It seems obvious that economic growth should reduce poverty, yet the issue remains controversial. Some scholars assert that economic growth does not eliminate poverty and may exacerbate the problems of the poor (United Nations 1997). For example, Dreze and Sen (1990) claim that economic growth does not generate benefits in terms of numerous nonpecuniary measures of well-being. Calls for increased government spending (Squires 1993) or other redistributions of wealth (Todaro 1997) are the logical extension of the argument that growth does not ensure the elimination of poverty.
Todaro (1997) labels the contention that growth actually reduces poverty as the "trickle-down theory." In the less than idealized state of affairs, there is not even a "trickle" downward. Simply put, general economic progress does not "improve the levels of the very poor" (Todaro 1997: 155). In fact, some development economists contend that the "growth processes" typically "trickle-up" to the middle classes and "especially the very rich" (Todaro 1997: 163).
A largely unexamined issue is the impact of the relative wealth of the rich and poor on the level of well-being. There is a substantial literature that asserts that improving the incomes of the poor has a greater effect on the average level of well-being in a country than on improving the incomes of the rich (Todaro 1997). That proposition, however, has not been exhaustively examined, and more careful analysis constitutes an important research agenda.
Wealth Distribution and Poverty
The first question regarding the relationship between the rich and the poor can be examined by estimating the relation of the incomes of the poor and rich to each other. For example, we can estimate the following equations:
(1) [Y.sub.ip] = [alpha] + [[beta].sub.r][Y.sub.ir] + [[epsilon].sub.i]
(2) [Y.sub.ir] = [alpha] + [[beta].sub.p][Y.sub.ip] + [[epsilon].sub.i],
where [Y.sup.p] and [Y.sub.r] represent the per capita incomes of the poor and rich, respectively, and [[beta].sub.p] and [[beta.sub.r] represent "class income transfer" coefficients. The [beta]s show the proportionate increase in one group's per capita GDP as a function of the other group's per capita GDP. For example, [[beta].sub.r] represents the change in the income of the poor attributable to the change in the income of the rich. If trickle-down is true, [[beta].sub.r] should be positive. If trickle-up is true, the coefficient should be negative.
In estimating equations 1 and 2, there is the potential problem of additional variables that conceivably also affect incomes. Variables that are important are examined in Barro and Sala-I-Martin (1995), and include human capital, institutions, and other variables. Presumably, those influences are reflected in the income of the other class. Thus, the only required additional variables would be variables often omitted in cross-national growth equations. The most conspicuous variables are geographic. Many of the poor countries of the world are located in tropical environments and are landlocked. Sachs (1997), Sachs and Warner (1997), and Sowell (1994) argue that these factors are especially debilitating for human well-being. Landlocked countries are often isolated from commercial practices, ideas and innovations, and market enhancing institutions. Tropical countries frequently experience diseases, lack of sanitation, and famine. Those conditions threaten the inhabitants' survival, adversely affect incomes, and perpetuate poverty. Finally, Lucas (1988) argues that urban economies entail higher productivity than rural economies because of externalities attributable to more productive human capital.
The second question regarding trickle-down is more direct and entails the relationship between poverty and the relative incomes of the poor and rich. Consider a simple model of the average level of human poverty (HP) in a country:
(3) H[P. …