“Contractual Structure and Wealth Accumulation ”

Dilip Mookherjee and Debraj Ray
IED Discussion Paper 107, June 2000

Whether historical levels of inequality and poverty constrain the current performance of economies is a matter of central importance to the economics of institutions, growth, and development. In recent years it has also been the subject of intense scrutiny by researchers who are interested in explaining the evolution, persistence, and macroeconomic consequences of wealth inequality. Following the early work of Loury, most of the subsequent literature explains persistence of wealth inequalities either via randomness in abilities, or a combination of capital market imperfections, investment indivisibilities and fixed savings rates. In this paper Mookherjee and Ray ask whether poverty traps and persistent wealth inequality can result even with a convex technology and endogenous savings behavior derived from optimization of discounted future utility. In this model, agents face credit constraints owing to a combination of moral hazard and limited liability. Asset accumulation relaxes future credit constraints, providing a strong incentive for poor agents to save. But at the same time, the presence of moral hazard may endogenously generate nonconvexities in the returns to savings, constraining savings incentives of the poor vis-à-vis nonpoor households.

Mookherjee and Ray demonstrate that relative bargaining power between borrowers and lenders in credit markets

can have a decisive impact upon the resulting pattern of asset accumulation and long-run inequalities in the distribution of wealth. They find that in the case where lenders have all the bargaining power, poverty traps can emerge. This is because the surplus that accrues to moderate wealth accumulation of the poor is entirely extracted by the lender. This generates a nonconvexity in returns to investment that sharply inhibits the poor agents’ incentives to save. In contrast, wealthy agents are able to extract a sufficient proportion of their surplus and thus have a sufficient incentive to save; consequently, their wealth drifts upward over time. The long run wealth distribution is polarized between a class of poor agents with zero wealth, and a class of rich agents with high wealth levels, with no mobility between the two classes. On the other hand, if there is a competitive supply of lenders, all the benefits of saving accrue entirely to the borrowers, independent of their level of wealth. This is sufficient to preclude poverty traps: agents accumulate wealth indefinitely irrespective of their initial wealth, rendering historical wealth distributions irrelevant in the long run.

The central lesson emerging from Mookherjee and Ray’s paper is that institutional characteristics of an economy may play a central role in limiting savings incentives of the poor, thus perpetuating poverty. This suggests the need for greater attention to allocation of bargaining power between borrowers and lenders in studies of credit markets in developing countries, as well as in policy analyses.