Immiserizing Deindustrialization: A Dynamic Trade Model with Credit Constraints

This paper argues that opening up an economy to trade can result in welfare reducing deindustrialization when agents are subject to credit constraints. The standard interpretation of such deindustrialization as a reallocation of resources across sectors with no aggregate long-term adverse consequences is shown to be misleading, suggesting that episodes like the deindustrialization of India in colonial times may have been quite damaging. The paper develops a two-sector, one-factor, open economy dynamic model with bequests and credit constraints. Agents differ only in their initial wealth and can produce in the agricultural sector, where credit constraints never bind, or in the industrial sector, where credit constraints could bind. Under autarky, a productive agriculture provides the funds for the needed investment in industry. If agriculture is productive enough, the bequests of agricultural workers over time are sufficient to enable their offspring to work where they wish. The economy converges to an equilibrium where credit constraints are not binding. Prices are such that agents are indifferent between occupations. If agriculture is not productive enough, the price of the industrial good must be high enough to allow the bequests needed for the offspring of industrial workers to remain in industry. At this bequest-sustaining price, everyone prefers to work in industry, but only the offspring of industrial workers can do so. In an open economy, if trade price falls below the bequest-sustaining level, deindustrialization occurs over time. If it remains above the occupational indifference level, this deindustrialization is also involuntary. Aggregate welfare falls discontinuously when price falls below the bequest-sustaining level as credit constraints become binding for all agents. Deindustrialization must be immiserizing when it is involuntary and the comparative advantage in agriculture is small so that gains from trade are small. The results are shown to extend in interesting ways for a large country and for a richer occupational structure.