social costs. This situation had changed by the mid- 1980s. First, a consensus gradually developed on the appropriate policy mix: the reduction of fiscal deficits, the elimination of price controls and trade restrictions, the introduction of realistic and unified exchange rate regimes and positive real interest rates, and a shift to relying on the private sector rather than the state to run most productive enterprises. As this consensus became clear, policymakers could begin to focus attention on social welfare issues (Williamson, 1994). Second, by the mid- 1980s, as the results of adjustment policies became clear, a broad debate began about their social costs. Concerns about these costs were heightened by the publication of a highly critical study by the United Nations Children's Fund (UNICEF) in 1987 (Cornia et al., 1987). Initially, this debate focused solely on the social costs of adjustment policies and resulted in numerous efforts to develop safety net measures within the multilateral development banks. Yet by the late 1980s, the debate had been altered once again by evidence from country experience. It had become increasingly clear that the poor fared far worse in the countries that failed to adjust than in those that adjusted in a timely manner.
The contrasting examples of Chile and Peru are illustrative. During the adjustment crisis in the early 1980s in Chile, the poor were effectively protected from declines in social welfare through targeted employment and nutrition policies. With the subsequent resumption of growth in the late 1980s and early 1990s, poverty decreased substantially, falling from approximately 45 percent of the population in 1986 to 28 percent in 1994. In contrast, in Peru, during a prolonged period of ''postponed" adjustment from 1985-90 that resulted in hyperinflation and economic collapse, per capita consumption fell 50 percent on average. It fell even further for the poorest two deciles-—over 60 percent. Poverty rates rose from 17.3 percent in 1985 to 54.7 percent in 1990 (Marcel and Solimano, 1994; Glewwe and Hall, 1994).2 In Africa, meanwhile, the few countries that successfully adjusted also had better records on the poverty reduction front than did nonadjusters (Demery and Squire, 1995; Bruno et al., 1996).3
Adjustment is a necessary but not a sufficient condition for poverty reduction, however, and there are still many unanswered questions. The first of these is the relationship between poverty and inequality, and here the evidence is far less convincing, although some recent work clearly suggests that poverty increases with inequality (Birdsall and Londoño, 1997). Another important area that is increasingly recognized as critical to poverty reduction, but in which we have very little empirical evidence, is that of institutional reform