Microcredit program, originating in Bangladesh in the late 1970s, has played an important role to meet the financing needs of the impoverished communities around the world. While the successes and failures of microcredit in lifting the poor out of poverty have been recorded in a wide array of literature, the employment outcome of participating in a microcredit program as a pathway to poverty reduction has been studied much less. Using two waves of longitudinal data on over 2000 households, we examine the employment impact of microcredit program in Bangladesh during 1998-2004. The longitudinal nature of data allows us fixed effects estimation of the effect of microcredit program participation on self-employment hours and household labor income isolating the biases that may result from non-random program placement, censoring in self-employment work hours and income data, and non-random sample selection of households or individuals as participants who already have entrepreneurial skills or pre-existing household conditions favourable to self-employment activities. The fixed effects estimate shows that households that participate in microcredit program work on average 245 hours longer in self-employment activities and earn 9.4% higher labor income than non-participant households. These extra hours are equivalent to around 7 weeks of employment for a person. The income effect of microcredit program participation is more discernible on household labor income than on total household income due to lack of direct link of microcredit program with non-labor income sources such as remittance. The participating households at the bottom of the income distribution appear to have gained more than those at the upper end suggesting equalizing effect of microcredit program participation over and above the positive effect on employment and income growth. Thus microcredit program in Bangladesh has succeeded in providing employment generating capacities to participants and raised the potential for income growth that contributed to poverty reduction.