This paper examines two issues associated with the impact of migration on household income and poverty. First, existing studies have typically overlooked a feature of migration that should be taken into account in estimating its impact, namely the fact that migration changes the size of the household. The 'corrected' impact that does take the change in household size into account is presented analytically and is estimated on the basis of data from Ghana's GLSS household survey. The corrected impact is shown to be three to five times larger for income and two to three times larger for poverty than is obtained from standard analysis. Second, existing studies examine migration's impact on the poverty of the entire sample. However, some policy questions require measures of the impact on the poverty of the migrant households themselves. The latter is shown to be twenty times larger for international migration and two to three times larger for internal migration, compared to the impact for the entire sample. It is further shown that these results hold whether the poverty measures are corrected for the change in household size.
migration income poverty underestimation reference groups