The Effects of Some Selected Variables on Child labour at Chapi Nawabganj District in Bangladesh- A Multivariate Analysis
.........................................................................................................................
Md. Rashed Alam
Lecturer
Department of Population Science and Human Resource Development
University of Rajshahi,
Rajshahi-6205, Bangladesh.
E-mail: mrasps_29us@uyahoo.com

Bangladesh is a developing country, about 50% of the household lives below the absolute poverty level. Employment of every young child is particularly an alarming problem. In 2002/03, the Bangladesh Bureau of Statistics (BBS) conducted the Second National Child Labour Survey (NCLS). This has been designed and conducted in the context of the commitments made by the government of Bangladesh, following the ratification of the International Labour Organization (ILO) worst forms of child labour convention (No.182) 1999. The child mortality rate under the age of 5 is estimated at 4.1 per 1000 live births and the maternal mortality ratio (MMR) is 3.8 per 1000 live births in 2001.

Considerable progress has been achieved through immunization programs. Percentages of boys and girls aged 12-23 months immunized against DPT (3 or more doses) are 76.8 percent and 71.7 percent respectively. Percentage of boys and girls aged 12-23 months immunized against polio (3 or more doses) are 92.1 and 88.4 respectively. On the other hand, children of the same age immunized against BCG: boys - 93.8 percent, girls - 90.4 percent and against measles (3 or more doses) for boys are 78.5 percent and girls 73.5 percent. National Child Labour Survey 2002-2003, BBS.

Extreme forms of poverty play a crucial role. Child labour is part of a vicious cycle, with poverty as a main cause as well as a main consequence. This implies that child labour cannot be addressed in isolation. Among factors contributing to child labour are rapid population growth, adult unemployment, bad working conditions, lack of minimum wages, exploitation of workers, low standard of living, low quality of education, lack of legal provisions and enforcement, low capacity of institutions gender discrimination, conceptual thinking about childhood etc. One or more of the above contribute to the large numbers of children working under exploitative or hazardous conditions. Child labour is a persistent problem throughout the world, especially in developing countries (ILO, 1997). It is especially prevalent in rural areas of those countries where poverty is widespread, coupled with the lack of capacity to enforce minimum age requirements for work and schooling. Among the variety of reasons for child labour, the most important is the pressure upon them to escape the plight of poverty (Ahmed, A. & Quasem, M. A. 1991).

Child labour is not a new problem, and there is a long history of international efforts to combat it. The International Labour Organization (ILO), for example, in 1919 developed the first Minimum Age Convention that regulated the age at which children could work. Then, in 1973, a more comprehensive Minimum Age Convention, Number 138, was adopted, and it remains the fundamental standard. Although not new and always a thorny problem, child labour has now become increasingly complex, assuming new forms as global realities and relations have changed. Among the underlying causes, poverty and economic disparities are of course, critical factors. For much of human history, children have contributed to family welfare in a variety of ways, but intensified urbanization and the breakdown of traditional economic systems have made even basic subsistence more precarious and put children at ever higher risk. The results of a nine-country survey in Latin America, for instance, showed that if teenaged children did not work, poverty rates would increase by 10 to 20 per cent (Ahmed, A. & Quasem, M. A. 1991).

The aims of this study are to investigate the socio-economic conditions and harmful side of child labour and how it can be solved.

Data sources
Data for this study was drawn from a survey, conducted under the authority of the Department of Population Science and Human Resource Development of Rajshahi University. The survey was carried out in Gomastapu Upazila, Chapai-Nawabganj district. To collect the data from the above mentioned areas a survey on child labour was conducted. In addition, fruitful discussion was made with the employers to study several of the functional aspects of child labour. First, a list of child labourers was collected from the studied area and then 200 child labourers were selected for detailed interview using a structured questionnaire.

Path Analysis
A path analysis is one technique of showing causal linkages among the interrelated variables. The technique of path analysis, which was developed during the 1920s by Sewall Wright as an aid to the quantitative development of genetics, gained popularity in social science studies with the further expositions (Alwin D.F et all., 1975).

Path analysis presumes the existence of a causal framework interlinking different predictor variables with the response variables. Such representation of the causal variables is called a path model and it is both stochastic and explanatory and is said to be an extension of the multiple regression model . It helps in estimating the magnitude of the linkages between interrelated variables and provides information about the underlying causal processes. This technique explores a chain of relationship among the variables by using standardized regression coefficients of a set of regression equations (Duncan, O.D. 1977). The fundamental to the path analysis is the path diagram which is the outcome of a set of linearly interrelated variables and the assumed causal relationship among them. In the path diagram the principles are as follows:

(i) the variables are arranged from left in such a way that all the endogenous variables are to the right of their exogenous variables
(ii) the unidirectional straight arrows called henceforth as 'causal paths' that go from left to right represent the endogenous variables, and
(iii) on the other hand, the two-headed curvilinear arrows represent the non-causal (correlated) relationship among the exogenous variables. This study employs a recursive path model relating to fertility and some of its determinants (Alam et all., 2004).

Methods and Model Specification for Path Analysis
Path analysis is a straightforward extension of multiple regressions. Its aim is to provide estimates of the magnitude and significance of hypothesized causal connections between sets of variables (Chandrasekaran and Hermalin, 1975). This analysis disentangles the specific mechanisms of the socio-economic factors affecting child labour by taking into consideration the intermediate variables involved in the analytical system. Moreover, path analysis provides a theoretical model specified as a system of simultaneous regression equations, which are linear, additive and usually recursive. This is best explained by considering a path diagram.

The path estimation equations are derived from the structural equations by applying the basic theorem of the path analysis. Thus, it is to be noticed that structural equations are different from the path estimation equations. According to the causal ordering of variables, we may divide the selected set of variables into three groups that are given below:

This model is a recursive path model in which each variable is assumed to be dependent upon all prior causal variables. From the path analysis the direct effects, indirect effects, joint effects, implied effects and total effects of each selected explanatory variables on Child labour are obtained.

From Table 2 we show that, most children whose age is less than 1o are involved in risky work; on the other hand we observed that most of the child are working for their own needs, not entertainment. Most of them fall prey to diseases and their age range is 13 to 17 years. We also observed that most of the children are illiterate and few have primary and class six and seven. Most of the children are working for their family's needs and their father's occupation is agriculture. We also showed that their income is very low and they do not have involvement with any NGO. Most of the child use their money for family needs. An important matter we observed is that children whose age is less than 10 years do not smoke but most of the children whose age is over13 and above smoke.

If poverty, as Nobel laureate Amartya Sen argues, is to be defined not merely in terms of low income but as a state of deprivation of basic capabilities, nothing illustrates that more forcefully than child labour. A result and also a cause of poverty, child labour is a prison that withers both capabilities and potential. The prevalence of stunting, under-weight and wasting in children aged 6-71 months has shown a modest decrease over the past decade. The prevalence of stunting amongst girl children has declined from 65.9 in 1989-90 to 49.1 percent in 2000. The prevalence of under-weight girls children declined from 67.8 percent in 1989-90 to 50.9 percent in 2000. The prevalence of wasting has also declined from 15.9 percent to12.0 percent for boys and 17.3 percent to 11.4 percent for girls during 1995-96 to 2000.

Non-causal effect = Total association- Total effect.

Figure 1: Path diagrams of factors affecting respondent occupation through other variables

From Figure 1 we see that age of the respondents is positively significant correlated with respondent income, purposes/use of money and respondent occupation are positively significant. Respondent education is negatively significant, correlated with respondent smoking, risky work and participation with NGO at 5% level of significance. Again participants in NGOs are negatively significant correlated with age of respondent, education of respondent, purposes/use of money, respondent income and smoking, at 1% level of significance.

According to Figure 1, we observe that there are 18 paths out of 13 hypothesized paths. In our study we have to mention the significant path coefficients only. And out of 7 variables, 3 are found to have a significant direct effect on the index of occupation. Among them are age of respondent (X1), Respondent education ( ), respondent smoking and risky work have a direct significant negative effect.

Total effect of respondent's education on occupation is found to be more pronounced in all the variables and respondent's income are positive effects. The total effect of respondent's education on the occupation (X8) is -0.391 of which about 14 % is transmitted through the age of respondent (X1) about 66% is transmitted through its implied effect in the same direction, then about 18% acts through the risky work (X7). Other indirect effects of respondent's education are via X6 and X7 and also the joint effects are negligible. Higher total positive influences of purposes of use of money on occupation and belongs to NGO. It is observed that the implied effect (P81) of age at respondent has contributed about 51%, 6% and 37 %of its total effect on occupation while the implied effect (P85) of smoking has contributed to about 55%, 3% and 82%of its total influence on occupation respectively.

Some results of path analysis deserve considerations from the viewpoint of policy implication. It has been found that respondent education and risky work have a direct negative influence on occupation. Thus raising age of respondent by implementing a minimum-age law may lower occupation and risky work also may indicate lower occupation since at that time children are risk free from reproduction. Again occupation has a direct positive effect by purposes of uses money and belongs to NGO.

Total effect of respondent education on occupation is found to be negative. Education may provide better employment opportunities outside the home and age of respondent can be raised through providing education. Based on the results of this section it may be suggested that attention should be focused on the need of providing educational facilities.

Ahmed, A. & Quasem, M. A.. 1991. Child Labour in Bangladesh. Department of Economics, Lund University, Sweden.

Ahmed, A. & Quasem, M. A.. 1997. Child Labour in Bangladesh, Bangladesh Institute of Development Studies (BIDS) 1997.

Alam M.R., Roy T.K. and Mondal D.K., 2004. Intensity of the effects of the selected socio-economic and demographic factors on fertility in Bangladesh. Man In India, Vol. No. 84(1&2):Pp 51-62.

Alwin D.F. and Hauser 1975. The decomposition of the effect in path analysis, American Sociological Review Vol. 40 Pp. 37-47.

Chandrasekaran, C. and Albert I.Hermalin,1975. "Measuring the Effect of Family Planning programs on Fertility. Paris": International Union for the Scientific Study of Population Published for the Development Centre of the Organization for Economic Co-operation and Development.

Duncan O.D. 1977. Path analysis: Sociological examples (Addenda), Causal models in social science, Chicago, Aldine-Atherton Inc.

ILO, 1991. International Labour Organization, Vol. No. 182,1991.