A Statistical Model for Simple, Fast and Reliable Measurement of Poverty

The primus inter pares of the UN Millennium Development Goals is to reduce poverty. The only internationally accepted method of estimating poverty requires a measurement of total consumption based on a time and resource demanding household budget or integrated survey over 12 months. Rather than measuring poverty only, say every 5th year, a model is presented to predict poverty based upon a small set of household variables to be collected yearly between two 12 months household surveys. Information obtained from the light surveys may then be used to predict poverty rates. The key question is whether the inaccuracy in these predictions is acceptable. The standard errors presented are lower than the sampling errors to the poverty estimates based on the 12 months household surveys. Predictions based on this sample also indicate that the problem of misspecifications of models is not large. It is recommended to test these models at the country level and if the test results are comparable to those here, apply the approach presented.