This research is put forward in the spirit of a practical policy recommendation for developing countries and agencies. This means, however, that these recommendations must be subject to appropriate feasibility conditions. There is no value in a description of a desirable goal if there are no feasible means through which to arrive at this goal. The most general way of putting the requirement of feasibility is this: a goal is feasible if, given everything we know about social and economic processes and the motives and actions of typical agents, there is at least one pathway through which agents could act in such a way as to bring about the goal.
The ultimate test of the poverty-first prescription, then, is this: Do these policies lead to the desired outcome? Are they feasible, given the constraints of individual motivations and the workings of political and economic institutions? And are these policies consistent with the long-term interests of humanity-the requirement of intergenerational equity?
One of the attractions of economic development as an occasion for thinking about justice is that it poses a set of problems that are immediately and essentially practical. It is not desirable or useful to offer idealized theories of justice in this context; policy makers need solutions that can be implemented given existing social arrangements, allocations of political power, and the like. Questions about feasibility are essential. We want to know, not what the ultimate just society would look like, but what avenues of change are possible given our starting point, and which among attainable outcomes is preferable on grounds of justice. It is possible, for example, that the prescriptions of an ideal theory of justice-Rawls's formulation of the difference principle, for example-are inaccessible from certain starting points and a given set of constraints on policy. These considerations dictate that recommendations in this area must be subject to appropriate feasibility conditions. There is no value in a description of a desirable goal if there are no feasible means through which to arrive at this goal. The most general way of putting the requirement of feasibility is this: a goal is feasible if, given everything we know about social and economic processes and the motives and actions of typical agents, there is at least one pathway through which agents could act in such a way as to bring about the goal.
The ultimate test of the policies associated with the poverty-first prescription, then, is this: Do these policies lead to the desired outcome? Are they feasible, given the constraints of individual motivations and the workings of political and economic institutions? And are these policies consistent with the long-term interests of humanity-the requirement of intergenerational equity?
Economic change unavoidably occurs as the net effect of the independent activities of vast numbers of agents: consumers choose commodities based on incomes and prices, producers choose products and techniques on the basis of their expectations about future prices, and government agencies act on the basis of perceived interests of the agency. It is not possible for policies to be mandated by government without attention to the incentives, opportunities, and constraints that the policy will create for participants. There are substantial costs of enforcement that rise steeply the more contrary a given policy is to the interests of the agents whom it affects.
The central constraint is this: we may assume that agents have private interests (income, security, wealth, access to health care) that have a substantial or even dominant motivational role, and that agents are opportunistic: they will select strategies that permit them to take advantage of existing institutions and markets in such a way as to advantage their private interests. Policy instruments and goals that require that individuals ignore their private interests are infeasible (absent substantial coercion).
Consider the example of collective agriculture in China in the 1960s. Collective farms were created involving large tracts of land and the pooled labor of 100 to 300 farm families. The goal of this system was to reduce rural inequalities and to enhance efficiency through economies of scale. However, collective farms turned out to be highly inefficient, due in part to incentive problems: family income was determined largely by the average productivity of the collective, and there was little relation between quality and intensity of work and income. Each worker had little incentive to work efficiently or intensively, since the added product that resulted from higher quality work had the effect of raising the social average only slightly. The result was low quality work of low intensity. ([Putterman, 1985, Restoration; Oi, 1989]; see also essays in [Nee, 1989] for elaboration.)
Policies are implemented by governments, and governments are subject to a variety of political constraints. Within a democratic electoral system governments are unavoidably concerned about the effects of various policies on the electoral blocks upon which they depend; so it is difficult for a government based on the support of rice farmers to abolish input subsidies. But there are political constraints within authoritarian regimes as well. Such regimes rarely have sufficient power internally to dominate the whole of civil society, but depend rather on the support-explicit or implicit-of various interest groups within society. Once again, it is difficult or impossible for such a government to implement policies that are inimical to the interests of the groups upon which it depends.
In this context we may distinguish between varying levels of state autonomy with respect to powerful elites. In some instances a powerful group-for example, a landholding elite-may be in a position to simply determine state policy-through its dominant presence in a legislature, its ability to capture the presidency, or through other familiar mechanisms. In a wider range of cases, however, powerful elites will be in a position to block various possible state initiatives, without being in general able to dictate terms to the state. (For general theoretical discussions of this point see [Nordlinger, 1981] and [Migdal, 1988]; for application of this approach to Indian politics see [Kohli, 1987].)
Political feasibility constraints do not entail a wholly conservative development agenda, or one that is unavoidably hostile to the interests of the poor. For it is possible for disenfranchised groups to acquire political influence, permitting them to pressure governments to adopt policies that offend previously dominant groups. In order for this progressive outcome to occur, however, it is mandatory that there be some pathway by which the interests of the disenfranchised are converted into effective political resources.
Consider an example derived from Adam Przeworski. Suppose that our theory of justice dictates that the just society requires democratic socialism. And suppose the central constraint is that policies must be adopted and implemented within the context of democratic electoral institutions. Then Przeworski offers an argument to show that the socialist outcome is inaccessible from a starting point in which capitalist property owners exist; for a socialist regime may be voted into office, but once it begins to implement a program of socialist reform it elicits a strategic reaction from capitalists that produces economic crisis; during the crisis a reform capitalist program can outbid the socialist reform program for electoral support; and the process of socialist reform is interrupted [Przeworski, 1985, Capitalism].
A third constraint on poverty-first development strategies is economic: we must have good analytical reasons to believe that a given set of policies will in fact have the desired economic effects. The problem of institutional design-the creation of property arrangements through which efficient, modern agriculture may proceed while serving the ends of equity and welfare-is a knotty one. Moreover, it is a commonplace that economic policies often have unforeseen and perverse consequences. For example, suppose that we are interested in raising incomes to unskilled rural workers. We might consider introducing minimum wage legislation in farm labor. Assuming that labor-hiring farmers are faced with choices of crops and techniques, this policy may well have the effect of reducing demand for unskilled labor, leading in the end to a decrease in demand for labor and lower overall incomes flowing to workers in this sector.
A second feasibility constraint concerns economic growth. Economies with low per capita incomes are incapable of improving the welfare of the poor very much. So economic policies that are chiefly redistributive will fail in alleviating poverty; greater equality will mean little more than spreading poverty more evenly. Instead, the policies that are chosen must have the effect of producing sustainable economic growth. This means, in particular, that development strategies must be productive of sufficient growth as to permit expansion in real incomes. If a given recommendation has the predictable effect that it will lead to economic stagnation, then it is disqualified by this constraint.
A final constraint is normative, but derives from the interest that each generation has in the wellbeing of the next. This is the requirement that the policies adopted should be broadly consistent with the interests of future generations; policies adopted today ought not be such as to guarantee harm for future generations. Several factors fall under this consideration: resource utilization and conservation; preservation of air and water quality; preservation of wetlands and rain forest; sustainable patterns of urbanization and urban development; and sustainable population policies.