12 SOCIAL SCIENTIST
exports, was made flexible in respect of the permissible imports and transferability of the licence. Exemption from customs duties in conjunction with licence allocation was introduced. In contrast to the 'rationing* approach of the 1960s, the approach in the last few years has been one of ensuring the supply of imported raw materials, intermediates and components to fulfil the entire requirements of actual users.
II
The dominant analytical framework within which the causes and effects of trade are assesed, and 'outward-oriented' policies prescribed, is the theory of comparative cost advantage in its modern neoclassical version. As a construct for assessing the efficiency of resource allocation, it has been criticised for its severely restrictive asumptions. In particular, it assumes away a range of factors which might be of special relevance in explaining cost differences between developed and developing countries : differences in natural resources, scale of operation, and technology—in fact everything except endowments of'factors of production' like labour and capital. Crucial developmental questions go by default, for endowments and technology are assumed to be exogenously given; all countries have the same technology; factors are perfectly mobile within, but immobile between them, being allocated domestically by a market mechanism which generates full employment 6fall factors. This involves not only a mistaken understanding of technology and the functioning of markets, but also requires the existence of social devices to suppress or compensate owners of factors which lose from trade—and in labour abundant economies, according to a standard neoclassical theorem, capitalists will be on the losing side.
In the rarefied setting, freer trade is shown to yield a finite increment in national income through improved allocation. Modern treatments of trade do recognise the scope it offers for exploiting economies, of scale but these are potentially inconsistent with the orthodox 'gains from trade' analysis. Economies internal to the firm are incompatible with competition; they also, along with those external to the firm but internal to the industry, hold out the possibility of a loss from trade if comparative advantage is such that industries with greater scale economies contract in the face of foreign competition ('Kemp-Negishi theorem'). In the context of liberalisation in developing countries, the enterprises fated to contract would be the capital-intensive legacies of'excessive' import-substitution; these would presumably be characterised by more marked technological scale economies than the expanding labour-intensive ones. The latter are likely to be vertically specialised within a global industry, so that external economies would be largely dissipated abroad.
Not surprisingly, even some eminent supporters of liberal trade policies have little use for neoclassical theory. According to a classic work by Haberler, the static gains "grossly underrate" the importance of trade for development.2 As another writer of the same persuasion admits : "It is fair to say that... the direct gains from trade through a more efficient allocation of resources provided the theoretical basis on which freer trade policies were advocated for the underdeveloped countries during the last two decades. Yet the casual links running from the direct gains from trade to economic development appear-to be rather weak."3
One possible dynamic extension emphasised by some authors is the stimulus provided by the saving out of the incremental income. Again, this is inconsistent with the logic of the neoclassical analysis. As mentioned above, trade is supposed to increase the share of wages in national income at the cost of profits. On the usual assumption that workers' propensity to save is lower than that of capitalists , a decline in the economy's rate of saving would follow. (Investment is of course determined by