ABSTRACT
In the familiar Heckscher-Ohlin-Samuelson (HOS) model of trade involving two commodities and just two primary inputs, land and labour, the comparative statics of industry-level input use per unit of output are very simple. Let commodity 1, say, be the land-intensive industry, so that when the terms of trade (p 1/p 2) ‘increase’ the real rent rate (wage rate) rises (falls). In each industry land use (labour use) per unit of output will fall (rise), so that for each primary input its use is inversely related to its real price. Suppose now, however, that we allow, realistically, for the fact that the two produced commodities are also used as inputs, as in Vanek (1963), Melvin (1969) and Woodland (1982, pp. 105–119) for example. How will the direct use of inputs, at industry-level, per unit of gross output (or ‘input use’ for short) now be related to real input prices? More specifically, will such relationships be qualitatively the same for the produced inputs as for the primary inputs? The familiar textbook treatment of ‘input demand in the long run’ suggests, after all, that no significant distinction need be made between the two classes of input (produced and primary)—but we shall see that that is not true in the present context.
