This paper analyses the growth and performance of the unorganized manufacturing (micro and small enterprises) sector in rural and urban India. The central concern would be to address the issues of employment potential of this sector. Based on NSS data on the unorganized manufacturing sector (40th Round, 1984–85; 51st Round, 1994–95; and the latest 56th Round, 2000–01) the paper looks at the growth performance of this sector during the pre- and post- reform period, and attempts to identify challenges posed and opportunities thrown by in globalising India. To find out the special locational constraints of rural enterprises, analysis has been made separately for rural and urban units. Further, to derive a firm picture about the specific sectors and sub-sectors that were rising or declining, especially in the wake of economic reforms, the analysis is extended to cover two-digit level of industrial classification for three variables, namely, number of workers, capital:labour ratio and per worker productivity. While the overall employment situation in the post-reform period is fairly depressing, the MSE sector has shown some sign of relief. It has witnessed a significant growth of employment in both rural and urban areas. But within the MSE sector, there has been a clear shift of manufacturing enterprises and employment from rural to urban areas and from tiny to bigger sized units. The levels of productivity are abysmally low for each of the three types of enterprises, but there are substantial differences among them and it is even more glaring between rural and urban located units. Thus, both locational and scale attributes are clearly at work. Moreover, there was a substantial change in the composition of workers in the post-reform period, i.e., part-time workers increased at a higher rate both in rural and urban areas. The paper unfolds the statistical delusion of post-reform high employment growth in the MSE sector and calls for a special policy attention towards technological improvement, promotion of sub-contracting and clusters.