Labour Supply in a High Unemployment Economy: Estimation Strategy
Variables related to household characteristics bring this perspective into the decision to supply labour to the market. The dummy on having a household member abroad is expected to test for the effect of remittances on labour supply decisions of non-migrant household members. Based on the neoclassical model of labour-leisure choice, remittances shift the budget line upwards and assuming that leisure is normal good then they reduce the labour supply of non-migrant household members (i.e. an income effect). However, in transition and developing countries characterised by an underdeveloped credit market, remittances are thought to be an important source of financing entrepreneurial projects. In line with this, some studies find that remittances increase the propensity of some non-migrant household members (males in particular) to engage in self-employment.
Regarding household incomes, our data allow us to distinguish between household labour and non-labour incomes. The regional dummies and the urban versus rural dummy account for any regional-specific effect in labour supply decisions. We expect that those from the region of Prishtina (the capital city) are more likely to participate in the labour force due to their higher chances of getting jobs. However, those from the western and southern regions (where agriculture is an important activity and engages many household members) are also more likely to participate compared to other regions. These regional dummies may be thought of as capturing any discouragement effect among the non-employed workers in their decisions to supply labour to the market. Given that when employed urban residents tend to hold formal full-time jobs, we would expect that education has a larger effect on their likelihood of participation compared to rural residents. To test for this we construct interactive dummies between residence and other explanatory variables.
Table A2 in the Appendix gives the mean of the explanatory variables for the whole sample, for those in the labour force and those inactive. There are three important observations from this data that are in line with our discussion above. First, those with higher potential wages and higher chances of getting a job (as proxied by the education level) are more likely to be in the labour force. Second, persons out of the labour force tend to be from households with higher non-labour incomes per capita. Third, females are less likely to participate, which is in line with their observed lower educational level (that affects their potential wages) and higher opportunity costs of employment (childcare).