Draft: Not to be quoted or cited Parenthood, Patriarchy and Penalty: Evidence from Indian Labour Market
SukanyaSarkhel1 St. Xavier’s College, Kolkata
Abstract
This paper estimates the extent to which parenthood influences wage income of working mothers and fathers respectively. In this context we try to assess the influence of family culture in explaining this gap using India Human Development Survey (IHDS) data. We argue that mothers from patriarchal family culture will work and earn less as compared to fathers and this will sharpen the wage gap in India. Using Oaxaca-Blinder method of decomposition for mothers and non-mothers we find that children have significant negative influence on earning of working mothers as compared to married females. The Oaxaca-Blinder decomposition of wage measures significant wage gap of 0.41 between married female without kid and mothers with one child. 28 percent of this gap is unexplained. On the other hand fathers are earning significantly more than married male coworkers and the gap is 0.33 and major portion of this is explained, which conforms that males become more productive after becoming father as breadwinners for the family and this fatherhood fetches more return for male. It appears that motherhood entails a wage “penalty” in the labour market and more importantly fatherhood is associated with a
“bonus” in terms of higher wage premium indicating differential impact of parenthood within a family.
JEL Keywords: J16, J23, J24, J31, J71
1. Introduction
Research shows that motherhood reduces women earning as they keep more time and energy for baby care and other household responsibilities. This paper captures the wage gap of working mothers and married women and working fathers with married men to see the extent to which motherhood and fatherhood influences wage income of working mothers and fathers respectively. We conjecture that there exist motherhood penalty and fatherhood bonus in Indian labour market and family culture might have some role in explaining this gap. That is mothers from patriarchal family culture will work and earn less as compared to fathers and this will sharpen this gap in India. Firstly motherhood makes women less productive as they reduce their work participation and effort due more time allocation in family and child care. Secondly, employers are biased against women with small children as they would be less efficient as they may take more leave etc. Fathers on the other hand become more efficient as they want to earn more for the children. As a result motherhood is likely to reduce women earning, i.e., penalty for mothers and fatherhood would increase male earning, i.e., fatherhood bonus of the otherwise similar group.
Research on motherhood penalty generally focuses on the pure effect of child on mothers wage ceteris peribus. But everything changes after a child birth. One change especially salient to labor economists is that many mothers exit the work force. Absences from the labor market are likely to reduce wages because general and firm specific skills depreciate and workers lose rents associated with good job matches. Low-skilled workers may be less vulnerable to such earnings erosion, since they have less human capital and their wages reflect less rent. If so, these workers may escape a motherhood wage penalty. On the other hand highly skilled women to experience the largest penalties for exiting the labor force to care for their children (Deborah J. Anderson, 2003) . . Korenman &Neumark, 1992), while yet others have found a modest wages premium for married women (Budig& England, 2001;
Waldfogel, 1997). Research examining the mechanisms underlying the wage penalty or premium for married women has been very limited (for an exception, see Killewald& Gough, 2013) and as almost all studies to date have been conducted in the United States, we know little to nothing about the relation between marriage and female wages in other countries. Not to mention what cross-national variations exist and how to explain them. Research among
men has found substantial cross-national variations in the effect of marriage on wages (De Hoon, Keizer, & Dykstra, 2013; Schoeni, 1995) and the same might be true for women.
A lab experiment and survey research found substantial wage gap between mothers and fathers.
Study indicates that employers are biased against mothers and thus they are penalized on perceived competence and recommended starting salary. Men were not penalized for, and sometimes benefited from, being a parent. The audit study showed that actual employers discriminate against mothers, but not against fathers. Mothers experience disadvantages in the workplace in addition to those commonly associated with gender. For example, two recent studies find that employed mothers in the United States suffer a per-child wage penalty of approximately 5%, on average, after controlling for the usual human capital and occupational factors that affect wages (Budig and England 2001; Anderson, Binder, and Krause 2003). A study measures the influence of unplanned children on unwed mothers labour force participation as compared to unmarried women. They found large short-term effects of un- planned births on labor-force participation, poverty, and welfare recipiency among unwed mothers, but not among married mothers (Grogger, 1994). Gash (2009) examines the extent of and the mechanisms behind the penalty to motherhood in six European countries. Each country provides different levels of support for maternal employment allowing us to determine institutional effects on labour market outcome. While mothers tend to earn less than non-mothers, the penalty to motherhood is considerably lower in countries with policy support for working mothers. The paper establishes the United Kingdom and West Germany to have the least policy support for working mothers as well as the largest penalties to motherhood.
Our conjecture is that mothers from patriarchal families are earning less than fathers. Mothers sacrificing their career growth for household duties and children are more likely to come from families with stronger patriarchal values. This allows us to link this work to a newly emerging field of culture and economic outcome which are pioneered by papers looking at the relationship between trust and trade (Guiso 2004), culture and effort (Ichino 2000}, religion and growth (Becker 2009, Weber, Tabellini 2010). A section of this strand of the literature also looks at the relationship between culture and economic outcomes in the historical context (Greif 1993, Botticini 2005).
Many studies reported the existence of such gap in the context of many countries. This present study particularly focuses on wage gap between females and males because of parenthood in the context of Indian labor market. This work is based on the Indian Human Development Survey (IHDS) data collected by the National Council of Applied Economic Research (NCAER) in 2004-05. The rest of the paper is organized as follows: section 2 describes the basic methodology of decomposition, section 3 describes the data and the construction of suitable variables for the purpose of our analysis, section 4 presents the major results and extends discussion of the same;
finally, section 5 concludes the paper by indicating the scope for further research.
2. Methodology: Decomposition of wage gap
There is vast literature on measuring gender pay gap that conforms significant amount of pay gap exists all over the world. Discrimination at the workplace is created by the employer’s bias against one group of employees Given that we cannot directly observe employers' attitude towards women and subsequent discrimination, usually gender wage discrimination is measured by measuring the occupation specific wage differential between male and female which cannot be explained by any of the observables such as education, experience etc. Measuring discrimination is an empirical challenge because existence of gender wage gap does not necessarily indicate discrimination and cultural bias against women. For example, a woman worker may get lesser wage than his male colleague just because she is less skilled. If the gap is due to any observable economic factor related to labor productivity and efficiency, then that can be explained by the regression analysis. However, if it is due to some unobserved socio-cultural influence like the social expectation regarding the role of women within the household and outside then the regression analysis will club these influences under unexplained residuals, unless proper control can be designed to account for their existence. So, one need to establish first the possible existence of such non-economic socio-cultural factors influencing the pattern of female labor force participation and then the subsequent analysis would be more insightful. The technique proposed to do this is known as the Oaxaca Blinder decomposition (Oaxaca 1973, Blinder 1973). This technique implicitly assumes that if the gender wage gap cannot be explained by any of the observed characteristics of the workers it must be because of the cultural bias of the employers. However, the unexplained wage gap may also arise if women are less
productive than their male counterpart and the productivity difference can only be observed by her office supervisor. If this explanation is true, gender wage gap that cannot be explained by observables, does not necessarily indicate gender discrimination in workplace. Such effort differential may be the result of family norms which require the women to take care of household chores, raise kids leaving them with less time and energy to put high effort in their work places.
To see how wage gap of these groups is explained by the explanatory variables and how much of this gap is unexplained, we use decomposition analysis of Oaxaca (1973), Blinder (1973).
We consider two groups working males ,M and working females, F and the outcome variable is their wage income. Here we consider log hourly wage, w, and a set of predictors,Xs are human capital indicators, education, experience, age, cultural indicator as patriarchy of the family.
We measure how much of the mean income difference, E(w) denotes expected value of the outcome variable, male female earning gap, is accounted for by the group differences (male- female difference) in the predictors. Male-Female Mean income gap :
𝐺 =𝐸 𝑤!"#! −𝐸 𝑤!"#$%! …..(1)
Based on linear model:
𝑤! =𝑤!′𝛽!+𝜖!, 𝐸 𝜖! =0, 𝑙= 𝑛𝑜𝑘𝑖𝑑,𝑤𝑖𝑡ℎ𝑘𝑖𝑑 ……(2)
Here w is the vector containing predictors and a constant, 𝛽 contains slope parameters and the intercept, and 𝜖, the error. The mean outcome difference can be expressed as difference in the linear prediction at the group specific means of the regressors.
𝐺 =𝐸 𝑤!"#! −𝐸 𝑤!"#$%! =𝐸 𝑥!"#$%′ 𝛽!"#$%−𝐸 𝑥!"#′ 𝛽!"#…….(3)
Since 𝐸 𝑤! =𝐸 𝑥!′𝛽! +𝐸 𝜖! = 𝐸(𝑥!)′𝛽! ; where 𝐸 𝜖! =0,𝐸 𝛽! =𝛽! Rearranging eq3: For mothers (m) group and married females (f) Group separately:
𝐺!= 𝐸 𝑥! −𝐸(𝑥!) ′𝛽!+𝐸 𝑥! ′ 𝛽!−𝛽! + 𝐸 𝑥! −𝐸 𝑥! ′ 𝛽!−𝛽! 2…(4a)
Gm= 𝐸 𝑥! −𝐸(𝑥!) ′𝛽!+𝐸 𝑥! ′ 𝛽!−𝛽! + 𝐸 𝑥! −𝐸 𝑥! ′ 𝛽!−𝛽! ….(4b) Gf = E + C + I
E = Expected change in female income if female group had male predictors;
C = Expected change in female outcome if female group had male coefficient;
I = Expected change in income;
3. Data
We use the India Human Development Survey Data (IHDS) of 2004-05 which is a database formed through a survey of 41554 households in 1503 villages and 971 urban areas for 35 Indian states and union territories conducted by Indian Council of Applied Economic Research (NCAER), New Delhi and University of Maryland. The survey consists of two parts, household questionnaire with household characteristics on demography, health, education, income, work, occupation, production, consumption, assets, social capital, fertility, children schooling etc. and individual questionnaire with work, income, gender relation, fertility decision, marriage practices, mass media, reading, writing skill etc.
In order to see the impact number of children on income of male and female workers we merged the household database with individual level information. The merged database thus pairs the individual information with household level information. Here, we attempt to unravel the influence of household level information like family structure on decision to participate in the labour market for individuals. As discussed in the theoretical section, work effort is likely to depend on the cultural aspects and our main hypothesis is work participation is inversely related with the degree of patriarchies in the family. The IHDS data consists of individual information about total hours of work in a year. This is given by a binary coded variable WORKANY that
2Winsborough and Dickinson 1971; Jones and Kelley 1984; Daymont and Andrisani 1984
takes a value one if the individual is working more than 240 hours in a year and 0 otherwise.
This has been used as an indicator for workforce participation and effort in the labour market.
We use age of the respondent(AGE), education(EDU), households total income (INCOME),IHDS measures total income of households summed across fifty components of income including wages and salaries, property income, net business income, farm income etc.
INCOME5 measures the distribution of household income in different quintiles, household asset (HASSET) counts the possession of number of valuable assets, that includes 30 items as goods and housing , sex(M, dummy coded 1 for male 0 otherwise ) , caste (caste_dum) 3variable grouped as low caste and high caste. In this we follow the specification of the usual earning equation.
Table: 3.1: Characteristics of the sample households in IHDS (2004-05)
Source :Based on authors calculation on IHDS(2005) data
3 Caste dummies have been incorporated as there were significant variation in workparticipation and patriarchy across caste groups. due to the preponderance of muslim households in the sample(12 percent as against 22 percent of high class), we have considered them as a separate social category as we treat dalit, adibashi and OBC as lower caste.
Household Characteristics Mean Value (Standard Deviation)
Rural Urban
Assets of the Household (HASSET) 10.25 (5.30) 16.48 (4.84)
Household Size (NPERSONS) 6.65 (3.29) 5.81(2.61)
Total Income of the households (INCOME) 48399.83 (81502.1) 79296.45 106978.5)
No. of children (NCHILD) 2.39 (1.90) 1.83 (1.50)
Highest Education of Adult (HEDM5) 6.83 (4.88) 9.81 (4.54) Highest Education of female (HHED5F) 3.61 (4.46) 6.96 (5.23) Highest Education of male (HHEDM) 6.54 (4.83) 9.34 (4.65) No. of married female in household
(NMARRIEDF)
1.52 (0.90) (0.75)
No. of married male in the household (NMARRIEDM)
1.45 (0.90) 1.31 (0.75)
Age of the respondent (RO5) 27.11 (19.67) 27.85 (17.84)
Education level of the respondent (ED5) 3.87 (4.26) 6.55 (5.04)
Dependent variable
We use different categories of annual earning of individuals from the IHDS dataset and constructed log hourly wage rate as our dependent variable.
Independent Variables
We use a set of explanatory variables from IHDS dataset that may influence individuals earning.
They are : Total number of children borne (FH6) as child dummy , 0 for no child, 1 for I child and 2 for two or more children, desire for more children, education (ED5) as education dummy for different levels, age of the individual, experience, constructed by age square, and the household culture as the index of patriarchy4 (PI), constructed from the IHDS dataset where they collect information about the most say of respondent in different household decisions.
Table 3. 2: Variables used from IHDS (2004-05)
Name of the Variables
IHDS Code Description Gender
(female,male)
RO3 Dummy: 0=female, 1=male
Age RO5 Years
Marital Status RO6 Spouse absent 0, married 1, single 2, widowed 3, separate/divorced 4, no gauna 5
Area URBAN Dummy: 0=Rural, 1=Urban
Annual Earning WS8ANNUAL Total annual earning
Hourly wage total WS8HOURLY Measured in rupees per year Hours of
Work/year
Wrkp1 Total hours of work per year
Work WORKANY If working yours per year is more than 240 then
4 We use the index of patriarchy, constructed by Basu (2015). The decision related information like who is having the most say in the family regarding cooking, marriage, fertility, child illness, expenditure, child marriage etc. are coded with 0 for women taking the decision and 1 for male members taking the decision. Combining six most say variable we computed an index of patriarchy in the line of Human Development Index4. Here the index can take a minimum value of zero where women take all the decisions and in case male members have the most say the index takes a value of 1 i.e. the maximum value. Thus, higher the value of the index the higher is the degree of patriarchy and lower is the autonomy of women. The indices are constructed follows: In case of autonomy index as we have taken six criterions the most the households can score in this regard is 6 and the minimum level is zero. Our proposed autonomy index would thus look like 𝐿𝑒𝑎𝑠𝑡_𝑠𝑎𝑦!=!!!!!!!. Where 𝜇! is the observed score of the household in terms of autonomy 𝜇 is the maximum value (i.e.6) and the minimum value (i.e. 0) is denoted by 𝜇.The other class of indices constructed in the same way would move in the opposite direction of patriarchy so we deduct them from unity.
participation decision
WORKANY=1, otherwise 0 Experience Age_sq Constructed Variable: age squre Educational
Attainment ED5
(individual) in years
Years of education (0 to 15 years)
Log Hourly wage L_hrw Constructed from the hourly wage total/yr Total No. of
Children borne
FH6_D Dummy : 0=no kid, 1=1 kid, 2=2 or more kids
4. Result
The graph shows distribution of log hourly wage earning of mothers and married without children. The mean eaning of mothers is less than that of married women.
Graph4.1: Log hourly wage of married women with and without children
We decompose mean earning of two groups of females, one with mothers’ and the other group is married women without kids. Table 4.1 reports the Oaxaca-Blinder decomposition that shows there exists significant amount of wage gap among mothers and non mothers in the labour
0.51
0 2 4 6 0 2 4 6
Nchildren=0 Nchildren>=1
Density
l_whr
Graphs by RECODE of FH6 (EH26 18.6 Total N children borne)
market of 0.41 and 0.42 for rural and urban areas respectively. Large share of this difference, 0.28 is being captured by the coefficient, which is called unexplained part or discrimination. This indicates that pay gap between married females without kid and married with kids might be due to discrimination against mothers.
Table 4.1: Oaxaca Decomposition of log hourly female wage for mother and potential mother
L_wghr RURAL URBAN
Married Females without children
1.92*** 2.61***
Mothers 1.51*** 2.20***
difference 0.41*** 0.42***
endowments 0.09*** 0.16***
coefficients 0.28*** 0.28***
interaction 0.03*** -0.02***
N 9467 2487
Group 1: Married without children (0), Group 2: Married with children (1 and more)5
Considering family culture as one of the explanatory variables we decompose the wage equation for the same groups. table 4.2 reports the decomposition result with patriarchy. Result reflects interesting fact. Pay differential between the groups expanded to 0.47 from 0.42 for urban area, and major portion of 0.32 is unexplained, which might be due to discrimination.
5 The results from decomposition are presented using Blinder's (1 973) original formulation of E, C, U and D. The endowments (E) component of the decomposition is the sum of (the coefficient vector of the regressors of the high-wage group) times (the difference in group means between the high-wage and low-wage groups for the vector of regressors). The coefficients (C) component of the decomposition is the sum of the (group means of the low-wage group for the vector of regressors) times (the difference between the regression coefficients of the high-wage group and the low-wage group). The unexplained portion of the differential (U) is the difference in constants between the high-wage wage and the low-wage group. The portion of the differential due to discrimination is C + U. The raw (or total) differential is E + C + U. The unexplained component is the difference in the shift coefficients (or constants) between the two wage equations. Being inexplicable, this component can be attributed to discrimination. However, Blinder also argued that the explained component of the wage gap also contains a portion that is due to discrimination.
To examine this Blinder decomposed the explained component into: (1 ) the differences in endowments between the two groups, "as evaluated by the high-wage group's wage equation"; and (2) "the difference between how the high-wage equation would value the characteristics of the low- wage group, and how the low-wage equation actually values them". Blinder called the first part the amount "attributable to the endowments" and the second part the amount "attributable to the coefficients", and he argued that the second part should also be viewed as reflecting discrimination: "[this] only exists because the market evaluates differently the identical bundle of traits if possessed by members of different demographic groups, [and] is a reflection of discrimination as much as the shift coefficient is." Conventionally, the high-wage group's wage structure is regarded as the "non-discriminatory norm", that is, the reference group. The average endowment differences are now weighted by the high-wage workers' estimated coefficients, and the coefficient differences are weighted by the mean characteristics of the low-wage workers. One can also do the reverse.
Table 4.2: Decomposition of log hourly female wage across mother and potential mother (With family culture)
Overall Rural Urban
Married Females without children
1.91*** 2.65***
Mothers
1.51*** 2.18***
Difference 0.40*** 0.47***
Endowments 0.09*** 0.12***
Coefficients 0.26*** 0.32***
interaction 0.06*** 0.04***
N 8368*** 2072
Group 1: Married without children (0), Group 2: Married with children (1 and more)
Having said that children have negative influence on mothers earning, we want to measure the extent to which it influence mothers earning. In order to measure that we take children dummy as 0 for no child, 1 for only one child and 2 for more than one children. The data reveals that log annual earning for marries women without child is higher than the females with one kid and that gap expands for more than one children (see fig 4.2).
Chart4.2: Log wage of mothers with kids
0.511.520.511.52
0 2 4 6
0 2 4 6
0 1
Density 2
l_whr
Graphs by RECODE of FH6 (EH26 18.6 Total N children borne)
Decomposition of wage for two groups of married females, group 1 with one kid and group two with more than 1 kids is reported in table 4.3. The result shows mean pay gap of women with 1 child and mothers of more kids is positive and significant and it is 0.49 for rural and 0.40 for urban. More interestingly major portion of 0.30 of this gap is unexplained for both rural and urban areas. But endowment effect is more for urban women. This might be because of fall in productivity of mothers of two or more kids relatively more than that of mothers with one kid.
Table 4.3: Decomposition of wage across married women with one kid and more than 1 kids
Overall Rural Urban
group_1 (women with 1 child) 1.99*** 2.54***
group_2 (mother >1 child) 1.50*** 2.14***
difference 0.49*** 0.40 ***
endowments 0.13*** 0.18***
coefficients 0.30*** 0.30***
interaction 0.069*** -0.08***
N 10239 4807
The literature shows that fatherhood increases male earning as males as breadwinners work more and become more productive to feed the children and family. This increases fathers earning more than married men. On the other hand employer may have bias in favor of married men relative to married females of males..The decomposition of wage earning for fathers and married men reflects that children have positive and significant influence on male earning.
Table 4.4 Decomposition of wage across Males : (Group 1=Married with 1 kid, Group2=married with two or more kids) (Deborah J. Anderson, 2003)
4.5 Decomposition of wage across Males
(Group1=Married Male Group2=Father of more than 1 kids )
overall Rural Urban
group_1=Married male 1.99***
2.54***
group_2=Father 2.09***
2.87***
difference -0.10***
-.33***
endowments -0.03***
-.18***
coefficients -0.06***
-.06***
interaction 0.001
.10***
N 17916 11523
overall Rural Urban
group_1=Married male 1.92***
2.61***
group_2=Father 2.09***
2.87***
difference -0.17***
-0.26***
endowments -0.08***
-.18***
coefficients -0.10***
-.10***
interaction 0.00
.02
N 16165 8871
5. Conclusion
Major contribution of this paper is to capture the issue of wage gap among women workers and influence of parenthood on wage earning of mothers and fathers in Indian context. There is hardly any study that addresses this issue for Indian labour market. Our result conform that mothers are earning less than childless women in India. And fathers are earning more than married males. It appears that motherhood entails a wage “penalty” in the labour market and more importantly fatherhood is associated with a “bonus” in terms of higher wage premium indicating differential impact of parenthood within a family. We also investigate farther the role of family culture in explaining this pay differential. There are few issues that need attention for context of this study. Earning differential between mothers and to be mothers may have potential endogeniety. This is because controlling for all other factors attitude and culture of married females could be different from mothers. We can address the problem of omitted variable bias by incorporating the variable, number of children desired as one of the explanatory variable.
Reference
Blinder, A. S. (1973). Wage discrimination: reduced form and structural estimates. Journal of Human resources, pages 436–455.
Botticini, M. and Eckstein, Z. (2005). Jewish occupational selection: education, restrictions, or minorities? The Journal of Economic History, 65(04):922–948.
Basu, S (2015), Gender Bias Work Participation and Socio-cultural Influences: A Study on Indian Labour Market, CBC Journal of Management Practices, Vol. 2 No. 2, 29-40
Deborah J. Anderson, M. B. (2003). The Motherhood Wage Penalty Revisited: Experience, Heterogeneity, Work Effort, and WorkSchedule. Industrial and Labor Relations Review, Vol. 56, No. 2 , 273-294.
Grogger, S. G. (1994). The Economic Consequences of Unwed Motherhood: Using Twin Births as a Natural . American Economic Review, Vol 84, No.5 , 1141-1156.
Ichino, A. and Maggi, G. (2000). Work environment and individual background: Explaining regional shirking differentials in a large italian firm. The Quarterly Journal of Economics, 115(3):1057– 1090.
Khanna, S. (2012). Gender wage discrimination in india–glass ceiling or sticky floor? Centre for Development Economics (CDE) Working Paper, 214.
Mincer, J. and Polachek, S. (1974). Family investments in human capital: Earnings of women. In Marriage, family, human capital, and fertility, pages 76–110. NBER.
Sean de Hoon, R. K. (n.d.). Marriage and Female Wages: DoMarried Women Pay a Penalty or Earn a Premium?
Shelley J. Correll, S. B. (2007). Getting a Job: Is there a Motherhood Penalty? American Journal of Sociology, Vol 112 no. 5 , 1297-1339.
Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International economic review, pages 693–709.