Parents’ labour force participation and children’s involvement in work activities: evidence from Thailand

ABSTRACT

This study provides a better understanding of children’s engagement in economic work and housework by examining its relationship with parents’ labour force participation. It also explores how parents’ employment type is associated with children’s involvement in work activities. Using Thailand’s Labour Force and Time Use Surveys, our multivariate probit regression results show that girls actively participate in economic activities when their mothers are employed, while boys’ involvement in such work is positively correlated to both parents’ employment. Girls’ housework participation is also positively associated with parents’ employment, suggesting that their assistance in household chores enable their parents to stay in the labour market. These positive relationships are prevalent particularly among children with either mothers or fathers working informally. Hence, the findings suggest that anti-poverty or expansionary policies aimed at increasing labour force participation without attention to job quality, social protection and care needs can adversely affect children by increasing their need to work.

KEYWORDS:

• Children’s work activity participation
• parents’ labour force participation
• informal employment
• Thailand

1. Introduction

Guided by the International Labour Organization (ILO), many countries have adopted legislation prohibiting or limiting the employment and work of children to protect them from any adverse effects, while at the same time recognising the benefits that children can gain (ILO, 2014). Thailand refined its labour laws in 2017 to comply with international labour standards. In particular, it set the minimum legal working age for general and agricultural work, and work performed at home, to 15 years. It specified the types of work prohibited for home-based workers under age 15, and increased penalties for the workplace exploitation of children (Inthonamuy, 2016). Yet, Thailand’s labour laws do not protect those working without a formal employment contract. Consequently, about 13% of Thai children under 15 are employed while 14.4% of children aged 7–14 combine work with school (U.S. Department of Labour, 2017). With poverty forcing many children into employment, 20% of secondary school-age children from the poorest households do not attend school, compared to 4% for children in the top wealth quintile (NSO and UNICEF, 2016).

While involvement in work activities allows children to acquire basic work experience and learn responsibility, it can hinder school attendance and academic performance (Ravallion & Wodon, 2000; Warren, 2007). Augmenting household productivity and earnings, children’s economic work involvement is a survival strategy among households that do not generate sufficient incomes for subsistence (Jacoby & Skoufias, 1997; Tzannatos, 2003). Their housework participation reduces adults’ household responsibilities, increasing the time available for labour market work (Susanli et al., 2016; Warren, 2007). Given that parents’ employment is vital for household welfare, a better understanding of the relationship between parents’ employment and children’s economic activity and housework participation can help in the formulation of policies aimed at reducing poverty, stimulating economic growth by increasing labour force participation, and improving children’s welfare.

This study examines the relationship between children’s involvement in work activities and parents’ labour market participation by answering the following questions: Does children’s involvement in labour market and household work correlate with their mothers’ and fathers’ labour force participation? How does parents’ employment type correlate with children’s involvement in these work activities? Answers to these questions can help policymakers make informed decisions, ensuring that anti-poverty and pro-growth policies expand employment opportunities without jeopardizing children’s welfare.

In our analysis, work activities refer to domestic chores and care tasks as well as paid and unpaid economic work, which includes working as unpaid labourers on farms and in family businesses. Parents’ employment includes formal and informal employment. The latter comprise economic activities that do not provide any employment security, legal protection, and employer-sponsored benefits, while formal employment offers legal protection through labour laws and social security coverage (NSO, 2017).

Our study is the first to investigate the relationship between both mothers’ and fathers’ labour force participation and children’s involvement in work activities at home and in the labour market, especially in the context of Thailand. Although several studies have found a negative correlation between parental earnings and children’s employment (Amin et al., 2006; Binder & Scrogin, 1999; Goldin, 1979; Holleran, 1997; Togunde & Carter, 2008; Wahba, 2006), few studies have examined the relationship between a parent’s employment and children’s overall work activity participation. In particular, studies by DeGraff and Levison (2009), Dayıoğlu (2008), Self (2011), and Susanli et al. (2016) focus solely on mothers’ and children’s employment, while Manacorda (2006) explores the effect of children’s employment on their parents’ labour supply. With the exception of Susanli et al. (2016) and Self (2011), most studies do not examine children’s involvement in housework. Moreover, we distinguish between parents’ formal and informal employment and explore whether children’s involvement in work activities correlates with the parents’ employment type, which the existing research does not explicitly examine.

We use the combined Labour Force Survey (LFS) and Time Use Surveys (TUS) collected by Thailand’s National Statistical Office (NSO). The LFS contains data on parental employment and household characteristics, while the TUS captures the extent to which children are involved in economic activities and housework. These surveys are nationally representative, thus allowing this study to generalize findings to the entire country. Earlier studies rely on either census data, such as the 1920 US Census (Manacorda, 2006), or special surveys, such as the 1994 and 1999 Child Labour Surveys (Dayıoğlu, 2008) and the 2003 Turkey Demographic and Health Survey (Susanli et al., 2016). These surveys, however, lack detailed information on children’s work activity participation and parents’ employment characteristics.

The rest of this paper is organized as follows: Section 2 explains the conceptual framework. Section 3 describes the LFS–TUS data. Sections 4 to 5 describe the methodology and results for each research question. Finally, Section 6 discusses the policy implications of the study’s findings.

2. Conceptual framework

Our conceptual framework extends the work of DeGraff and Levison (2009). We start with household labour supply decisions, which are derived from household utility maximization involving the household’s consumption of goods, leisure, and child education, subject to time and budget constraints (Basu & Van, 1998; DeGraff & Levison, 2009). Child education and leisure are vital for well-rounded child development and can be adversely affected by children’s significant involvement in work activities. The outcomes of the household utility maximization are demand functions in reduced form. The full theoretical model is located in the paper’s appendix.

We focus on the relationships among selected outcomes: parents’ employment decisions and children’s participation in each type of work activity. These decisions are assumed to be made simultaneously by parents. Parents decide whether to work and in what type of employment, while deciding whether to engage their children in economic and/or household work. Hence, parents’ employment and children’s work activity participation are determined jointly.

2.1. Parents’ employment and children’s economic work

The relationship between parents’ employment decisions and children’s involvement in economic activities is a net outcome of the substitution effect, income effect, or added worker (complementary) effect. The substitution effect captures a situation in which an improvement in the labour market opportunities of a family member, such as a parent, increases his/her opportunity cost of not working. Thus, s/he spends more time in economic work and less time in other activities, reducing the need for children to work, implying a negative correlation between children’s work and parent’s employment.

On the other hand, the income effect posits that an increase in a family member’s or parent’s income enables the parent to work less and spend more time on non-work activities. This allows other family members, namely children, to also work less. The outcomes of the income effect with regards to parent’s and children’s participation in work activities move in the same direction and therefore yield a positive association between parent’s labour force participation and children’s work activity participation. In the added worker (complementary) effect, family income depends on the family enterprise or farm and is positively correlated with the number of family members working in that enterprise or farm. In this case, the outcomes also move in the same direction such that, when parents’ employment increases, children’s participation in economic activities also increases. This also results in a positive association between parents’ employment and children’s economic work. Children’s employment is found to be positively correlated with mothers’ employment in Brazil (DeGraff & Levison, 2009), India (Self, 2011), and Turkey (Dayıoğlu, 2008).

In our analysis, the negative substitution effect is likely to be weaker for informally employed parents than for formally employed parents. The positive income effect, however, would hold regardless of the parents’ employment type. The added worker (complementary) effect is likely to apply to those in informal employment but not in formal employment. While the former allows children to work as family workers without contracts or fear of inspection and fines, the latter prohibits or discourages the hiring of family members. This distinction regarding employment type is pertinent in Thailand, where informal employment is the source of earnings for 62.4% of the employed population (ILO, 2015). With its lack of income security, informal employment yields lower earnings than formal employment, and most informal workers belong to the lowest income group (Chen et al., 2004; Dasgupta et al., 2015).

Children’s economic work involvement is likely to be positively associated with the informal employment of parents, given that they provide additional labour at little or no cost on farms and in informal enterprises (Dasgupta et al., 2015; Francavilla et al., 2013). Children’s participation in economic work can be a coping strategy among those relying on informal employment yielding variable earnings. Informally employed parents, particularly mothers, may have their children work alongside them as alternative childcare because they are unable to afford other childcare options (Bertulfo, 2011).

2.2. Parents’ employment and children’s housework

The relationship between parents’ labour supply and children’s involvement in housework is influenced by the income or substitution effects. As a parent’s income increases, household members spend less time in economic work and more time in other activities, including housework. Thus, parents’ employment is negatively correlated with their children’s participation in housework under the income effect. With the substitution effect, an improvement in parental earnings creates an incentive for the person to increase his or her labour supply, reducing their time for other activities. Other household members may be required to supplement the time devoted to domestic and care tasks, indicating a positive relationship between parents’ employment and children’s housework participation.

Other factors also influence these relationships. Societal attitudes and gender norms permeating household decision-making processes and labour markets influence how sons and daughters are raised and the position of women and men within the labour market. For instance, filial piety, a central value in Thai culture, encourages their children, namely daughters, to help their mothers perform household chores (Kanchanachitra, 2014; Limanonda, 1995). This is especially true if the parents are unable to afford domestic help. Gender-based discrimination in labour markets can also propel women to become self-employed or engaged in home-based work in the informal sector. This implies a positive correlation between parents’ informal employment and children’s housework participation, particularly among mothers and daughters.

Additionally, some Thai families emphasise child development whereby the parents avoid having children perform housework and encourage them to be involved in activities essential to their development, such as studying, reading, or playing a musical instrument (Petchprasert, 2014; Tulananda & Roopnarine, 2001). This is possible for the formally employed with stable income streams, suggesting that parents’ formal employment and children’s housework may be negatively correlated.

The relationship between parents’ employment and children’s involvement in housework is understudied. The few extant studies suggest an ambiguous relationship. Susanli et al. (2016) find, for example, that children’s participation in housework and mothers’ employment are negatively correlated in rural areas but that the opposite holds for those in urban areas. The relationship between children’s work activities and parents’ employment type also remains unclear in the literature.

This study contributes to our understanding of children’s work activity participation by testing the following hypotheses. First, we expect that the relationship between parents’ employment and children’s engagement in economic work and housework is ambiguous and depends on the relative strength of the income/added worker (complementary) effect and substitution effect. Second, we expect parents’ informal employment to be positively correlated with children’s work activities, and formal employment to be negatively correlated.

3. Data

We use the combined Labour Force Survey (LFS) and Time Use Survey (TUS), which were implemented between June and September 2009 in 75 provinces and the Bangkok Metropolitan Area (NSO, 2009). The LFS provides demographic characteristics of all household members and information on labour participation of people aged 15 and over. The TUS records the activities individuals aged 10 and above engage in, and its duration, using a 24-h diary divided into 10-min slots. These activities range from agriculture and subsistence production and paid employment to unpaid caregiving and domestic activities. The TUS also collects each respondent’s basic demographic characteristics.

The LFS sample was based on a stratified two-stage sampling method, while the TUS adopted a stratified three-stage sampling method. The first stage involves a random selection of 5,976 enumerated areas, and then a total of 79,560 households were chosen using a systematic sampling method in the second stage. The LFS interviewed all individuals from each of these chosen households. The TUS uses an additional random sampling method to select one member who is at least 10 years old from each household to complete a time-use diary (NSO, 2009; Yokying et al., 2016). Given that the surveys interviewed identical households, we merged the surveys using the household and individual characteristics available in both surveys. The combined dataset provides the employment and socio-economic characteristics of adult family members and data on the children’s time spent in work activities, which the LFS or TUS, alone, cannot provide.

We focus on children aged 10 to 18 years living with their parents in a couple of household who were interviewed for both surveys during a weekday and have non-missing household labour income data. The total number of children 10 to 18 years old who were interviewed for both surveys is 7,417. We exclude those who were not the child of the household head because of a lack of information about their parents’ characteristics. We exclude children in single-parent households, since children’s time allocation and single parents’ labour supply differ from those of children living with both parents. Given that most children (80.44%) were interviewed during a weekday and that children’s time use is likely to differ during weekends, we omit children with weekend time diaries. Lastly, we exclude children from households with missing total household labour income data. These exclusions reduce the analysis sample to 3,888 children.

The child is involved in housework and economic work if his or her time in the System of National Accounts (SNA) activities¹ and domestic and caregiving (non-SNA) activities, respectively, is at least 10 minutes per day. Table A5 in the appendix provides descriptive statistics on children’s participation rates and mean time in each activity. The parent is employed if s/he is identified by the LFS as such, even if the time spent during the interview period may be zero.² The parent works informally if s/he does not have any employment security, legal protection and employer-sponsored benefits. The opposite holds for those formally employed.

4. Parents’ employment and children’s work activities

4.1. Methodology

Our analysis starts from the premise that household labour supply decisions are interdependent and made simultaneously as part of the household’s resource allocation process. We first examine the individual and household characteristics affecting the likelihood that mothers are employed, fathers are employed, and children are engaged in each type of work activity. After estimating these decisions jointly, we rely on correlations among unobserved characteristics to determine whether the relationships between children’s work participation and each parent’s employment decisions are positively or negatively correlated. These relationships can be expressed in the following reduced-form equations:

(1) $M_i^{\ast }={\alpha }_1{W}_i+e_i$(1)

$M_i=1if{M}_i^{\ast }\ge 0,and{M}_i=0$ otherwise

(2) $F_i^{\ast }={\beta }_1{X}_i+u_i$(2)

$F_i=1if{F}_i^{\ast }\ge 0,and{F}_i=0$ otherwise

(3) $L_i^{\ast }={\gamma }_1{Y}_i+{\mu }_i$(3)

$L_i=1if{L}_i^{\ast }\ge 0,and{L}_i=0$ otherwise

(4) ${\mathrm{H}}_{\mathrm{i}}^{\ast }={\mathrm{b}}_1{\mathrm{Z}}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$(4)

$H_i=1if{H}_i^{\ast }\ge 0,and{H}_i=0$ otherwise

where $M_i^{\ast }$ and $F_i^{\ast }$ are latent variables capturing the probability of being employed for the mother and father of child i. $M_i$ and $F_i$ are the observed binary variables equal to one if child i’s mother and father, respectively, are employed. $L_i^{\ast }$ and $H_i^{\ast }$ are latent variables reflecting the probability that child i performs economic and household work, respectively, but we only observe whether or not child i is involved in economic work ($L_i)$ and housework ($H_i$). W_i and X_i are vectors of the individual and household characteristics of child i’s mother and father. The parent’s individual characteristics include age, age squared, parental age difference, education, and spousal education. Y_i and Z_i represent child i’s individual and household characteristics affecting the probability that s/he is engaged in economic work and housework, respectively. The child’s individual characteristics are age, age squared, and birth order. The household characteristics included in all the equations are household composition, a dummy for male-headed household, exogenous household income, a dummy for living in a rural area, minimum daily wage rate for each province,³ and region of residence dummies. $e_i$, $u_i$, ${\mu }_i,$and ${\epsilon }_i$ are the error terms, which are jointly distributed. The correlation between the two corresponding error terms is ρ.

Equations 1(1) $M_i^{\ast }={\alpha }_1{W}_i+e_i$(1) –4 are estimated using multivariate probit regressions. The unknown parameters of the above equations and correlation coefficients are treated as free parameters and estimated using a mixed-process maximum likelihood procedure (Roodman, 2011). Using this approach has several advantages. This model does not require a hierarchic specification of work participation decisions and allows the unobserved characteristics to be correlated, enabling us to treat mothers’, fathers’, and children’s involvement in work activities as jointly determined. The sign and statistical significance of the estimated correlation coefficients between two error terms provide information about whether each parent’s employment and children’s participation in each type of work activity is positively or negatively correlated.

We conduct separate estimations for girls and boys. Given that Thai gender norms place different societal expectations on boys and girls and that mothers typically serve as household managers and caregivers, mothers are expected to train their daughters to perform housework and assign them certain chores. Conversely, boys are not expected to do housework and are given fewer household responsibilities. The likelihood-ratio Chow test rejected the hypothesis that the coefficient estimates for boys’ and girls’ equations are equal, indicating that the gender-specific coefficient estimates yield useful information.

4.1.1. Child and parental characteristics

In our study, we also take into account children’s level of development and maturity, proxied by age, age squared, and birth order. Children are given more economic work and housework as they grow (Hofferth & Sandberg, 2001). First-born children are more likely to perform work activities, as they are put in charge of household tasks (Self, 2011). We include these children’s characteristics in our regression analyses alongside parents’ characteristics that include their lifecycle stage, captured by age and age squared, age gap and educational attainment. The parents’ likelihood of working increases as they become older (DeGraff & Levison, 2009), but it diminishes as they reach a much later lifecycle stage. The parental age gap is a proxy for each spouse’s relative household bargaining power. A larger age gap suggests that fathers are likely to take their breadwinner role more seriously and have a greater say in the mothers’ employment decisions (Mapraneet, 2009).

Furthermore, each parent’s education level impacts earning potential and can affect their employment decisions. Spousal education can, therefore, influence a person’s labour force status and employment type. Mothers with educated husbands are less likely to work since their husbands’ earnings can adequately maintain the families (Detboon et al., 2017). Better-educated mothers are likely to have spouses who are formally employed. Mapraneet (2009) and Detboon et al. (2017) explain that this dynamic maintains a male-dominant gender relation, which can be threatened by a woman’s higher educational achievement.

Given that fathers are household-head breadwinners, their characteristics (age, age squared, education attainment) along with parental age gap and the mothers’ education are incorporated as determinants of children’s participation in work activities. Children are more likely to be involved in work activities when they have an older father. Parental education can also influence attitudes towards children’s activities; educated parents’ children tend to spend more time on activities essential for development, such as reading and leisure (Hofferth & Sandberg, 2001; Wacharasin et al., 2013). Thus, children of better-educated parents are expected to be less involved in work activities.

4.1.2. Household characteristics

We also control for pertinent household characteristics. Household composition includes the number of females and males of similar, older, or younger age cohorts. Households with more young children are likely to have a higher demand for care and higher expenses, increasing the need for children to participate in work activities. The opposite is true for households with more working-age adults. Elderly household members can provide help with domestic chores and childcare, reducing the need for children to perform these chores. However, elders can increase the demand for care if they have limited functionality. Consequently, children may be asked to provide elderly care. Given that fathers are the breadwinners and mothers are household managers, having more children increases the fathers’ probability of working and decreases the likelihood that the mothers work. Employed mothers may be drawn to informal employment since it offers greater flexibility, allowing the mothers to cope with household responsibilities.

We use the logarithm of the exogenous household labour income of non-parent adult household members to capture household wealth.⁴ We also control for province-level minimum daily wages in the formal sector, a rural dummy, and regional-level location dummies reflecting unobservable differences across regions that can influence child, maternal, and paternal employment characteristics.

4.2. Results

4.2.1. Descriptive statistics

Table 1 presents the relevant characteristics of children and their parents. The mean age (14.1 years) of the children is below the legal minimum age of employment (15 years). The mean age difference between their parents is 3.3 years. Half of the children have mothers who did not attend school, while slightly more than half have fathers who have completed at least primary education. About 70% live with informally employed mothers. More than three-fourths (80%) reside with working fathers, and 60% have fathers who are informally employed. The average monthly household labour income is 5,837.3 Baht (around 176.4 USD), and more than half of these children live in rural areas.

Table 1. Descriptive statistics of individual and household characteristics

Characteristic	Mean	Standard Deviation
Children’s age	14.1	2.4
Children’s sex:
Girls	0.5
Boys	0.5
Rank of birth order	1.7	0.9
Mother’s age	47.5	12.1
Father’s age	48.6	10.8
Age gap (father–mother)	3.3	5.2
Mother’s education:
No schooling (base)	0.5
Primary school	0.2
Middle school	0.1
High school	0.1
Associate’s degree or above	0.1
Father’s education:
No schooling (base)	0.4
Primary school	0.2
Middle school	0.1
High school	0.1
Associate’s degree or above	0.1
Mother’s labour force status:
Not in the labour force/unemployed (base)	0.3
Employed	0.7
Mother’s employment type:
Formal (base)	0.3
Informal	0.7
Father’s labour force status:
Not in the labour force/unemployed (base)	0.2
Employed	0.8
Father’s sector of employment:
Formal sector (base)	0.4
Informal sector	0.6
Household composition:
Children 0–3 years old	0.2	0.4
Children 4–6 years old	0.1	0.4
Children 7–15 years old	1.2	0.8
Children 16–18 years old	0.5	0.6
Adults 19–59 years old	2.2	0.7
Elderly 60 years old and up	0.3	0.6
Exogenous labour income (Baht)	5,837.3	8,214.9
Minimum daily wages (Baht)	161.6	14.6
Location:
Urban (base)	0.5
Rural	0.5
Total	3,888

Table 2 compares the characteristics of children who are and are not involved in economic activities (see Columns 1 and 2) and housework (see Columns 3 and 4). Children involved in housework are slightly older than those who are not and are more likely to be girls. Working children spend less time on educational activities and have less-educated parents than non-working children. On average, non-working children live in households with a higher monthly labour income. Work activity participation is greater for children living in rural areas than for those living in urban areas.

Table 2. Selected characteristics associated with working and non-working children

Characteristic	Economic Work		Housework
	No Work	Work	No Work	Work
Children’s mean age	13.7	15.4	13.9	14.2
Children’s sex:
Girls	53.0%	34.4%	29.2%	59.2%
Boys	47.0%	65.6%	70.8%	40.8%
Mother’s education:
No schooling	48.0%	56.3%	43.7%	53.0%
Primary school	23.1%	29.3%	22.1%	25.6%
Middle school	9.2%	8.6%	10.2%	8.5%
High school	8.7%	3.6%	8.5%	7.1%
Associate degree or above	11.1%	2.3%	15.4%	5.9%
Father’s education:
No schooling	40.9%	55.9%	40.2%	46.5%
Primary school	20.7%	24.1%	17.4%	23.7%
Middle school	11.3%	8.4%	10.4%	10.9%
High school	11.3%	8.4%	11.6%	10.1%
Associate degree or above	15.7%	3.3%	20.5%	8.9%
Household composition:
Children 0–3 years old	13.2%	16.7%	12.1%	15.0%
Children 4–6 years old	12.6%	11.5%	12.5%	12.3%
Children 7–15 years old	39.9%	44.3%	41.9%	40.3%
Children 16–18 years old	11.6%	13.7%	13.3%	11.4%
Elderly 60 years old and up	26.6%	27.1%	25.4%	27.4%
Exogenous labour income (Baht)	8,059.8	5,367.8	8,187.3	7,068.2
Location:
Urban	54.1%	42.9%	57.4%	48.6%
Rural	46.0%	57.1%	42.6%	51.4%
Total	3,034	854	1,332	2,556

4.2.2. Econometric results

Table 3 reports the average marginal effects⁵ from the multivariate probit estimations for girls and boys. As hypothesized, Columns 1 and 2 show that educated parents’ daughters are less likely to be working. A greater number of working-age adults in the household reduces the probability of working for girls, and the number of elderly persons has a negative effect on girls’ housework participation.

Table 3. Multivariate probit estimates (average marginal effects) of children’s work activity involvement and parents’ employment by sex

	Girls				Boys
Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Economic Work	Housework	Mothers’ Employment	Fathers’ Employment	Economic Work	Housework	Mothers’ Employment	Fathers’ Employment
Child’s age	−0.035	0.133***			−0.047	0.130**
	(0.045)	(0.047)			(0.050)	(0.059)
Child’s age^2	0.002	−0.004**			0.004**	−0.004*
	(0.002)	(0.002)			(0.002)	(0.002)
Child’s birth order	0.020	0.039			0.002	−0.050
	(0.031)	(0.030)			(0.030)	(0.036)
Mother’s age			0.046***				0.047***
			(0.007)				(0.007)
Mother’s age^2			−0.001***				−0.001***
			(0.000)				(0.000)
Father’s age	−0.000	−0.005		0.016***	0.003	−0.009		0.027***
	(0.007)	(0.007)		(0.004)	(0.007)	(0.009)		(0.004)
Father’s age^2	−0.000	0.000		−0.000***	−0.000	0.000		−0.000***
	(0.000)	(0.000)		(0.000)	(0.000)	(0.000)		(0.000)
Age gap	0.002	0.001	−0.004**	0.009***	0.003*	−0.001	−0.002*	0.008***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.002)	(0.001)	(0.001)
Mothers’ primary school	0.015	−0.075***	0.102***	0.015	−0.002	−0.033	0.046	0.043*
	(0.024)	(0.028)	(0.030)	(0.021)	(0.029)	(0.035)	(0.030)	(0.023)
Mothers’ middle school	0.035	−0.114***	0.029	−0.023	−0.023	−0.059	0.010	0.064**
	(0.032)	(0.035)	(0.037)	(0.026)	(0.039)	(0.047)	(0.038)	(0.030)
Mothers’ high school	−0.073*	−0.110***	0.070*	0.063	−0.070	0.005	0.057	0.062*
	(0.040)	(0.037)	(0.042)	(0.039)	(0.047)	(0.053)	(0.043)	(0.034)
Mothers’ associate degree	−0.130**	−0.146***	0.138***	0.054	−0.131**	−0.141**	0.262***	0.002
	(0.052)	(0.040)	(0.047)	(0.036)	(0.054)	(0.055)	(0.051)	(0.031)
Fathers’ primary school	0.010	−0.009	−0.004	0.085***	−0.038	0.063	0.024	0.056**
	(0.026)	(0.030)	(0.031)	(0.024)	(0.030)	(0.038)	(0.031)	(0.023)
Fathers’ middle school	−0.039	−0.029	0.001	0.041*	−0.096***	0.066	0.012	0.008
	(0.032)	(0.034)	(0.036)	(0.025)	(0.037)	(0.044)	(0.036)	(0.025)
Fathers’ high school	0.020	−0.084**	0.003	0.025	−0.155***	0.034	0.008	0.020
	(0.032)	(0.034)	(0.038)	(0.027)	(0.040)	(0.046)	(0.039)	(0.028)
Fathers’ associate degree	−0.076*	−0.162***	−0.086**	0.088***	−0.214***	−0.087*	−0.076*	0.035
	(0.043)	(0.036)	(0.041)	(0.033)	(0.049)	(0.052)	(0.043)	(0.032)
Children 0–3	0.029	0.015	−0.116***	−0.039***	0.021	0.115**	−0.101***	−0.056***
	(0.035)	(0.038)	(0.022)	(0.014)	(0.037)	(0.045)	(0.021)	(0.014)
Children 4–6	−0.044	−0.025	−0.018	−0.018	0.004	0.055	−0.034	−0.036**
	(0.040)	(0.040)	(0.027)	(0.018)	(0.041)	(0.048)	(0.024)	(0.016)
Children 7–15	−0.014	−0.027	−0.032**	−0.013	0.026	0.047	−0.007	−0.010
	(0.028)	(0.026)	(0.013)	(0.009)	(0.028)	(0.033)	(0.012)	(0.008)
Children 16–18	0.018	−0.005	0.018	−0.005	0.034	−0.047	−0.007	−0.003
	(0.022)	(0.023)	(0.018)	(0.012)	(0.025)	(0.030)	(0.017)	(0.012)
Adults 19–59	−0.019*	−0.060***	0.001	0.013**	−0.005	−0.076***	−0.005	0.012*
	(0.011)	(0.011)	(0.011)	(0.006)	(0.012)	(0.014)	(0.012)	(0.007)
Adults 60 and up	−0.015	−0.040*	0.003	0.035**	0.052**	−0.033	0.002	0.015
	(0.022)	(0.023)	(0.024)	(0.015)	(0.023)	(0.028)	(0.023)	(0.014)
Male head	−0.006	0.042*	−0.015	0.159***	0.016	0.014	−0.024	0.151***
	(0.022)	(0.023)	(0.025)	(0.013)	(0.025)	(0.030)	(0.026)	(0.014)
Household income (log)	−0.012	−0.010	0.187	0.083	−0.013	−0.014	0.042***
	(0.009)	(0.007)	(7.159)	(8.534)	(0.010)	(0.010)	(0.014)
Minimum wage	0.000		−0.002*	−0.001*	−0.002**		−0.002**	−0.001
	(0.001)		(0.001)	(0.001)	(0.001)		(0.001)	(0.001)
Rural	−0.009	0.034*	0.069***	0.040***	0.059***	0.033	0.057***	0.027**
	(0.017)	(0.018)	(0.019)	(0.013)	(0.019)	(0.023)	(0.019)	(0.013)
${\rho }_{1,3}$	0.319***
${\rho }_{1,4}$	−0.017
${\rho }_{2,3}$	0.108**
${\rho }_{2,4}$	0.145**
${\rho }_{5,7}$					0.247***
${\rho }_{5,8}$					0.224***
${\rho }_{6,7}$					0.024
${\rho }_{6,8}$					0.095
Location dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	1,903	1,903	1,903	1,903	1,985	1,985	1,985	1,985

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.10. ρ is a correlation coefficient between child’s work activity participation equation and parent’s employment equation. Coefficient and standard error for the household income variable is missing due to low variation in household income among boys with fathers who are unemployed or not in the labour force.

Columns 5 and 6 indicate that educated parents, especially those with at least an associate’s degree, are less likely to have their sons involved in work activities. A greater number of toddlers increase the likelihood of doing housework for boys but not for girls. This may reflect a gradual shift in gender role perception in Thailand whereby housework is perceived as the responsibility of both genders (Wasikasin et al., 2008). A greater number of working-age adults lower the probability of boys performing housework. Although the effect is miniscule, a higher minimum wage reduces the probability of boys being involved in economic activities because it is likely to reduce the economic necessity of having boys take on a supplemental breadwinner role. Living in rural areas positively affects boys’ involvement in economic work, presumably in agriculture.

Table 3 shows the regression estimates of the determinants of employment for mothers (see Columns 3 and 4) and fathers (see Columns 7 and 8). While there are some differences in coefficient sizes, the results are similar for boys’ and girls’ parents. As mothers and fathers age, their probability of employment increases at a slower rate. Educated parents are more likely to be employed. A greater number of children, particularly toddlers, reduce mothers’ and fathers’ employment probability. A minimum daily wage increase in the formal sector slightly reduces the likelihood that parents work, reflecting the fact that more than half of the working population in Thailand is employed informally (ILO, 2015). Rural mothers and fathers are more likely to be employed than their urban counterparts because rural areas provide farm employment opportunities that absorb unskilled labourers and those unable to find non-farm employment (ILO, 2015).

The correlation between unobservable factors is reported at the bottom of Table 3. The large and statistically significant correlation coefficients confirm that household labour supply decisions are made simultaneously. For girls, we find evidence of the income/added worker effect on the relationship between their involvement in economic activities and mothers’ employment. The substitution effect dominates the income effect in the association between girls’ involvement in housework and parents’ employment. The positive correlation between mothers’ employment and girls’ economic work is particularly strong, as shown by the large magnitude of the correlation coefficient (ρ_1,3 = 0.319). One possible explanation is that working mothers are likely to be self-employed and have their daughters work as unpaid family workers in their businesses and perform housework. However, the correlation between the unobservable factors affecting fathers’ employment and girls’ economic work (ρ_1,4) is insignificant.

For boys, the correlation coefficient between the unobservable factors affecting their involvement in economic work and mothers’ employment (ρ_5,7) is positive and statistically significant. It is also positive and significant in the case of fathers’ employment and boys’ economic work (ρ_5,8). These results imply that unobservable factors, such as household income or consumption shocks, that increase the likelihood of parents being employed also increase the probability that sons work. Thus, we find evidence of dominant income/added worker effects in the relationship between parents’ employment and boys’ participation in economic work. On the other hand, the correlation coefficients between each parent’s employment and boys’ participation in domestic and care work are statistically insignificant.

5. Parents’ employment type and children’s work activities

5.1. Methodology

We further examine whether parents’ employment type is correlated with children’s involvement in work activities. As discussed in Section 2, we assume that parents choose their respective employment type when they decide whether their children need to work. We focus on children with employed parents and replace reduced-form Equations 1(1) $M_i^{\ast }={\alpha }_1{W}_i+e_i$(1) and 2(2) $F_i^{\ast }={\beta }_1{X}_i+u_i$(2) above with Equations 5(5) $M{S}_i^{\ast }={\delta }_1{W}_i+{\tau }_i\hspace{0.17em}$(5) and 6(6) $F{S}_i^{\ast }={\theta }_1{X}_i+{\nu }_i$(6) below. We then estimate Equations 3(3) $L_i^{\ast }={\gamma }_1{Y}_i+{\mu }_i$(3) –6 jointly using multivariate probit regressions:

(5) $M{S}_i^{\ast }={\delta }_1{W}_i+{\tau }_i\hspace{0.17em}$(5)

$M{S}_i=1ifM{S}_i^{\ast }\ge 0,andM{S}_i=0$ otherwise

(6) $F{S}_i^{\ast }={\theta }_1{X}_i+{\nu }_i$(6)

$F{S}_i=1ifF{S}_i^{\ast }\ge 0,andF{S}_i=0$ otherwise

where $M{S}_i^{\ast }$ and $F{S}_i^{\ast }$ are latent variables reflecting the likelihood that child i’s mother and father, respectively, are informally employed. $M{S}_i$ and $F{S}_i$ are the observed dummies and are equal to one if child i’s mother and father work in informal employment. On the other hand, these observed dummies are equal to zero in the case of child i’s mother and father work in formal employment. W_i and X_i are vectors of the relevant characteristics affecting the employment type of the working mother and father of child i. We use the same individual and household characteristics used in Equations 1(1) $M_i^{\ast }={\alpha }_1{W}_i+e_i$(1) and 2(2) $F_i^{\ast }={\beta }_1{X}_i+u_i$(2) to estimate Equations 5(5) $M{S}_i^{\ast }={\delta }_1{W}_i+{\tau }_i\hspace{0.17em}$(5) and 6(6) $F{S}_i^{\ast }={\theta }_1{X}_i+{\nu }_i$(6) . ${\tau }_i$ and ${\nu }_i$ are jointly distributed disturbances. Estimating Equations 3(3) $L_i^{\ast }={\gamma }_1{Y}_i+{\mu }_i$(3) –6 jointly yield correlation coefficients between each pair of error terms, providing insight into whether the informal or formal employment of each parent is positively or negatively correlated with children’s involvement in each type of work activity.

The correlations for informal employment can be used to infer the relationship between parents’ formal employment and children’s work activity participation, since the outcomes of Equations 5(5) $M{S}_i^{\ast }={\delta }_1{W}_i+{\tau }_i\hspace{0.17em}$(5) and 6(6) $F{S}_i^{\ast }={\theta }_1{X}_i+{\nu }_i$(6) are the flip side of the latter. For example, if working in informal employment is positively correlated with children’s work activity participation, then working in formal employment is negatively correlated. On the other hand, if working in informal employment is negatively correlated with child’s participation in work activity, then working in formal employment is positively correlated.

5.2. Econometric results

Table 4 presents the regression estimates of the determinants of parents’ informal employment (see Columns 3 and 4 for girls’ parents, and Columns 7 and 8 for boys’ parents) and the relationship between children’s engagement in work activities and parents’ informal employment. As expected, better-educated parents are less likely to work in informal employment, implying that they are more probable to be formally employed. Household wealth decreases the likelihood of being informally employed and increases the probability of being formally employed. A higher minimum daily wage in the formal sector decreases the likelihood of working in informal employment for mothers and fathers. It, in turn, increases the probability of working formally for the parents. This result is consistent with Hohberg and Lay (2015) finding of positive effects of minimum wage increases on formal employment. Residing in rural areas has a positive impact on fathers being employed informally, indicating that it has a negative effect on fathers’ formal employment.

Table 4. Multivariate probit estimates (average marginal effects) of children’s work activity involvement and parents’ informal employment by sex

	Girls				Boys
Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Economic Work	Housework	Mothers’ Informal Employment	Fathers’ Informal Employment	Economic Work	Housework	Mothers’ Informal Employment	Fathers’ Informal Employment
Child’s age	−0.028	0.168***			−0.058	0.124*
	(0.061)	(0.055)			(0.064)	(0.072)
Child’s age^2	0.002	−0.005***			0.004*	−0.004
	(0.002)	(0.002)			(0.002)	(0.003)
Child’s birth order	0.017	0.001			0.024	−0.020
	(0.042)	(0.038)			(0.039)	(0.045)
Mother’s age			−0.037**				0.011
			(0.016)				(0.012)
Mother’s age^2			0.000***				−0.000
			(0.000)				(0.000)
Father’s age	−0.008	−0.010		−0.013	−0.015	−0.005		−0.012
	(0.012)	(0.012)		(0.015)	(0.012)	(0.015)		(0.014)
Father’s age^2	0.000	0.000		0.000	0.000	0.000		0.000
	(0.000)	(0.000)		(0.000)	(0.000)	(0.000)		(0.000)
Age gap	0.001	0.003	0.005**	0.004	0.002	−0.002	0.003	−0.004
	(0.002)	(0.003)	(0.002)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Mothers’ primary school	0.011	−0.123***	−0.016	−0.033	0.013	−0.014	0.023	0.054
	(0.031)	(0.034)	(0.035)	(0.036)	(0.035)	(0.043)	(0.035)	(0.038)
Mothers’ middle school	0.019	−0.131***	−0.074*	−0.044	−0.015	−0.028	−0.036	0.002
	(0.042)	(0.043)	(0.044)	(0.046)	(0.047)	(0.057)	(0.045)	(0.049)
Mothers’ high school	−0.069	−0.140***	−0.076	−0.078	−0.072	0.013	−0.084*	0.022
	(0.049)	(0.045)	(0.047)	(0.049)	(0.056)	(0.063)	(0.049)	(0.054)
Mothers’ associate degree	−0.161**	−0.208***	−0.374***	−0.120**	−0.170***	−0.160**	−0.479***	0.025
	(0.063)	(0.046)	(0.048)	(0.053)	(0.065)	(0.066)	(0.050)	(0.058)
Fathers’ primary school	−0.017	−0.014	−0.021	0.005	−0.044	0.061	0.007	−0.014
	(0.032)	(0.034)	(0.035)	(0.036)	(0.035)	(0.043)	(0.035)	(0.038)
Fathers’ middle school	−0.053	−0.051	−0.032	−0.071*	−0.109**	0.065	−0.006	−0.126***
	(0.040)	(0.039)	(0.042)	(0.042)	(0.044)	(0.052)	(0.043)	(0.045)
Fathers’ high school	−0.003	−0.091**	0.007	−0.118***	−0.159***	0.065	0.039	−0.189***
	(0.039)	(0.038)	(0.042)	(0.041)	(0.046)	(0.055)	(0.045)	(0.046)
Fathers’ associate degree	−0.116**	−0.134***	−0.053	−0.389***	−0.215***	−0.075	0.003	−0.412***
	(0.054)	(0.043)	(0.048)	(0.047)	(0.059)	(0.063)	(0.052)	(0.053)
Children 0–3	0.078	0.035	−0.002	−0.108***	0.004	0.113*	0.022	−0.028
	(0.049)	(0.048)	(0.034)	(0.034)	(0.050)	(0.060)	(0.034)	(0.035)
Children 4–6	−0.020	0.062	−0.028	−0.036	0.008	−0.027	0.032	0.063*
	(0.055)	(0.051)	(0.035)	(0.036)	(0.053)	(0.061)	(0.033)	(0.036)
Children 7–15	0.009	−0.010	0.001	0.002	0.017	0.031	−0.015	−0.004
	(0.039)	(0.034)	(0.017)	(0.017)	(0.036)	(0.041)	(0.016)	(0.017)
Children 16–18	0.024	0.005	0.003	0.028	0.047	−0.032	0.008	0.001
	(0.030)	(0.029)	(0.022)	(0.023)	(0.032)	(0.037)	(0.022)	(0.024)
Adults 19–59	−0.006	−0.045***	0.076***	0.068***	−0.008	−0.039*	0.038*	0.018
	(0.017)	(0.017)	(0.023)	(0.022)	(0.018)	(0.022)	(0.021)	(0.021)
Adults 60 and up	−0.010	−0.041	0.029	0.068*	0.050	−0.043	0.083**	0.043
	(0.032)	(0.031)	(0.037)	(0.038)	(0.032)	(0.039)	(0.035)	(0.037)
Male head	0.004	0.041	0.051*	0.008	0.003	0.013	−0.030	−0.004
	(0.029)	(0.027)	(0.029)	(0.031)	(0.033)	(0.037)	(0.031)	(0.033)
Household income (log)	−0.018**	−0.016**	−0.063***	−0.059***	−0.017*	−0.023**	−0.038***	−0.038***
	(0.009)	(0.007)	(0.009)	(0.009)	(0.010)	(0.011)	(0.009)	(0.010)
Minimum wage	0.000		−0.004***	−0.002**	−0.003**		−0.002**	−0.001
	(0.001)		(0.001)	(0.001)	(0.001)		(0.001)	(0.001)
Rural	−0.016	0.014	0.028	0.050**	0.041*	0.021	0.025	0.091***
	(0.022)	(0.023)	(0.024)	(0.025)	(0.024)	(0.029)	(0.024)	(0.025)
${\rho }_{1,3}$	0.417***
${\rho }_{1,4}$	0.229***
${\rho }_{2,3}$	0.292***
${\rho }_{2,4}$	0.384***
${\rho }_{5,7}$					0.187***
${\rho }_{5,8}$					0.233***
${\rho }_{6,7}$					0.254***
${\rho }_{6,8}$					0.354***
Location dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
	1,196	1,196	1,196	1,196	1,282	1,282	1,282	1,282

Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.10. ρ is a correlation coefficient between child’s work activity participation equation and parent’s informal employment equation.

The correlation coefficients in Table 4 confirm our hypothesis that parents’ informal employment is positively correlated with children’s participation in work activities. The reverse is true for formally employed parents. Specifically, the correlation coefficient between involvement in each work activity and each parent’s informal employment is positive for both girls and boys, signifying that the parent’s formal employment and their children’s work activity participation are negatively correlated. The magnitude of the correlations is particularly large between mothers’ informal employment and girls’ economic work participation (ρ_1,3 = 0.417) and also between fathers’ informal employment and children’s participation in housework (ρ_2,4 = 0.384 for girls and ρ_6,8 = 0.354 for boys). Thus, the income/added worker (substitution) effects dominate the relationship between parents’ informal (formal) employment and children’s involvement in economic activities. In terms of housework, the substitution (income) effect prevails among children whose mothers or fathers are informally (formally) employed. The positive relationship is consistent with social expectations that children should help at home and in family enterprises (Naksung & Piansoongern, 2018). Other factors that motivate parents to work informally can also propel parents to have their children work as unpaid family workers or paid labourers to augment household labour and income (ILO, 2013).

That said, our empirical investigation is based on a reduced-form model estimating the likelihood that the mother, father, and child are employed. As with any reduced-form models, it does not uncover the true model generating the data and does not give much information on behavioural responses to policy changes.

6. Conclusion

This study seeks to understand how mothers’ and fathers’ labour force participation correlates with children’s economic and household work participation in the context of Thailand. It also explores how parents’ employment type is associated with children’s involvement in work activities. In line with the findings of DeGraff and Levison (2009), Dayıoğlu (2008), and Self (2011), our results show that mothers’ employment is positively correlated with girls’ and boys’ economic work participation. Fathers’ employment is also positively associated with boys’ involvement in economic activities. On the other hand, girls’ participation in domestic and care work is positively related to both mothers’ and fathers’ employment. No correlations are found between boys’ housework and parents’ employment. Girls’ and boys’ participation in work activities is positively (negatively) associated with parents’ informal (formal) employment. These findings are largely robust to specifications of alternative measures of children’s work activity participation.⁶

The results indicate that policies aimed at increasing adults’ labour force participation can increase the need for children to participate in economic work and housework. Establishing social protection programs and affordable childcare services may allow working children to spend more time on activities beneficial to their development. For example, providing childcare subsidies to informally employed parents with dependents could help ease their consumption costs and reduce the need for children to engage in economic activities and domestic and care tasks. Promoting gender equity and addressing gender discrimination in labour markets through gender-aware labour policies can help mothers gain access to formal employment and better pay. Hence, policies aimed at reducing poverty or stimulating growth by increasing labour force participation among women and men should incorporate complementary programs designed to mitigate the negative spillover effects of these policies on children and ensure that child welfare is not compromised.

Acknowledgments

The authors would like to thank Natalia Radchenko, Thomas Hungerford, Yana van der Meulen Rodgers, Jefferson Fox, Nicholas Reksten, and Minja Choe for their insightful feedback. They would also like to thank Thailand's National Statistical Office for providing access to the merged LFS-TUS data. They are also grateful to the participants of the panel sessions in the Pathways to Gender Equality Conference held in November 2018 at American University, Washington DC and in the Midwest Economics Association Annual Meeting held in March 2019 as well as the attendees of the East-West Center’s Noon Seminar held in February 2019 for their helpful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Notes on contributors

Phanwin Yokying

Phanwin Yokying is a fellow at the East-West Center. Her research focuses on development and gender issues affecting the livelihoods of females and males in developing countries.

Maria S. Floro

Maria S. Floro is a Professor of Economics at American University in Washington DC. She also leads the Care Work and the Economy Project, which aims to develop gender-sensitive macroeconomic models for policy analysis in order to reduce gender gaps in economic outcomes.

Notes

1. These activities include working for corporations, non-profit institutions, and government; engaging in household primary or non-primary production activities or construction work; and working for household providing services for income.

2. This is because the parent may not be working on the interview day.

3. We do not use minimum daily wages as a determinant of children’s participation in housework because we assume that they do not affect children’s housework participation.

4. We assume that these members’ employment characteristics are not jointly determined with those of the mother, father, and child.

5. It is the average change in probability of the dependent variable as a function of a change in an explanatory variable. They are calculated by estimating the derivative of the conditional mean function with respect to the explanatory variable for each observation using other covariates as they are observed, then by calculating the sample average of individual marginal effect to obtain the average marginal effect.

6. Given that longer work hours may compete with the time children have for educational and leisure activities, we used three alternative thresholds for each work activity: working for more than 1 hour, 1.5 hours, and 2 hours per day. The sensitivity analyses results are available upon request.