This paper analyses the structure, employment, and productivity growth in the unorganized manufacturing sector in India between 1984/5 and 1994/5 using National Sample Survey data for three periods, 1984-85, 1989-90 and 1994-95. The size, structure of employment and manufacturing investment, and levels and changes in the partial factor productivity and capital intensity are examined at the two-digit industry level. A simple Cobb-Douglas production function is employed to analyze the output elasticity of labour and capital. Evidence at the aggregate level shows that the size of the unorganized sector, measured in terms of number of enterprises, employment, fixed capital and value added, has declined during the period under consideration. A more disaggregated analysis at the twodigit level points to the increasing importance of food and beverages, and repair services in the unorganized manufacturing sector, and declining importance of textiles. It is also found that there is a symmetric relationship between real value added and employment in this sector. This suggests that it would be desirable to give impetus to real output generated in the sector to have the desired impact on employment generation. Evidence also indicates that the rate of growth varies widely across the two-digit industries but the variation in growth rate is smaller in the 90s. The major employment providing industries in the sector such as the manufacture of food products, beverages, cotton products and wood products have recorded a fall in employment and manufacturing investment. However, the partial factor productivity analysis shows that these industries have recorded a better performance in terms of enterprise productivity, labour productivity, and capital productivity. Our analysis shows that the workers in most industries do not receive emoluments commensurate with their productivity or the contribution they made in the value added. An exception is jute industry where the labour productivity is the highest and the emoluments per worker is also the highest. In contrast, the beverages industry’s contribution to value added per worker is relatively high but the emoluments per worker is very low, lower than the national average. Cobb-Douglas Production Function fit to the data demonstrates that labour and capital significantly contribute to output growth. Overall, we find that the efficiency of the production process has improved in the unorganized manufacturing sector in India.