For comparability of the poverty estimates based on the panel data with the ―official‖ estimates based on the full FIES, the panel income data are calibrated in such a way that the poverty-incidence estimate from the panel data for 2006 is approximately equal to that from the full 2006 FIES data. All poverty estimates are based on official poverty lines for 2006. For consistency, these lines are held fixed in real terms. By construction, the resulting poverty estimates are not strictly comparable with officially published poverty estimates which are based on time-varying poverty lines (i.e., the welfare standard for poverty comparison varies from one survey year to another).