Children’s Own Time Use and its Effect on Skill Formation

ABSTRACT

Using time use data from a longitudinal survey (covering Ethiopia, India, and Vietnam), the present study examines how the amount of time children spend on different activities impacts their acquisition of cognitive and noncognitive skills. Modelling the skill formation production function of children and extending the set of inputs to include the child’s own time inputs, the study finds that child involvement in work activities (paid or nonpaid) are associated with a reduction in both cognitive and noncognitive achievements. The results imply an indirect adverse effect of child work on skill development through the reduction of hours of study.

Acknowledgements

The data used in this study come from Young Lives, a 15-year study of the changing nature of childhood poverty in Ethiopia, India (Andhra Pradesh), Peru, and Vietnam. Young Lives is funded by UK aid from the Department for International Development (DFID), with co-funding by the Netherlands Ministry of Foreign Affairs and Irish Aid. The views expressed here are those of the author and not necessarily those of Young Lives, the University of Oxford, DFID, or other funders. Financial support from the Charles University Grant Agency is gratefully acknowledged.

I wish to thank Patrick Gaule, Stapan Jurajda, Alan Krueger, Alex Mas, Cecilia Rouse, Orley Ashenfelter, Randy Filer, and seminar participants at CIREQ Montreal, DIAL Paris, Princeton University, and SSPC Porto for their helpful comments. All remaining errors are mine.

Disclosure statement

No potential conflict of interest was reported by the author.

Supplementary data

Supplementary Materials are available for this article which can be accessed via the online version of this journal available at https://doi.org/10.1080/00220388.2018.1499893

Notes

1. For a detailed derivation and in-depth analysis of this model, see Todd and Wolpin (2003).

2. Access to the data and permission to use them for this study is granted by the UK Data Service. The last round of survey has not yet been publicly archived by the survey administrators while this paper was written. Data from Peru is not considered in this study as a large part of the time diary data were not properly cleaned.

3. The data are clustered and cover 20 sites in each country across rural and urban areas. Sites were chosen purposively to reflect the diverse socio-economic conditions within the study countries and therefore are not statistically representative for the country: comparisons with representative datasets like the DHS samples do show, however, that in each of the countries, the data contain a similar range of variation as nationally representative datasets (Barnett et al., 2012).

4. See Fiorini and Keane (2014) for a simple illustration of how analysing few inputs in isolation conveys only partial and potentially misleading information as one cannot characterise the trade–off between inputs.

5. The full list of estimation results is available from the author on request.

6. Fiorini and Keane (2014) conducted their studies on a sample of Australian children. The pattern of time allocation of the children in their sample is quite different from the ones considered in this paper.

7. These results are not reported here, but can be obtained from the author on request.

Additional information

Funding

This work was supported by the Czech Science Foundation Grant P402/12/G130; and by the Charles University Grant Agency - GAUK (grant number 578314).