increase in household consumption, considerably more than the gain under a targeted cash transfer policy of equal cost. However, given that the petroleum subsidy does not specifically target the poor, it only reduces poverty by a fraction of what a targeted cash transfer would. Moreover, as illustrated by the experiences of other countries, increases in a fuel subsidy, once introduced, are likely to be very difficult to reverse. This suggests that if the major policy goal is to reduce poverty, a direct cash transfer would be more effective than the other policy options considered here. Combining these policies, however, would be even more effective than any single intervention, reducing poverty incidence by around 2.5 million people, and thereby preventing nearly all of the potential increase in poverty resulting from global price shocks.