Harold Coulombe
Thomas Otter

National Statistical Office (NSO)
United Nations Development Programme (UNDP)

This paper documents the construction and shows some results of a poverty map based on data from the 2002/03 Living Standards Measurement Survey (LSMS) and the Housing & Population Census 2000. Based on a methodology developped by Elbers et al. (2002, 2003), we calculate poverty indicators at low levels of aggregation, using the detailed information found in the survey and the exhaustive coverage of the Census. Results at region, aimag and soum levels are presented.

In the past decade poverty profiles have been developed into useful tools to characterize, assess and monitor poverty. Based on information collected in household surveys, including detailed information on expenditures and incomes, those profi les present the characteristics of the population according to their level of - monetary and non-monetary - standard of living, help assessing the poverty reducing effect of some policies and compare poverty level between regions, groups or over time. While these household survey-based studies have greatly improved our knowledge of welfare level of households in general and of the poorer ones in particular, the approach has a number of limitations. In particular, policy makers and planners need finely disaggregated information in order to implement their anti-poverty schemes. Typically they need information for small geographic units such as city neighbourhoods, towns or even villages.

Telling a Mongolian policy maker that the neediest people are in the rural areas would not be too impressive as that information is well known and not useful since it would be too vague; telling them in which aimags or even soums the poorest households are concentrated would be more useful and convincing! Using aimag-level information often hides the existence of poverty pockets in otherwise relatively well-off aimag which would lead to poorly targeted schemes if soum-level information is not used. Having better information at local level would necessarily minimize leaks and therefore permit more cost-effective and effi cient anti-poverty schemes. Poverty indicators are needed at a local level as spatial inequalities can be important within a given region.

The methodology used have been developed by Elbers, Lanjouw and Lanjouw (2002, 2003) and should be seen as more sophisticated than other methods3 as it uses information on household expenditure, is fully consistent with poverty profi le fi gures, and permits the computation of standard aerrors of those poverty indicators. Since those types of poverty maps are fully compatible with poverty profi le results, they should be seen as a natural extension to poverty profi les, a way to operationalise poverty profi le results. This report documents the construction of the Mongolia poverty map but the map would reach its full potential once a series of applications under consideration would be undertaken.

The remaining of this paper is structured as follow: we first present the methodology in layman words, follow by a description of the data used. The paper ends by a discussion of the results – including gender-specifi c ones – and on furthers work to undertake. A more technical presentation of the methodology can be found in Annex 1; along with some more detailed results (Annexes 2 to 4)