This paper utilises the expenditure data from 2004–05, 2009–10, and 2011–12 to impute household expenditure into a survey of durable goods expenditure conducted in 2014–15. The model’s predictions are comparable to the World Bank’s current adjustment method for the rural areas but imply a slower rate of poverty reduction for urban areas. In two validation tests, using past data, three alternative model specifications perform worse than the preferred model. The analysis indicates that survey-to-survey imputation, when feasible, is a preferable alternative to the current method of adjusting survey-based poverty estimates to later years.
