preview

In Addition To Saving And Bequest Channels, The Proposed

Decent Essays

In addition to saving and bequest channels, the proposed model in this paper identifies two additional channel of child labor. These are heterogeneity in child 's ability, which causes heterogeneity in the cost of education, and differential future expectation of parents depending on their socio-economic background. I attempt to capture these four channels using a number of variables. For saving, I use income of household head and for bequest, I use whether the household owns any house. Here income of household head may be under-reported or may have some types of measurement errors, so, we use several other variables to instrument it. Ownership of a house is used to capture bequest as housing price is very high almost across the country …show more content…

Secondly, the socio-economic group in which the household head belongs to is also crucial in forming household head 's expectation. Thus, I use household head 's gender ($Head_female$), Household head 's sector of employment ($Sector$), religion of household head ($Religion$), and the location of household ($Urban$). Using these variables I estimate the following models: Where $y$ is child 's working hours per day, $ extbf{X}={Urban, Sector, Female, Head\_literacy, Religion, House,$ $ Head\_income, Head\_female, Dropout\_ability, Dropout\_cost, Dropout\_income}$, $ extbf{Z}={Urban, Head \_sector,$ $ Female, Head \_literacy, House, Land, Num \_hhchld, Head \_literacy, Head\_age, Dropout\_ability,$\ $Dropout\_cost, Religion, Dropout\_income, TV, Head\_female}$, $Q_{N}(eta) = g_{N}(eta) 'W_{N}g_{N}(eta)$, $g_{N}(eta)=frac{1}{N} sum_{i=1}^{N}Z 'epsilon$, and $W$ is a $N imes N$ positive semi-definite matrix. In addition, I also estimate logit and probit model of child labor, where I attempt to estimate the probabilities of being a child labor with respect to the variables mentioned above. Results of these models are shown in the following table:

In the above table second column reports the OLS estimates. Dummy $Urban$ has positive coefficient, which indicates that child labor on an average work 0.4 hour more in

Get Access