This paper proposes a two-sector model of child labor to argue that income channels are necessary but not sufficient to explain the high incidence of child labor that we currently observe in less developed countries. In addition to income channels, the model in this paper uses two extra channels: parental optimism regarding the future and child's scholastic ability, which affects the costs of education in terms of money and time. The model's predictions are tested using data from the child labor survey of Bangladesh. Empirical results provide strong support in favor of the model's predictions that non-income channels have significant effects on the amount of time worked by child workers.