This paper is motivated by the predicament of capital goods production in Bangladesh which has been penalized during past import policy regimes. Duty concessions applied to capital goods imports combined with high tariffs on imported inputs have contributed to negative effective protection. Current import policy has corrected this by moving towards a more uniform tariff structure which includes rates applied to capital goods. Under the new tariffs, however, the capital goods sector has continued to be depressed and has provided little employment opportunity for the relatively skilled labour used more intensively in this sector. Using the presumption that this sector is also an important source of skills dissemination, this paper examines the optimum tariff structure and its determinants.

Because of its relevance for many developing countries, the analysis is kept at a general level. Having made the presumption that there is learning by doing in the capital goods sector, the optimal tariff structure is examined. A two period model with intersectoral linkages demonstrates the dynamics. The optimal tariff rate on capital goods imports is also solved for. While machinery imports can lead to the acquisition of technology, the dissemination of such technology requires domestic capabilities and endowment of skilled labour. Domestic production of capital generates learning by doing and thereby enhances resource productivity in the future by increasing the endowment of skilled labour.

The tariff structure is found to depend on the extent of learning and intersectoral linkages. An agreegated numerical model using data from Bangladesh also confirms this result. The conclusion is that encouraging capital goods production will increase the availability of skilled labour as well as the level of welfare.