Abstract
Poverty is considered to be the most important cause of child labour and it deprives children of schooling and acquiring human skill. The poor children grow as unskilled workers and earn low wages in adulthood. So, poverty persists and the parents are forced to send their children to work and a child-labour trap is formed. The econometric analysis using state level panel data in the Indian context demonstrates that poverty, illiteracy and child labour are significantly correlated. The results show that poverty adversely affects children’s schooling and education and results in persistence of poverty thereby creating a child-labour trap.
Introduction
The problem of child labour is a serious socio-economic problem that has drawn huge attention from all quarters of the society throughout the globe and the international organizations like the UN, UNICEF and the ILO; and national governments of the countries are in search of effective policy measures to solve this problem. The initiatives in this regard are broadly classified under three headings: law, direct intervention and market-based schemes. Among the factors behind this problem, poverty is considered to be the most important one. Basu (1999) has shown how a child labour trap is formed due to persistence of poverty in a dynamic perspective. In fact, poverty is closely associated with lack of schooling and incidence of child labour in most of the cases. As the poor parents send their children to work instead of sending them to school they are deprived of education and acquiring skills; and in effect, the children grow as unskilled adult workers. Naturally, they remain poor in the next generation as well, and they are forced to send their children to work again to supplement the family income. Thus the chain of poverty persists and a child-labour trap is formed in the system (Basu, 1999, 2000; Basu and Van, 1998; Basu and Zarghamee, 2009). Educational failure is cited as another important cause of child labour although all educational failures cannot be attributed to the incidence of child labour (Dumas, 2007; Ravallion and Wodon, 2000).
The ILO states that child labour deprives the children of their childhood and their dignity and it hampers their access to education and acquisition of skill (Basu, 1999; Burra, 2005; Venkatanarayana, 2004). The First World Summit for children in 1990 showed that education could play a key role in reducing child labour and in its eventual abolition (Grootaert and Kanbur, 1995). According to the ILO estimates, child labour in the world in the age group 5–14 years has declined to 153 million in 2008 from 250 million a decade ago. According to the provisional estimate of Census 2010, the number of child labour in India is 12.62 million against 13.6 million in 1981 indicating that the decline in the absolute number is not very significant over a period of three decades. There are two major schools over the debate on child labour. The first school blames poverty as the main cause of child labour and believes in its regulation. The second school makes the lack of educational outcomes responsible for it and advocates for complete prohibition of child labour. Accordingly Prohibitive and Regulatory Acts have been passed in almost all the countries and policy measures have been adopted to increase literacy and compulsory education for the children. The Child Labour (Prohibition and Regulation) Act was passed in India in 1986. This regulatory measure has been supplemented by a number of schemes for greater schooling and universal literacy of the children. Neither of the schemes has achieved any significant success. Our argument is that poverty and educational failure are not two different issues and they are very closely associated with each other resulting in the formation of a child-labour trap.
Burra (2005) notes that out of 203 million child population in the age group 5–14 years in India, almost 100 million are out of the school system. But referring to all out-of-school children as child labourers is not very convincing. Although working as child labourer and schooling are not always mutually exclusive, lack of schooling is a major cause as well as consequence of child labour. Therefore, schooling and education are the prime requirements for the children to grow as skilled workers. Those who cannot attend school due to economic, health or school-related problems can be referred to as educationally deprived children (Venkatanarayana, 2004). So, schooling and human capital formation are considered as the most effective solutions to the problem of child labour. Many scholars are in favour of a complete ban of child labour. But in countries with mass poverty and weak governance, the enforcement of such regulations is neither feasible nor always desirable. Basu (1999) is opposed to the idea of a legal ban and compulsory education if they make the conditions of the children worse. It is demonstrated that if family income falls below some critical subsistence level, the parents will choose to send their children to work instead of sending them to school. The dynamic implication of not sending a child to school or denying the child the opportunity of acquiring skills through education creates a child-labour trap due to persistence of poverty (Basu, 1999; Basu and Van, 1998).
There are theoretical reasons and empirical evidence to believe that improvement in the condition of adult workers may lead to a decline of child labour. So, one route to curbing child labour can be intervention in the adult labour market. But, the effect of minimum wage legislation and government intervention in the labour market on child labour may be more complex. It may so happen that as a result of the enforcement of a minimum wage Act, employment of adult labour declines and use of child labour rises (Basu, 2000). Basu and Zarghamee (2009) have considered the idea of controlling child labour using labour standards in international trade and the boycott of child labour-tainted products in export markets. The authors apprehend that the presence of a consumer boycott will command a lower price for the tainted products resulting in a reverse reaction which will cause child labour to rise rather than fall. Baland and Duprez (2009) also are not sure whether the condition will be better if the labelling programme is introduced for child labour-tainted products in the international market. The parents’ income is definitely very important for children’s education and skill-formation and this income is dependent on capital investment, overall economic growth and institutional aspects of the countries along with other things. In this context, Bhaumik and Banik (2006) show that the Caribbean economies have not been able to take advantage of the physical proximity of developed countries in spite of high levels of human development in terms of literacy, education and skill formation due to lack of vision about development of some institutional aspects. Not just literacy, continuous upgradation of educational skills and training of workers is also necessary in the present global economic perspective.
It is important to understand how the decisions regarding children’s education are taken by the parents and when the parents find the schooling of the children economically meaningful. Basu (1997) and Galor and Zeira (1993) have demonstrated using overlapping generation models that expenditure on education has to cross a minimum threshold level in order to impart skill in labour and make the investment on education yield any benefit. These studies also show that in a two-period (child and adult) model, such investment takes place if inheritance or bequest exceeds some critical value. However, the earning of the poor parents does not permit them to make the critical minimum level of expenditure. In this context, the observation of Glomm (1997) is worth mentioning. He remarks that an individual, when young, can make his/her own decision regarding investment on education. But at the school level, it is the parent who decides whether the child should go to school and for how many years. The paper by Ermisch and Francesconi (2001) concludes that parents’ education and income do matter in a child’s education. The study of Post (2011) in the context of some Latin American countries highlights the quality of education as an important factor in this regard. Therefore, schooling and quality education for poor children are very important to break the child-labour trap and mere free education in public schools, free mid-day meals or some ordinary benefits from the government are not sufficient for solving the problem.
The schooling of the children has both supply- and demand-side problems. The study of Khasnabis and Chatterjee (2007) in the slum areas of the Kolkata Metropolitan City of India shows that despite government attempts, retaining the students from very poor economic backgrounds in a formal school is far more difficult than enrolling them. The National Family Health Survey of India, Government of India (1998–1999) and Report on Elementary Education in Rural India, 2005 reveal that the drop-out rate in the age group 6–9 years is 30.04 per cent. The respective rate for the age group 10–14 years is 65.5 per cent. High drop-out rates at relatively higher ages has some special meaning. As the child grows, his earning capacity rises and he leaves school to earn for the family. Handa (2002), on the other hand, finds that supply-side facilities of schooling are more effective than the interventions for raising the income of the poor households, in raising enrolment of poor students. Sasmal (2006), in his study of demand-side problems in primary education, argues that mere literacy does not add any skill to the child and a minimum number of years of schooling is necessary to acquire skill. He maintains that just free education, mid-day meals or other ordinary benefits are not plainly sufficient to create any strong incentive for the parents to send their children to school.
In the Indian context, we find that there is public intervention for the schooling of poor children. The government has introduced a number of schemes for literacy and free education in the country like, ‘Operation Blackboard’, ‘Sarva Siksha Abhiyan’ (universal literacy campaign), free mid-day meal for students and so on. These programmes have been coupled with a bunch of poverty alleviation schemes in addition to a Constitutional Provision for compulsory and free education up to 14 years and Prohibitive Acts on child labour. Despite these initiatives, India remains a land with the highest number of illiterate and poor people in the world. About 310 million people are illiterate (Population Census, 2011), more than 300 million are living below the poverty line and 12.62 million children are working as child labour (Source: Ministry of Labour, Government of India). The study by Jalan and Panda (2010) informs that the Mid-Day Meal Programme has a mixed response in increasing children’s attendance in school but its observation on the quality of education in public schools is very revealing. It reports that in a large number of cases a student of standard IV cannot work out a problem of standard I. Therefore, just schooling is not sufficient; the quality of education is also very important. But poor governance, corruption, lack of proper planning and management have made the government measures grossly unsuccessful. When the failure of the market is very clear and the interventions of the government remain ineffective, we may talk about corporate social responsibility in this regard. But Raman (2006) shows that very few studies have looked at the issue of corporate social responsibility in developing countries like India. In fact, the corporate sector could have helped the implementation of labour laws and spread of education and training among the poor children.
The objective of this article is to examine the relationship between poverty, literacy and child labour in the Indian context. The hypothesis is that poverty is the cause of child labour, and educational failure due to persistence of poverty creates a child-labour trap. The hypothesis has been tested using econometric analysis based on state-level panel data of the Indian states.
Econometric Analyses on the Relationship between Poverty, Literacy, Per Capita Income and Child Labour in Indian States
Data and Methodology
The panel regression has been done to examine the relationship between poverty, literacy, state per capita income and child labour using state-level panel data for the states and union territories (UTs) of India. The data on percentage of child labour and literacy rate for 27 Indian states and union territories have been collected from the Census of India for the years 1971, 1981, 1991 and 2001 and the website of the Ministry of Labour, Government of India. Census 2011 gives data on literacy rates but state-wise data on child labour for the year are not yet available. The data on poverty (head count ratio) have been taken from ‘Poverty in India and Indian States: An Update’ by Gourav Datt, Food Consumption and Nutrition Division (FCND), International Food Research Institute, Washington D.C., USA, FCND Paper No. 47, July 1998 for 15 states and for the years 1971, 1981 and 1982–1983. In the study of Datt (1998) there are separate estimates of poverty for rural and urban areas of the Indian states. Since data on overall poverty are not available and child labour comes mainly from rural areas, rural poverty has been taken as a measure of poverty to explain the problem of child labour and illiteracy in our econometric analysis. Further we note that the estimate of poverty for the year 1981 is not available. So, the estimate for the year 1982–1983 has been used as a proxy for the year 1981 in our panel regression. We have also used poverty measures (head count ratio) of Tendulkar and Ravi for the years 2004–2005 and 2009–2010 respectively for the major states of India from Ahluwalia (2011) as proxies for the years 2001 and 2011 respectively due to non-availability of data for the relevant years. The data on state per capita income (NSDP) at constant prices (1980–1981 = 100) have been taken from the Handbook of Statistics on the Indian Economy, Reserve Bank of India, 2009–2010. In each regression there are four to five waves of panel data. The whole exercise of panel regression has been done using the software STATA. Both fixed effects and random effects models have been used for panel regression. The Hausman test has been conducted to decide which model is more appropriate in the estimation. Following Wooldridge (2009) the regression models have been specified as below.
The equation for the fixed effects model is:
where, Yit is the dependent variable with value of the ith individual observed in time t, i = 1, 2, …, n
Xit is the ith independent variable (IV) observed in time t.
β is the coefficient for IV
ui is the unobserved individual heterogeneity of ith entity of the dependent variable.
εit is the error term of the ith entity in time t.
In the fixed effects model: E (Xit, ui) ≠ 0.
That is, Xit and ui are correlated.
The Random effects model is:
where, Yit, Xit, ui, εit and β are the same as above. But here,
In addition to the value of R2, the F-statistic has been computed to test for joint significance and determine overall significance of the regression (Wooldridge, 2009).
Regression Results
The figures in Table 1 give the number and percentage of child labour in Indian states over time. It does not give any particular trend or pattern. We have tried to get some idea about the relationship between the variables from regression analysis. The results of the panel regressions have been presented in Tables 2 to 7. The F-statistic confirms the overall significance of the regression analyses in all the cases. The hypothesis that poverty is the cause of child labour has been established by the results in Table 2. The regression results in Table 2 show that percentage of child labour is positively related with poverty ratio and the coefficient is significant at the 5 per cent level.
The results in Table 3 show that percentage of child labour is negatively related with literacy rate and the result is significant at 5 per cent level. In Table 4 literacy rate has been regressed on child labour and it is found that there is a significant negative relationship between the two. It signifies that the incidence of child labour adversely affects children’s education. The results in Table 5 establish the negative relationship between poverty and literacy. The hypothesis that poverty is the cause of child labour and child labour causes educational failure of the child is empirically verified by these econometric results.
The results in Tables 6 and 7 show that child labour has a negative relationship and literacy has a positive relationship with state per capita income. It gives a clear impression that in states with higher per capita income, literacy is higher, poverty and child labour are lower.
Administrative and Policy Failures
Despite various legislative and policy measures, child labour has not declined in India, in absolute numbers, in the last decade. It requires special mention that the ratio of girl child labour is higher than the boy’s ratio in almost all occupations (see Table 8). It is an indication that administrative and legal measures have grossly failed to protect the girl children.
Distribution of Working Child Labour, in the Age Group 5–14 Years in Major Indian States and Union Territories
Panel Regression of Child Labour on Poverty
Dependent Variable: Percentage of child labour in the state/UT.
Explanatory Variable: Percentage of poverty (H–C ratio) in the state/UT.
Number of Groups (State/UT): 15
Number of Observations: 60
Time Period (T): 4
Panel Regression of Child Labour on Literacy
Dependent Variable: Percentage of child labour in the state/UT.
Explanatory Variable: Percentage of literacy in the state/UT.
Number of Groups (State/UT): 27
Number of Observations: 108
Time Period (T): 4
Panel Regression of Literacy on Child Labour
Dependent Variable: Percentage of literacy in the state/UT.
Explanatory Variable: Percentage of child labour in the state/UT.
Number of Groups (State/UT): 27
Number of Observations: 108
Time Period (T): 4
Panel Regression of Literacy on Poverty
Dependent Variable: Percentage of literacy in the state/UT.
Explanatory Variable: Percentage of poverty (H-C ratio) in the state/UT.
Number of Groups (State/UT): 15
Number of Observations: 60
Time Period (T): 4
Panel Regression of Child Labour on State Per Capita Income (NSDP)
Dependent Variable: Percentage of poverty (H-C ratio) in the state/UT.
Explanatory Variable: State Per Capita Income at constant prices.
Number of Groups (State/UT): 26
Number of Observations: 78
Time Period (T): 3
Panel Regression of Literacy on State Per Capita Income (NSDP).
Dependent Variable: Literacy in the State/UT.
Explanatory Variable: SNDP at constant prices.
Number of Groups (State/UT): 24
Number of Observations: 96
Time Period (T): 4
Hausman test accepts fixed effects.
Sectoral and Sex Composition of Child Labour as a Percentage to Total Workforce in Rural Areas of India, 2010
Despite much hyped national policy for compulsory education of the children and constitutional provision and large scale arrangements for universal literacy, the reports of the Household Population Survey in India gives a very dismal picture. Twenty-five per cent of girl children and 23 per cent of male children in the age group 6–9 years are not at all enrolled in school. The education level is less than 5 years for 33.6 per cent girls and 35.5 per cent boys in the age-group of 10–14 years. The drop-out is so high that only 0.4 per cent of both girls and boys complete their class 10. National Child Labour Projects (NLCP) was launched in 1988 in the country under the Project Based Plan of Action. Under this scheme, special schools were set up for formal and non-formal education along with vocational training, stipends, nutrition and regular health check-ups of the children to prepare them to join regular mainstream schools. The basic purpose was to bring the children under mainstream education system and social life. But the scheme has failed to achieve success. Out of 12.66 million child labourers (Census, 2001) only 53,009 could be mainstreamed under NCLP.
Summary and Conclusions
This article addresses the problem of the child-labour trap caused by poverty. The existing literature and empirical evidence show that if the household income falls below a critical subsistence level, the parent is forced to send his child to work to supplement the family income. As the child works as a labourer, he is deprived of schooling and acquiring skill and grows as an unskilled worker. As an unskilled adult labourer he earns a low wage in his adulthood and as a result, poverty persists and he is again forced to send his child to work. Thus the chain of poverty persists and a child-labour trap is formed in a dynamic perspective.
It is suggested that there is scope for fruitful public intervention and in fact, there are policy interventions in different forms. But the studies on child labour reveal that the legal ban, government regulations, compulsory and free education, free mid-day meals for poor students and various poverty alleviation measures have failed to make much of a dent on this problem. This article argues that in cases where poverty is the cause of child labour, the child-labour trap cannot be broken unless poverty is eliminated. This is because poverty deprives the children of education and acquiring human skill. The educational failure due to poverty causes a child-labour trap. This article has done econometric analysis using state level panel data in the Indian context and found that child labour is significantly associated with poverty, illiteracy and the per capita income of the states. In states with higher per capita income, the incidence of poverty and child labour is low and literacy rate is high. So, it is suggested that top priority should be given to reduction of poverty and economic growth so that the incidence of child labour declines. The article has also highlighted the administrative failure to implement the policy measures to reduce poverty and eliminate child labour.
Footnotes
Acknowledgements
The authors are grateful to the anonymous referees of the journal for their extremely useful suggestions to improve the quality of the article. They also express their gratitude to Sugata Marjit and Ritwik Sasmal for their help in writing this article. The usual disclaimers apply.