In the 1990s, poverty increased in Pakistan and the differences in mean earnings of males and females remained quite significant. In 1998-99, the ratio of average female earnings to average male earnings declined to 19 percent from 24 percent in 1992-93. In nominal terms the earnings difference decreased to Rs 2386.76 in 1998-99 from Rs 1177.52 in 1992-93. This shows that, gender-based earnings gap is high and it increased in Pakistan during 1992–1999. The earnings gap could be a result of the difference in male-female education and rising involvement of females in the labour market. Another reason could be the change in occupational composition of males and females. The study decomposes the earnings differences in productivity differences and market discrimination for the male and female worker belonging to poor and non-poor households. The earnings differential could be the result of unfair hiring practices leading to labour market discrimination. This implies that full potential of the work force is not utilised, which will lead to economic inefficiencies, providing a justification for government intervention. However, if the earnings gap is the result of individual’s own occupational choice, then government intervention to ensure ‘equal pay’ will create distortions. Therefore, it is important to understand the main factors contributing to the differences in the earnings gap, keeping in view the occupational concentration of males and females. The workers in Pakistan, particularly females, are crowded in a few occupations like sales and service workers, agricultural workers, or production workers. However, within these occupational categories the distribution of females has changed significantly over time. The study examines the following questions: Is occupational concentration different between the poor and non-poor households? How far do the characteristicsadjusted earnings differential decline when we adjust earnings profiles for occupational choice? The results of the study give us important insights into the earnings decomposition based on gender and poverty-gender. The results support the view that education and experience contribute significantly to the earnings of females and males. For females, the returns to education are higher, whereas the returns to experience are higher for males. The estimates of the rate of return are also sensitive to the occupational structure, implying a close linkage between education and occupational choice. The results are more sensitive to occupational concentration among the non-poor households. Furthermore, the residual earning gap is present for both groups, implying that the females of poor and non-poor households, both, face labour market discrimination. The analysis shows that poverty is an important dimension to examine occupational choice and gender discrimination. In order to integrate gender into poverty analysis, the following steps are critical: (1) ensuring that gender is addressed in all four dimensions, viz., education, occupational concentration, poverty of opportunity, and poverty of capability; (2) documentation of the experience of poverty in all four dimensions mentioned above; (3) undertaking of gender analysis of the data gathered and integrating the findings into the country’s poverty diagnosis; and (4) ensuring availability of reliable data with gender-based information.
