摘要：In developing countries poverty is generally measured with expenditure data. Such data are difficult and costly to obtain and it is generally recommended to collect them every 3-5 years. In between surveys, however, there is a clear need to provide policymakers with information for the monitoring of poverty trends. The paper reviews several such methods and compares the poverty estimates and trends resulting from their application to a panel dataset for Albania. The results are broadly consistent across methods and point to an overall improvement in welfare conditions over time, although the magnitude of the changes differs by locale, with urban areas showing a larger improvement than their rural counterparts. However, given the sensitivity of the predictions to model specification, it is worthwhile to construct alternative welfare indicators and triangulate results. Lacking a gold standard measure, the use of a suite of welfare indicators, if duly validated, could be a viable approach to monitor poverty trends; some caution should be exercised in drawing conclusions about the actual magnitudes of the changes.
