DETERMINANTS OF FARM LABOUR MARKET PARTICIPATION AND LABOUR SUPPLY OF FEMALES IN MALAWI

Abstract

Female labour supply has been rigorously researched in the literature. In Malawi, however, studies on female labour supply do not exist. The objective of this study is to explore the determinants of farm labour supply of females. The recent IHS survey for 2005 has shown an increase in participation of females in the labour force. In the recent past women have been marginalized interms of access to employment opportunities henceforth they participated less in the labour market in comparison with men. An integrated household survey conducted by the Government of Malawi through National Statistics Office from 2004 to 2005 provides a data source used for this study. This study uses Tobit Type II model also known as Heckit after Heckman (1979) because the Heckit Model is simple to apply (Hill, 2003) and secondly, because hours of work as a continuous dependent variable in the sample is censored. The problem of censoring occurs for the dependent variable because it is observed only for labour market participants. Simple use of OLS will be biased and inefficient due to presence of censored observations which may result in selectivity bias. However, using OLS within the Heckit framework allows us to control for selection bias. The model has been adapted to suit local economic conditions since Malawi as a developing country labour markets may not be well developed hence we focus on agricultural labour where hours of work are flexible. Comparison with a simple Tobit is made in order to check for sensitivity of parameters of farm labour supply. The major findings from this study confirm the hypothesis that there is high labour market participation of females in the rural than urban and there is a presence of a positively sloped female farm labour supply curve with backward bending characteristic. An own female wage elasticity of about 16% and the income compensated cross substitution effect of a spouse’ wage is around 3% although not statistically significant at 10%. This indicates that leisure time of the woman is taken jointly with her husband’s leisure.