The decade of 1990s was marked by rising unemployment in Pakistan due to low rates of economic growth and investment. Youth unemployment rate fluctuated between the low of 7.6 to the high of 13.6 per cent during 1990-91 to 2003-04. This paper attempts to explore both the macro and micro dimensions of youth unemployment issues in Pakistan at regional and gender level during the 1990s. Results from the simple bivariate regression model based on 9 years of data from the Labour Force Surveys reveals that (a) unemployment of youth only begins to decrease if the annual growth rate of GDP is greater than 4.25 per cent per annum; (b) as compared to the growth rate of GDP, the growth rate of services sector GDP has greater impact on reducing female unemployment; and (c) the growth in private sector investment is more conducive than public sector investment in reducing youth unemployment. Using household micro level data from the latest Labour Force Survey (2003-04), the preliminary and exploratory results of probit modelling indicate that skill acquisition and vocational training do not increase the chances of employment. Migration history of unemployed youth does not support the hypothesis that higher youth unemployment in urban areas reflects the transmission of open unemployment in the rural areas.