
H2. There is a negative correlation between religiosity and social insurance preference
H3a. There is a positive correlation between government trust and social insurance preference
H3b. There is a positive correlation between social insurance trust and social insurance preference
H3c. There is a negative correlation between private insurance and social insurance preference
H4. There is a positive correlation between age and social insurance preference
Based on the determinants of life and non-life insurance demand identified in previous studies, we developed the following hypotheses. In particular, we focused on trust and financial literacy, variables that have received limited attention in prior research. Previous studies have consistently established a significant positive correlation between income and life insurance consumption (Browne and Kim, 1993; Outreville, 1996; Beck and Webb, 2003; Li et al., 2007; Lee et al., 2010; Elango and Jones, 2011; Park and Lemaire, 2011; Sen and Madheswaran, 2013; Dragos, 2014; Millo and Carmeci, 2015; Luciano et al., 2016; Zerriaa and Noubbigh, 2016; Lee et al., 2017; Lin et al., 2017; Dragos et al., 2020). When a high-income earner dies, the expected utility loss for dependents is greater; this, in turn, increases the value of holding life insurance (Li et al., 2007).
In general, individuals with higher levels of education possess a better understanding of risk management and risk recognition. That is, highly educated people are more likely to be exposed to information regarding the benefits of insurance, leading to a greater perceived necessity of insurance (Truett and Truett, 1990; Browne and Kim, 1993; Outreville, 1996; Liu and Chen, 2002; Li et al., 2007; Feyen et al., 2013). Financial literacy refers to the ability to process information and make decisions about financial planning, wealth accumulation, debt, and pensions (Lusardi and Mitchell, 2014).
H1. There is a positive relationship between income and life insurance preference
H2. There is a positive relationship between education level and life insurance preference
H3. There is a positive relationship between financial literacy and life insurance preference
H4. There is a positive relationship between financial trust and life insurance preference
H5. Age is positively correlated with life insurance preference
H6. Gender (male) is positively correlated with life insurance preference
We did frequency analysis to understand the general socio-demographic characteristics of research subjects. The socio-demographic characteristics by country, and characteristics of insurance and insurers by country, are reported in Appendix B.
We conducted an exploratory factor analysis (EFA) to verify the validity of the measurement items. After excluding two items (no. 27_2 and no. 27_3) that impeded validity, the factor loadings for all remaining constructs—Social Insurance Trust (SIT), Private Insurance Trust (PIT), Religiosity (REL), and Government Trust (GOVT)—exceeded the 0.4 threshold. Furthermore, we assessed internal consistency using Cronbach’s alpha. The reliability coefficients for all key variables were above 0.7, confirming that the measurement tools are statistically reliable. Detailed results of these analyses are provided in Appendix C.
To analyze the effect of Income (lnINC), religiosity (REL), government trust (GOVT), social insurance trust (SIT), private insurance trust (PIT), and age (AGE) on insurer preference (social/private), we first classified the dependent variable into three groups: social insurance preference, private insurance preference, and neutral. Table 4 summarizes the multinomial logistic (MNL) regression results for Myanmar, Vietnam, and Indonesia.
MNL Results: Insurer Preference
| Myanmar | Vietnam | Indonesia | ||||
|---|---|---|---|---|---|---|
| Independent Variable | Social | Private | Social | Private | Social | Private |
| Intercept | 1.973 | −5.703 | −1.112 | −2.446 | 2.475 | −2.951 |
| lnINC (Income) | −0.207 | 0.338 | 0.031 | 0.119 | −0.337* | 0.240 |
| REL (Religiosity) | −0.088 | 0.050 | 0.003 | −0.024 | −0.007 | −0.047 |
| GOVT (Govt. Trust) | −0.013 | 0.013 | 0.008 | 0.037 | 0.054* | 0.027 |
| SIT (Social Ins. Trust) | 0.005 | −0.069* | −0.008 | −0.040 | 0.017 | −0.080** |
| PIT (Private Ins. Trust) | −0.007 | 0.097** | −0.012 | 0.042 | −0.046 | 0.096** |
| Age | 0.020 | 0.014 | 0.045*** | 0.006 | 0.003 | −0.012 |
| Pseudo R2 (Nagelkerke) | 0.119 | 0.074 | 0.105 | |||
| −2 Log Likelihood | 429.685 | 788.519 | 942.601 | |||
| N | 242 | 401 | 488 | |||
Note: The reference category is “Neutral”. Coefficients shown are B values. Significance levels:
p<.05,
p<.01,
p<.001.
First, regarding Myanmar, social insurance trust is negatively related to private insurance preference, while private insurance trust is positively related. In other words, lower trust in social insurance and higher trust in private insurance increase the likelihood of preferring private insurance compared to the neutral option. Specifically, a one-unit increase in social insurance trust decreases the probability of preferring private insurance by 6.68%, whereas a one-unit increase in private insurance trust raises it by 10.15%.
In the case of Vietnam, age is positively related to social insurance preference. The results indicate that older respondents are more likely to prefer social insurance over the neutral option, with the probability increasing by 4.59% for each additional year of age. On the other hand, for private insurance preference, no independent variables were found to be statistically significant.
Finally, for Indonesia, income is negatively related to social insurance preference, while government trust shows a positive relationship. Also, the social insurance trust is negatively related, and the private insurance trust is positively related to private insurance preference. As income increases, the likelihood of preferring social insurance decreases by 28.62%, whereas higher government trust increases this likelihood by 5.56%. Furthermore, as the social insurance trust rises, the preference for private insurance drops by 7.65%, while a higher private insurance trust boosts the preference for private insurance by 10.13%.
To analyze the effect of Income (lnINC), education (EDU), financial literacy (FLI), financial trust (FINT), age (AGE), and gender (GEN) on insurance product preference (life/non-life), we classified the dependent variable into three groups: life insurance preference, non-life insurance preference, and neutral. Table 5 presents comparative results for each country.
MNL Results: Insurance Preference
| Myanmar | Vietnam | Indonesia | ||||
|---|---|---|---|---|---|---|
| Independent Variable | Life | Non-life | Life | Non-life | Life | Non-life |
| Intercept | 0.129 | −5.198 | −3.526 | −3.090 | −2.414 | −1.436 |
| lnINC (Income) | 0.098 | 0.276 | 0.240 | 0.068 | 0.133 | −0.170 |
| EDU (Education) | −1.351* | 0.235 | 0.094 | 0.093 | −0.070 | −0.421 |
| FLI (Fin. Literacy) | −0.096 | 0.349** | −0.006 | 0.017 | 0.256*** | 0.302* |
| FINT (Fin. Trust) | 0.113 | 0.157 | −0.008 | 0.164 | −0.013 | 0.027 |
| Age | −0.011 | 0.020 | 0.051*** | 0.020 | 0.006 | 0.003 |
| GEN (Gender/Male) | 0.178 | 0.403 | −0.113 | 0.150 | 0.792*** | 0.520 |
| Pseudo R2 (Nagelkerke) | 0.185 | 0.087 | 0.103 | |||
| −2 Log Likelihood | 431.513 | 739.780 | 853.183 | |||
| N | 242 | 401 | 488 | |||
Note: The reference category is “Neutral”. Coefficients shown are B values. Significance levels:
p<.05,
p<.01,
p<.001.
In Myanmar, education levels above a bachelor's degree are significantly and negatively related to life insurance preference; those with a college degree or higher are 74.09% less likely to prefer life insurance than those with a lower high school diploma. Conversely, financial literacy is positively related to non-life insurance preference, with the likelihood increasing by 41.77% as financial literacy scores rise. For Vietnam, age is positively associated with life insurance preference. As age increases, the probability of preferring life insurance rises by 5.21%. In Indonesia, gender is a significant predictor, being positively related to life insurance preference. Additionally, financial literacy is positively associated with a preference for both life and non-life insurance.
This study investigated the determinants of insurer preferences (social vs. private insurance) and insurance product preferences (life vs. non-life insurance) among individuals in Myanmar, Vietnam, and Indonesia—three ASEAN countries representing different levels of insurance market maturity. Using survey data and multinomial logistic regression, this study identifies both shared and country-specific drivers of insurance preferences, offering empirical insights into consumer behavior in early-stage and transitioning markets.
The results indicate that socioeconomic and demographic variables, along with behavioral factors such as trust and financial literacy, shape insurance choices in distinct ways across the three countries. Income emerged as a significant determinant in Indonesia, where higher-income respondents were less likely to prefer social insurance. This is consistent with prior findings that income-linked contributions in national health insurance systems can encourage higher-income households to opt out in favor of private coverage. Trust also plays a central role in shaping insurer preferences. In Indonesia, trust in government institutions increases the preference for social insurance, whereas trust in private insurers enhances the preference for private insurance in both Myanmar and Indonesia. These results reaffirm the inherently trust-based nature of insurance markets and the importance of institutional reliability for consumer decision-making.
Educational attainment and age further influence insurance preferences, but their effects differ by country. In Myanmar, higher education is associated with a lower likelihood of preferring life insurance, suggesting that general educational attainment alone may not translate into increased awareness of life insurance needs. In Vietnam, however, both life insurance and social insurance preferences rise with age, highlighting demographic differences in risk perception and household financial planning. These findings point to the heterogeneity of consumer behavior across ASEAN markets and underscore that insurance demand cannot be understood through economic variables alone; institutional, behavioral, and cultural contexts are equally important.
The implications of this study reflect the varied levels of insurance development across the three markets. Myanmar, still in the earliest stage of insurance market formation, would benefit from initiatives that build consumer trust and improve awareness, particularly in relation to life insurance products. Given the limited availability of products and low public familiarity, insurers operating in Myanmar must prioritize clear communication, claims transparency, and consumer education as foundational steps toward expanding the market.
Vietnam presents a different set of challenges and opportunities. With rising income levels and a growing insurance sector, Vietnam's insurance demand is shaped by demographic factors, particularly age. As older consumers show stronger preferences for both social and life insurance, policy reforms aimed at improving accessibility and benefiting adequacy for these groups may be necessary. Simultaneously, insurers seeking to attract younger consumers should emphasize product innovation, financial education, and targeted marketing strategies.
Indonesia, the largest and one of the fastest-growing markets among the three, exhibits a dual system in which social and private insurance coexist as both substitutes and complements. The influence of trust on social insurance participation highlights the importance of strengthening transparency, governance, and administrative efficiency in the national health insurance system. Private insurers, operating in a rapidly expanding but trust-sensitive environment, must address capacity constraints by improving product expertise, distribution quality, and claim-handling processes. The strong association between financial literacy and insurance preference in Indonesia further suggests the need for broader financial education initiatives to support informed consumer decision-making.
Overall, this study contributes to the limited body of literature examining consumer-level insurance behavior within ASEAN. The findings demonstrate that insurance preferences in emerging markets cannot be explained solely through traditional socioeconomic indicators; instead, behavioral characteristics such as trust, financial literacy, and demographic structure play significant roles. For policymakers, the results highlight the importance of enhancing institutional reliability, promoting financial literacy, and designing policies that reflect demographic and cultural differences. For insurers, the findings provide strategic guidance for market entry, product development, and consumer engagement, particularly in markets undergoing rapid socioeconomic transformation.
Finally, this study has limitations regarding sample representativeness. Given that the survey was conducted online and distributed primarily in major urban centers (Yangon, Ho Chi Minh City, and Jakarta), the sample may over-represent younger, more educated, and internet-savvy individuals compared to the general population. While this demographic constitutes the emerging middle class, the primary target market for commercial insurance in these developing economies, the findings may not fully reflect the preferences of rural or lower-income populations. Future research should aim to incorporate offline surveys in rural areas to address this sampling bias and enhance the generalizability of the results.