This study produced three main findings. The first and most important one is that, when the analysis simultaneously included the three dimensions of informal employment, not having social security coverage was the strongest predictor of poor health status for both women and men. Second, when these dimensions were examined separately, not having social security coverage, being self-employed, and being employed with an oral or no contract, were strongly associated with poor health outcomes in both sexes. Lastly, among employees of both sexes with social security coverage, those with an oral or no contract were more likely to report poor mental health.
Our results are consistent with previous studies, which have found that workers in informal employment settings have poorer health status. One study reported that being an informal worker was associated with common mental disorders compared with formal workers . Other studies also observed that women working in informal employment were more likely to report poor mental health [17, 19]. Another study found that female housemaids (mostly with informal job contracts) had worse mental health indicators, including depression and anxiety symptoms, than women with other occupations (principally those with formal job contracts) . Some studies carried out in South Africa, which constructed a complex formality index [30, 31], have also shown that informal employees were more likely to report poorer health status, although it depends on the interaction with earnings. Finally, another study reported that living in a household with at least one informally employed person was associated with poor self-perceived health status, regardless of individual socioeconomic factors and housing characteristics . However, most prior research on the Latin America region has compared only formal versus informal employment, aggregating several categories of informal employment that can actually have different meanings and therefore different impacts on health [17, 19, 20]. In considering different categories of informal employment, we have identified a lack of social security coverage as a key issue.
There are several non-mutually exclusive mechanisms that might explain the relationship between informal employment and health status. Firstly, it is important to note that our reference employment profile is basically the Fordist notion of standard employment (employees covered by social security with a written contract) , which has never represented the main labor market relation paradigm in Central America. Social security coverage and formal employment were more prominent in the region before the economic crisis in the 1980s. Since that time, formalization and salaried processes have declined. The informal economy took hold as a result of unemployment (mainly of formal workers) and the flexibility and precariousness of the labor market, which characterized the neoliberal process of structural adjustment of this crisis coupled with the globalization of markets . In this context, a mechanism that could be operating is the health-related features of the wider phenomenon of employment precariousness. In Central America, there is considerable employment precariousness, particularly affecting informal workers .
Employment precariousness is characterized by employment insecurity, economic vulnerability, temporality, low collective bargaining power, low earnings, and a lack of social protections; all of which have the potential to affect a worker’s health [35, 36]. In many cases, these characteristics are present in informal employment, so we cannot rule out the hypothesis that the poor health of informal workers runs through the health-related features of employment precariousness.
Employment precariousness in turn is related to wider social precariousness, including denied access to health care, which may also affect the health status among workers in informal employment. In Central America, between 13 % (Guatemala) and 42 % (El Salvador) of people do not have health service coverage . Informal employment and poverty were among the access barriers identified as well as the structure and organization of health systems, which differentiates countries regarding the quality and the diversity of health services offered and that are accessible to people who need them, as pointed out in different studies [38–40]. Therefore, precarious employment could also result in more precarious lives, and the persistence of poverty and poor living conditions [41, 42].
Secondly, although we adjusted for occupation as a proxy of working conditions, differences in working conditions between formal and informal workers could still persist. Working conditions in informal employment are usually poorer than those in formal employment . This is illustrated by taking the occupation of vendors as an example. In Central America, the percentage of informal vendors far exceeds the percentage of formal ones (75 % versus 25 %, respectively) . Also, in our sample there was a huge proportion of vendors, primarily informal self-employed and informal employers, comprising around 70 % and 72 % of women and 41 % and 34 % of men, respectively. Most of the formal vendors were not working on the street, but almost 20 % of women and 30 % of men who were informal vendors were street vendors (results not shown). As Marcelli et al. illustrate , “Selling oranges in a grocery store is a formal economic activity. Selling them on a highway exit ramp in Los Angeles County to passing motorists is an informal activity”. Some of the working conditions that may be different between the two kinds of vendors are that street vendors are exposed to long working hours, unsafe workplaces, traffic pollution, musculoskeletal problems, inclement weather, and even sexual harassment among women [45–47]. Lastly, a remarkable difference in working conditions between formal and informal workers, not only in vendors, could be the exposure to long working hours, which have a harmful impact on workers’ health and with different gender patterns . In our study, most of the informal employment profiles were more exposed to long working hours (exceeding 48 h per week) for both, women and men. This is particularly the case of employees with no social security coverage with an oral or no contract (41 % of women and 45 % of men), self-employed with no social security coverage (34 % and 38 % of women and men, respectively), and employers with fewer than five employees with no social security coverage, among whom 48 % of women and 36 % of men worked more than 48 h per week. On the contrary, only 19 % of women and 28 % of men employees with social security coverage and a written contract were exposed to long working hours (results not shown). These results are consistent with previous studies which have found a relationship between socioeconomic vulnerability and acceptance of obligatory long working hours and other poor social and economic conditions .
Additionally, there is a close relationship between employment and working conditions and living conditions outside of work. Poor working and employment conditions, such as those that often characterize informal employment, are connected to poverty, with occupational hazards and poor living conditions combining in nonadditive ways . Hence, we also suggest that poverty and social exclusion could be another possible mechanism explaining poorer health outcomes among people performing informal work [51–54]. Despite the fact that not all informal workers are poor (one study shows there is a proportion of the informal sector that on average earns more than its formal counterpart ), lots of poor workers are informally employed, with an over-representation of women in low- and middle-income countries . Likewise, for a large proportion of people in poverty or on the border of social exclusion, informality is a survival strategy as it is the only way to enter into the labor market. At the start of the century, the informal sector in Latin America accounted for around 70 % of employment among the urban poor (approximately 80 % in Guatemala, Honduras and Nicaragua, 74 % in Panama, 62 % in Costa Rica, and 36 % in El Salvador) . In addition, almost four out of ten households in Central America are estimated to be in a situation of exclusion (approximately three out of ten in urban areas and five out of ten in rural areas). People living in such households enter the labor market through subsistence self-employment (95 %), without social security coverage (more than 99 %), and more than half have not completed primary education .
Hence, there could be a vicious circle of informality, labor precariousness, poor working and living conditions and poverty that may generate and perpetuate health inequalities among many Central American workers by the different axes of health inequalities. The employability of certain groups more prevalent in the informal economy, such as women or indigenous people, is even more difficult and is mediated by exclusion dynamics, which makes it very hard for members of these groups to break out of this cycle .
Likewise, labor precariousness and poor working conditions do not only affect to informal workers in Central America. The effects of globalization and structural adjustment programs on the quality of employment could also have led to a systemic precariousness of formal employment in the region (specifically for paid employment) and a deterioration of working conditions, regardless of the formal or informal nature of the job [10, 11, 57]. Therefore, they could explain the significantly poorer mental health found for employees with social security coverage and an oral or no contract. Contrarily, there was no significant association with poor self-perceived health in this employment profile. As previously mentioned, informal employment could be a barrier to accessing good quality health services [38–40]. So we suggest that these employees, covered by social security, could more easily access good quality of health services, and therefore they have a better perception of their health. Finally, it is important to notice that this employment profile only accounted for around 4 % of workers, and among them 61 % of women and 53 % of men were from Costa Rica (results not shown). Although the analysis were adjusted by country, it may be possible that this finding is showing a specific reality of this country, which should be studied more deeply in the future.
There were small gender differences, either in the prevalence of different types of informality or in the pattern of association between informal employment and poor health status. The prevalence of informal profiles was slightly higher among women, contrary to results from previous research. Future studies will be needed to determine whether the pattern is changing in Central America or whether perhaps the ECCTS is not fully representative of informal employment in the region.
On the other hand, regarding the pattern of association between informality and poor health status, there was only one remarkable gender difference. Female employees, without social security coverage but with a written contract, were not different from the reference category (covered by social security and with a written contract). For men in the same situation, the odds of poor health outcomes were significantly higher. Possible explanations for this finding could be the interplay among the axes of social inequality in provoking health inequalities. In this case, gender inequalities could be mediated by social class, as these women are clearly in a more favorable social position than men in the same employment profile. Whereas a large proportion of these women were in non-manual or skilled occupations (56.6 %), men were more often represented in manual and unskilled ones (47.2 %). In addition, 60.2 % of the women had a university or secondary education versus only 41 % of the men. Moreover, the number of weekly working hours also differed. While 21.7 % of women worked more than 48 h per week, 35.4 % of men worked more than 48 h. Finally, another noteworthy result is that 61.8 % of women were young, aged between 18 and 30 years old, unlike the men with only 44.8 % were in this age group (results not shown).
It is well-known that there are large health inequalities by gender, but they also depend on interactions with social class and other axes of inequality [58, 59]. In our study, the employment profile of employees without social security coverage but with a written contract applied to groups whose characteristics differed by gender, in that it was more favorable for women, who had better health than men (probably due to their more favorable social class, as noted above). Therefore, the meaning of these dimensions, as well as their influence on health, differed by gender and social class. This employment profile is an interesting example, as the observed differences may be due to these women being in a more favorable social class than the men with this employment profile, as most of them are young women with high education and skilled non-manual jobs, who have good health.
In order to fully understand the gender dimension of informality and its impact on health, future studies should examine the interaction between informal employment and family characteristics, since their relation with the labor market for women, often through informal employment, is still mediated by the woman’s role at home and the general socioeconomic situation of the household . For example, one study shows that 21.7 % of women in Latin America with the highest household incomes only engage in unpaid work (at home) in contrast with 46.5 % of women in lower income households . This study also reported that in Honduras, the labor market participation rate of women from households with no children under six years old is 42.6 %, while for those having three or more children in that age range, the rate decreased to 26.6 % (differences were seen regarding educational level, as the participation rate of the latter women is 23.3 % for those with less than four years of education, but 72.6 % for those with thirteen or more years of education). Finally, the interaction and the intersectionality between informal employment, gender and other axes of inequalities such as age group, immigration , ethnicity , social exclusion , and territory  should also be analyzed in detail in future studies .
Of note, social class is also a mechanism that could be operating not only on the gender differences observed, but transversally in the different mechanisms involved. Since formal employment in Central America remains exceptional, the selection into formal employment is mainly subject to a medium and high level of social closure. In other words, it is practically reserved for individuals of an advantaged social class, well-positioned in society, and enjoying resources that facilitate their life course, including access to higher education, health care, and good living conditions . As a consequence, these workers enjoy a better health status. Therefore it is possible that this category of formal workers, as the reference employment profile, is in fact a highly selective well-off sample. This could partly explain the strong health inequalities found. In that sense, the distinction between formality and informality could become a proxy for class inequalities. Just as informality is the path for many of the poor, formality could be the path for many of the favored social classes.
Furthermore, a main characteristic of this reference employment profile of formal workers is the availability of social security coverage, with the security and all the job benefits that it represents for them including retirement pension, paid vacations, sick leave, maternity/paternity leave, weekends off, personal/family leave, or breastfeeding time for women. The ILO strategy of formalization of employment highlights the importance of extending social protections to all workers to reach the goal of decent work in the immediate term . Promotion of this strategy for the working poor has also shown that it constitutes a key pathway to reduce poverty and improve their working and living conditions, with an emphasis on women because of their large numbers in the informal economy . Therefore, if we consider that work is a central element of life for many people, giving them access to economic resources and opportunities for achieving good health , it would be essential to encourage decent work by the extension of social security coverage for all workers.
Strengths and limitations
Among the strengths of the study, it is important to remark that, as far as we know, this is the first time that reliable and homogeneous information has been gathered in Central America about informal employment and health status. Most research has neglected the potential association with poor health status. Moreover, the study is based on a large and representative sample of Central American workers. Additionally, we have constructed a complex employment profile with different dimensions of formal and informal employment, which advances our understanding of the complex universe of informality instead of simply dichotomizing between formal and informal employment. Nevertheless, this study has several limitations. For example, we could not exactly apply the classification proposed by ILO  since the survey did not allow us to distinguish between contributing family workers or members of producers’ cooperatives for jobs, or between households for the type of production unit. Moreover, we had to use a proxy of the informal sector (fewer than five workers) because the questionnaire did not ask about the legal registry of the company. Furthermore, we could not separate employees according to different informal sectors because of the limited sample size. Since this is a cross-sectional study, we cannot rule out the possibility of reverse causation, whereby rather than the experience of informal employment leading to poor health status, it may be that people with poorer health are more likely to work in informal job arrangements. Moreover, people with good health may be more likely to work in formal employment due to their favorable social status. Finally, since the analysis was carried out for the entire Central American region, we cannot rule out potential differences between countries derived from their political and cultural differences. Despite this, our results are consistent with our hypotheses, and we may assume that they could be transferable across countries of the region. Nevertheless, future studies will have to be performed but using countries separately in order to deepen the specificities of each one.