Labour Supply in a High Unemployment Economy: Findings
In this section, we discuss our findings from estimating Equation. Given the observed large gender differences in the labour market and our prior expectations regarding different effects of explanatory variables on labour supply of males and females explained above, we estimate separate regressions by gender. Findings are presented in Table 2. The estimated coefficients on age and its squared value have the expected signs and are significant for males and females, giving the normal concave shape of the LFP curves with respect to age. We have insufficient evidence to reject the null hypothesis of similar shapes of the LFP curves with respect to age for urban and rural residents.
In line with the evidence elsewhere, the estimated coefficients on educational dummies have the expected positive signs and are significant. The significant and positive coefficients on the interactive dummies between residence and education (for both genders) support our expectations of a greater role for education on the labour supply of urban residents compared to their rural counterparts. Our estimate indicates that being married lowers the likelihood of supplying labour to the market for females ceteris paribus, but no significant difference is found between urban and rural females. The estimated additional affect on marital status for urban males is positive and significantly different from the rural rate. For urban areas this provides some support for our expectations of an increased labour supply of married males to provide additional incomes for the extended family, which is also in line with the findings elsewhere.
We do not find evidence for the expected negative effect on participation of females engaging in childbearing. Examining the descriptive statistics we find that 31 percent of households are composed of 8 members or more, while 17 percent have 10 members or more that eases the constraints of having children on females’ labour supply, since older family members can take care of them. Having a household member abroad is estimated to affect negatively female’ participation, as expected, though the coefficient is significant only at the 10 percent level. For males, the coefficient on the interactive dummy with urban residence is positive and significant, but again only at 10 percent level. Regarding the effect of household labour incomes, our findings for females is of an unexpected positive significant effect in rural areas, although for urban areas the interactive dummy is significant and negative. With regard to the expected negative effect of household non-labour incomes on participation, our estimates support this only in the case of urban males.
Table 2: Estimates from the Probit Model for the Labour Force Participation
|The dependent variable equals 1 if in the labour force (employed or unemployed) and 0 if inactive|
|Female age 25-40||-0.096||-0.90|
|A household member is abroad||-0.092||-0.86||-0.178||-1.81|
|Household labour incomes per capita||0.001||0.30||0.007||5.10|
|Household non-labour incomes per capita||-0.0001||-0.22||-0.001||-0.58|
|Urban resident – Age||-0.043||-1.04||0.034||0.80|
|Urban resident – Age squared||0.0004||0.92||-0.0004||-0.67|
|Urban resident – Upper-secondary education||0.472||2.92||0.443||3.34|
|Urban resident – Higher education||0.862||3.34||0.739||2.25|
|Urban resident – Married||0.570||2.53||0.239||1.62|
|Urban resident – Female age 25-40||-0.196||-1.24|
|Urban resident – A household member is abroad||0.367||1.86||-0.217||-1.33|
|Urban resident – Household labour incomes per capita||-0.002||-0.96||-0.010||-6.46|
|Urban resident – Household non-labour incomes per capita||-0.004||-3.07||-0.002||-1.54|
|Likelihood Ratio test, %2(25)||415.2||666.7|
|Pseudo R squared||0.201||0.214|
|Mean dep. variable (observed)||0.816||0.419|