Martin Ravallion

The economy is rarely a passive vehicle for transmitting a policy initiative to its
target. People react to policy changes. A policy aimed at one class of people can
easily bring with it unexpected gains and losses to others. Nor is it always clear
how the target group will be affected, once the economy as a whole has reacted.

In targeting policies towards the needs of the poor in a market economy there
will generally be some prices and quantities which appear in their budget
constraints but are not controllable by the policy maker. An assessment is
called for of how those prices and quantities will respond to the policy
intervention before one can determine how welfare of the target group will be
affected. Policy design may also be constrained by effects on the welfare of
non-target groups. Only policies which keep these effects within certain
bounds may be politically feasible in specific settings. For example, in
designing policies aimed at raising the welfare of the rural poor one may be
constrained to leave the post-tax wage of urban workers unaffected.

This chapter surveys and analyses a broad range of anti-hunger policies in
market economies, looking particularly at the way market responses can
influence policy design. The method of analysis draws on both theoretical and
empirical arguments with the latter applying almost exclusively to South Asia,
particularly Bangladesh. The policies examined include direct transfer pay-
ments, public employment ('relief work' hereafter), food pricing policies,
public buffer stocks, external trade policies, and public information policies.
The survey is far from even handed. One policy is given more emphasis than
the others, namely relief work, while another potentially important one is not
discussed at all, namely land reform. This bias seems in keeping with current
emphasis amongst policy makers and this is understandable; in the main, the
political pre-conditions for successful land and tenancy reforms have not been
present, while there seems more immediate hope for relief work and similar
income transfer policies.

A theme of this study is that anti-hunger policies should be concerned with
two conceptually distinct aspects of individual consumption: its level over a

-241-