Family formation and economic insecurity among youth in South Africa

ABSTRACT

Youth unemployment continues to be a burden and concern for the South African government. Being economically insecure, the situation is dire with the government needing to provide resources to a population who should be economically independent. There is a need to look at the social determinants of economic insecurity among youth in South Africa. Family formations could either promote or inhibit economic well-being. This article aims to assess whether economic security improves as youth enter into unions and/or have children. The South African National Income Dynamics Study is used. Unmarried youth with no children are measured at baseline (2008) and followed up over time to examine whether economic security status changes as union status changes. Results show that while economic security, employment (from 7.61% to 25.67%) and net income per month (from 19.48% to 32.79%) increase over time, youth who marry but have no children have the lowest risk of economic insecurity (relative risk ratio = 0.02, p < 0.05) compared with those who remain unmarried but have children. Special attention needs to be given to youth who have children and are unmarried and among those who marry and have children soon after.

KEYWORDS:

• Family formation
• economic insecurity
• South Africa
• longitudinal
• National Income Dynamics Study
• relative risk ratios

1. Introduction

Income insecurity is high among youth in South Africa. An estimated 36.9% of youth are unemployed, while 21.1% are employed on limited duration contracts and 32.9% are not in employment, education or training (Statistics SA, 2015b). A further 34.3% of youth aged 15 to 19 years live in households with no employed members and almost 13% are accessing at least one government-disbursed social grant (Statistics SA, 2015a, 2015b).

Despite these economic challenges, youth also face several health and social challenges including HIV/AIDS and gender inequality in access to services. Amidst all of this, youth are fulfilling social roles such as family formation. In South Africa, pathways to family formation are unconventional with fertility sometimes preceding and even replacing marriage. Approximately 48.4% of the female population are never-married mothers and a further 2.2% are single mothers due to divorce or widowhood (Statistics SA, 2013). Research has adequately identified the economic constraints that single and never-married mothers face; these include poverty and financial challenges and reduced residential stability, and even child health outcomes are known to suffer in these households (Thomson & McLanahan, 2012; Carrion et al., 2015; Ribar, 2015).

In South Africa, however, it is not only single mothers who are affected by economic insecurity; young married couples face similar challenges too. According to national statistics, 47.75% of young married couples are unemployed and almost 2% are living on social grants from the government (Statistics SA, 2015a). Insecure employment, such as limited duration and contract work, as well as job loss prior to and during the early stages of marriage are the causes of this. Young married couples are also financially independent from their parents, and in newly formed households with two (or more) residents; there is therefore a need for sufficient income to cover the basic needs of everyone in the household. Such financial pressures on young couples are known to result in stress and violence within the union, and also marital dissolution (Dew et al., 2012; Schwarzer & Schulz, 2012; Daoud et al., 2015). While much is known about the economic consequences of early marriage and single motherhood, these are largely treated as autonomous statuses that are not interlinked. Therefore, what remains unknown is the relationship between economic insecurity and marriage with children, marriage without children and remaining unmarried but having children. Each of these is a different form of family types and this study examines the likelihood of economic insecurity by each of these formations using longitudinal data.

The relevance of this research lies in its attempt to understand the relationship between social processes, in this case family formation, and economic security in South Africa. For many youth, marriage and/or childbearing are normative transitions to adulthood and are even recognised as rites of passage. When these transitions occur it is vital that youth are financially fit to secure their entry into adulthood. In conducting this research, vulnerable youth who are forming families and are at risk of economic insecurity are highlighted. From this point, it is important that policies and programmes in the country acknowledge the risks and address these challenges directly.

The aims of this research are therefore to measure the levels of economic insecurity among youth in South Africa over time and to examine the relationship between diverse types of family formations and economic insecurity over time.

2. Methods

2.1. Data

Data were taken from two panel waves of the South African National Income Dynamics Study (SA-NIDS), the first national panel study in South Africa. The SA-NIDS was undertaken by the South African Labour and Development Research Unit based at the School of Economics of University of Cape Town. This study uses data from 2008 (Wave 1) and 2012 (Wave 3). A stratified, two-stage random cluster sample design was employed to sample households for inclusion at baseline and was stratified based on the 53 district councils in the country. From each district council, primary sampling units were drawn (Leibbrandt et al., 2009). The household-level response rate was 69% and the individual response rate within households was 93%. The baseline SA-NIDS survey provides data for 28 247 individuals (including children) from 7301 households (Leibbrandt et al., 2009).

2.2. Study population for secondary data analysis

The study population are youth between the ages of 15 and 24 years in the first wave (2008). Further among this age group, only those who were unmarried and did not have children in 2008 were included and followed up over time. The initial sample size was 3134, and 2536 individuals were successfully followed up to 2012. This shows an 80.92% follow-up rate or a 19.08% attrition rate (Figure 1). Both males and females are included in the analysis. The sample size therefore includes all youth who were successfully followed up from Wave 1 to Wave 3 (2536 individuals).

Figure 1. Cohort follow-up.

2.3. Outcome variable

Economic insecurity is a composite measure consisting of the following components in the survey: employment status (yes or no); type of employment (permanent, limited duration contract or unspecified); monthly income (arranged around the mean); and social grant reliance (if the couple rely on social grants to ensure livelihoods). Respondents who were employed, on permanent contracts, with an above-average income and not accessing social grants are considered ‘economically secure’. Respondents who are not employed, have less than average or no income and are accessing at least one social grant are considered ‘economically insecure’. Respondents who are a combination of these two extremes are considered ‘partially secure’.

2.4. Explanatory variables

At baseline, respondents who are not married and have no children are selected. In the outcome period (2012), the formation of three types of families among the baseline respondents are measured. These include whether the respondents went on to: marry and have children; marry without having children; or remained unmarried but now have children. Those in the sample who remained unmarried and childless in the latter period are used as the reference category.

Other determinants used as control variables in the study are age, gender, race or population group, education status, religion and health status.

2.5. Statistical analyses

Analyses were performed using Stata software version 13. Clustering as well as survey design effects were accounted for using sample weights to correctly estimate the determinant standard error and hence significance.

Frequency and percentage distributions are used to describe the study population at baseline and follow-up period. Multinomial logistic regression is used to identify the determinants of economic insecurity among youth in South Africa.

3. Results

3.1. Sample characteristics

Table 1 presents a description of the demographic characteristics of the youth who participated in the study during the first wave. Results from the table show that the study population consists of more males (54.77%) than females (45.23%) and, with respect to the age groups, more participants are between age 15 and 19 years (59.46%) compared with those between age 20 and 24 years (40.54%). Among the different racial groups, results show that the African participants constituted 88.29%, as compared with the other groups (11.71%). The most dominant religious groups is Christianity (81.39%), with Traditional religion constituting 3.71% while groups that fall under the category ‘other’ constitute the lowest percentage (2.40%). The highest education level attained is secondary education and contains the highest percentage (85.29%), with a significantly low percentage constituting participants with no education (0.79%). With regards to health status, the majority of the participants indicated excellent health (75.75%), while 5.05% indicated poor health.

Table 1. Percentage distribution of demographic characteristics among youth, Wave 1 (2008).

Characteristic	Frequency (n = 2536)	Percentage
Sex
Male	1 389	54.77
Female	1 147	45.23
Youth groups
15 to 19 years	1 508	59.46
20 to 24 years	1 028	40.54
Race
African	2 239	88.29
Other	297	11.71
Religion
None	313	12.5
Christian	2 038	81.39
Traditional	93	3.71
Other	60	2.40
Education
None	20	0.79
Primary	353	13.92
Secondary	2 163	85.29
Health
Poor	128	5.05
Good	487	19.20
Excellent	1 921	75.75

In Wave 1, 7.61% of the respondents indicated that they were employed, with this percentage increasing to 25.67% in Wave 3 (Table 2).

Table 2. Percentage changes in economic security over time, Wave 1 (2008) and Wave 3 (2012).

Economic security measure	Wave 1 (2008) (%)	Wave 3 (2012) (%)
Employment status
Yes	7.61	25.67
No	92.11	74.30
Missing	0.28	0.03
Type of employment^a
Limited	22.28	14.51
Unspecified	42.49	28.50
Permanent	34.20	56.99
Net income per month^a
≤ZAR 2404.46	80.52	67.21
≥ZAR 2404.44	19.48	32.79
Social grant recipient
Yes	2.09	20.81
No	97.91	79.19

^aOnly of those employed.

Furthermore, 92.11% in Wave 1 indicated unemployment, while this percentage was reduced in Wave 3 (70.30%). In terms of the type of employment of the respondents who are working, 22.28% indicated limited type in Wave 1 with this number decreasing to 14.51% in Wave 3. Those who indicated an unspecified type of employment decreased from 42.49% in Wave 1 to 28.50% in Wave 3. Respondents who are permanently employed also increased from 34.20% in Wave 1 to 56.99% in Wave 3. For net income, 80.52% of respondents indicated receiving a net income of less than or equal to ZAR R2404.46 per month, compared with 67.21% in Wave 3. Approximately 19.48% of respondents in Wave 1 receive a net income of more than ZAR 2404.44 per month (the average income of the cohort), while in wave 3 approximately 32.79% of respondents received the same amount. Social grant recipients increased from 2.09% in Wave 1 to 20.81% in Wave 3, while those who are not recipients of the social grant decreased from 97.91% in Wave 1 to 79.17% in Wave 3.

3.2. Relationship between economic insecurity and family formation

Figure 2 shows the percentage distribution of the sample at Wave 3 (2012) by family status and economic security status. The figure shows that, among the unmarried, economic security is highest for those who do not have children (90.23%). Among those married without children, about 64% are economically secure and 4% are economically insecure. Among those married with children only 7.14% are economically secure, with the vast majority (95.83%) being insecure.

Figure 2. Description of the study sample at Wave 3 by family status and economic security status.

Partial economic security is selected as the base outcome in Table 3. The table shows that the odds of complete economic security are lowest for females (odds ratio [OR] = 0.65, p < 0.005), primary school education (OR = 0.73, p < 0.005) and those who are unmarried with children (OR = 0.22, p < 0.005). The likelihood of being economically secure is highest for those with secondary (OR = 1.03, p < 0.005) or tertiary (OR = 1.11, p < 0.005) education and for those married with children (OR = 0.51, p < 0.005), married with no children (OR = 2.64, p < 0.005) and unmarried with no children (OR = 1.96, p < 0.005). The table also shows the likelihood of economic insecurity and shows that females (OR = 2.35, p < 0.005) and those unmarried with no children (OR = 1.09, p < 0.005) have higher likelihoods of being economically insecure. Further, coloured race (OR = 0.59, p < 0.005), secondary (OR = 0.21, p < 0.005) and tertiary (OR = 0.14, p < 0.005) education, and married with children (OR = 0.68, p < 0.005), married with no children (OR = 0.02, p < 0.005) and unmarried with no children (OR = 0.02, p < 0.005) have lower odds of being economically insecure.

Table 3. Multinomial logistic regression showing the odds of economic security by union status.

			95% confidence interval
Wave 3	Relative risk ratio	p value	Lower	Upper
Partial security (base)
Complete
Sex (RC = males)
Female	0.65	0.054	0.41416	1.00744
Age group (RC = 18 to 25 years)
26+ years	2.69	0.000	1.90200	3.79883
Race (RC = African)
Coloured	2.51	0.000	1.62265	3.87614
White/Asian/Indian	0.58	0.606	0.0755	2.50883
Religion (RC = none)
Christian	1.60	0.113	0.89466	2.87317
Other	2.06	0.055	0.98601	4.31759
Education (RC = none)
Primary	0.73	0.0045	0.08298	2.4663
Secondary	1.03	0.036	0.01273	2.40526
Tertiary	1.11	0.022	0.01312	2.44448
Health status (RC = excellent)
Good	0.28	0.199	0.87985	1.84899
Poor	0.57	0.358	0.60067	4.09338
Family type (RC = unmarried no children)
Married with children	0.51	0.177	0.18439	1.36436
Married no children	2.64	0.035	0.83314	2.37755
Unmarried with children	0.22	0.003	0.00044	0.38532
Insecure
Sex (RC = males)
Female	2.35	0.045	1.02084	5.42242
Age group (RC = 18 to 25 years)
26+ years	1.07	0.553	0.85273	1.34686
Race (RC = African)
Coloured	0.59	0.005	0.41319	0.8545
White/Asian/Indian	0.83	0.844	0.13386	5.17306
Religion (RC = none)
Christian	0.67	0.094	0.4247	1.06938
Other	0.68	0.191	0.38863	1.20735
Education (RC = none)
Primary	0.33	0.163	0.06833	1.56976
Secondary	0.21	0.045	0.04305	0.96854
Tertiary	0.14	0.013	0.01139	0.66091
Health status (RC = excellent)
Good	0.87	0.27	0.67063	1.11812
Poor	0.54	0.15	0.85572	2.77409
Family type (RC = unmarried no children)
Married with children	0.68	0.000	0.00807	0.04589
Married no children	0.02	0.000	0.0022	0.13932
Unmarried with children	1.09	0.003	0.00397	0.88717

Note: RC = reference category.

4. Discussion

The economic security of youth is important on several levels. First, on a micro-level youth who are economically secure are able to provide at least the basic needs for themselves, their children and any elderly or ill family members in their care. Second, on a macro-level these youth contribute to the development of the country, and are less dependent on the state to meet their basic household needs. For these reasons, ensuring that youth are economically stable is a pivotal goal for developing countries, including South Africa where about 37% of the population are youth (15 to 24 years old) (Statistics SA, 2015a:60). Further, an important milestone for youth is family formation, and the current study’s contribution lies in examining the economic risks youth face as they form families for the first time.

The relationship between family formation and economic security is complex. Economic security status changes as youth enter marriage and again as children are added. Youth who remained unmarried (single) but have children are the least economically secure, while youth who enter marriages and do not have children right away tend to have more economic security. Single parents globally suffer worse economic outcomes than their married counterparts (Duncan & Magnuson, 2013; Duncan et al., 2013). In South Africa, this is no different. Single parents are known to suffer chronic poverty, to have lower levels of education and literacy, and to suffer worse health outcomes than dual-parent households (Aliber, 2003; Spaull, 2013; Meinck et al., 2015). The explanation for this is straightforward; a single income does not provide as much support to multi-person households as dual incomes. While social grants are available to assist, the values of these do not substitute additional employment wages and benefits (Crush & Caesar, 2014).

Young married couples, however, are able to assist each other financially, hence increasing the economic security of individuals in the relationship (Shapiro et al., 2013; Nakano, 2016). This is true of an unemployed (or limited contract) partner who marries an employed spouse. This is even truer of two employed persons who get married and remain employed during the union. In the latter situation, where there are no children, there are fewer expenses and hence more economic security. However, as this study has shown, economic security decreases when children are added to the marriage. This is because there are three or more family members relying on the income of one or both spouses. This creates a deficit, which at least temporarily sets the family unit back.

Overall, employment, permanent contract employment and net monthly income increased among youth over time. The study also shows that about 85% of the sample had at least a secondary education as baseline. Given this high rate of education and the progressing age of the cohort, increased employment is expected. Related to this, the South African government has prioritised the employment of youth in the country, even offering tax incentives to private companies who specifically hire youth (Ranchhod & Finn, 2014). The majority of youth also report excellent health at baseline. The absence of ill-health improves productivity and broadens the type of employment opportunities for youth in the county.

The study is not without its limitations. First, as is the case in all longitudinal or panel studies, is the issue of loss to follow-up. In this study about 19% of the respondents from Wave 1 to Wave 3 were not followed up. Given the relatively small percentage, this did not greatly affect the study. Second, the measures of economic security are based solely on employment-type indicators, which do not fully encompass security. Given the nature of the data and the emphasis on socio-economic factors, they were selected as the best fit for the research question. However, future research which interviews respondents on their economic needs and perceptions of economic security would assist in creating a more holistic depiction of economic security. Finally, the family formations used in this study are few and do not include forms of cohabitation.

A few strengths of the article are also worth noting. First, the study uses longitudinal data and methods, allowing for causation to be determined. Second, the sample size is fair and this has enabled more accurate empirical results to be deduced from the study. Finally, the data allow for more diverse measurement of family formation types. These types mirror the current trends in the country and are therefore relevant to the context.

In conclusion, the plight of unmarried youth with children and economic instability is highlighted in this study. Shown to be the most economically insecure of the different family types, particular attention should be paid to improving the economic stability of this group. Herein, policies such as the Labour Relations Amendment Act of 1992 and the Skills Development Act of 1998, which aim to reduce employment inequalities and improve the skills of young people in the country, could benefit from the results of this study. National programmes, such as the career guidance and Operation Fikelela programmes offered by the Ikamva Youth organisation, which seek to address the skills shortages in the country by empowering youth, should note the particularly precarious situation young, unmarried couples face and introduce specialised programmes for this sub-population.

Within this field of study, a few future research recommendations can be made. First, qualitative studies which determine partner selection criteria for marriage and parenting could help in better understanding how families are formed, based on what criteria, and determine whether financial stability before marriage is essential to partner selection. Second, there should be more research on other diverse forms of families not captured in this study, including but not limited to polygynous households and same-sex families. The current research did not account for the make-up or structure of the couple and the recommended study would inform economic stability differences by types of unions which are uncommon but nonetheless relevant. Finally, other social factors that affect economic security need to be addressed over longer periods of time. For example, a study which looks at the impact number of children and other living arrangements, such as resident grandparents, aunts, uncles and so forth, would give a better indication of economic standing by household structure and over time.

Acknowledgements

Demography and Population Studies Programme, Wits University; WZB-ISSC Global Fellowship Programme 2016. Opinions expressed and conclusions arrived at are those of the author and are not necessarily to be attributed to the Centre of Excellence in Human Development.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

The support of the National Research Foundation Centre of Excellence in Human Development towards this research is hereby acknowledged.