Are there dynamic gains from a poor-area development program?

_s I6S POLICY RESEARCH WORKING PAPPER 1695 Are There Dynamic Gains Are Drograms from a Poor-Area Vir afs.ordo Development Program? gains Jyotsna Jalan Martin Ravallion The World Bank Policy Research Department Poverty and Human Resources Division December 1996  l POLICY RESEARCH WORKING PAPER 1695 Summary findings Poor-area developnment programs -in which the household level. Second, whether an area is targeted government transfers extra resources to unusually poor appears to depend heavilly on those same area areas - have been widely used in efforts to fight characteristics. The-re is a large omitted-variable bias in poverty. There has been some researcn on such prograrns assessments of program irmpact that do not take account but we know little about their impact on household of differing local conelitions. living standards over time. jalan and Ravatiun use the Generalized Method of Jalan and Ravallion address the issue for a sizable Marments to get a consistent estimate of a dynamic poor-area development program in rural China. China's model of cons-urrpt on growth at the farm-household program had a significant imripact on rural living level. Growth is a funct.on of whether the county of standards in the targeted areas of the sample. But this residence is included ir the program, plus community fact was not evident by comparing the growth rates of and household Ive eit al conditions and covariate consumption in targeted areas with growth rates shocks to fanr y elds elsewhere. A simple comparison of the growth rates The consurnprzon-growth model suggests that suggested that although the programz was successfully households living ir areas targeted by the program had a targeted to poorer areas, it did nothing to improve their higher rate of consurr.ption growth than one would have position relative to other areas. There were strong signs otherwise expected '07ithout the programi, the initial of unconditional (absolute and relative) divergence over conditions in those arcas appear to nave been so time between the areas covered by the program and unf1avorable that ;ev w-ould have seen a decline in those that were not. average livi-ng star.dCads over the period studied. The There are two reasons why these signs of gains were erough to ontevnt arn absolute decline, but unconditional divergence are so misleading about not enough over five ,-ears to reverse the strong program benefits. First, there are external effects of underlying divergent terindncies in the rural econony. initial area characteristics on consumption growth at the This paper -a product of the Poverty and H-umran Resources Division, Poiicy P.esearch EDeprtment -is part of a larger effort in the department to evaluate tlhe welfare impact of public programns. -he st-udy wlsas 'unded by the Bank's Research S,upport Budget under the research orciect "Policies for Poor Areas" (RPO 678-,).