Page 246 - Contributed Paper Session (CPS) - Volume 3

CPS2010 Rodrigo L. et al.
an index based on household characteristics, durables, and household heads’
education. Their results show declines in poverty during the nineties decade,
mainly due to advance achieved in rural areas. However, Sahn and Stifel (2003)
investigate urban-rural differences using several living standards indicators,
including an asset index, and find gaps that are not diminishing over time.
Finally, Booysen et al (2008) create an asset index to examine changes in seven
African countries and conclude that progress is mainly associated to
accumulation of private assets, as compared to the setback in the access to
public services. This paper will close a gap in the literature by carrying out a
similar analysis for the Latin America region using census microdata.

2. Methodology
This study uses census microdata from the IPUMS International project,
the largest repository of international census samples. Household wealth is
observed across the 1990, 2000, and 2010 census rounds. The data include 24
census samples from 10 countries, which cover at least for two of the
referenced rounds. The following censuses are employed for the analysis:
Argentina 1991, 2001, and 2010, Bolivia 1992 and 2001, Brazil 1991, 2000, and
2010, Chile 1992 and 2002, Colombia 1993 and 2005, Ecuador 1990, 2001, and
2010, Paraguay 1992 and 2002, Peru 1993 and 2007, Uruguay 1996, 2006, and
2011, and Venezuela 1990 and 2001. The samples for each of these censuses
represent 5 to 10% of the country's population. The data provided by IPUMS
are harmonized (i.e. it uses a consistent coding structure for the same
variables), which facilities the construction of indicators and enhances
comparability for the analysis.
Household wealth is defined using dwelling characteristics and human
capital. We identified a set of indicators at the household level that are
common across all datasets. The choice of these indicators is based on
previous literature (Booysen et al, 2008; Lovaton et al, 2014; Sahn & Stifel,
2000) and it also follows some indicators comprised within the Sustainable
Development Goals (SDGs) framework. The calculation of indicators not only
uses consistent definitions across countries and census rounds to achieve
higher comparability, but it is also based on harmonized microdata offered by
IPUMS. The set of nine non-monetary wealth indicators analyzed are: 1)
whether the dwelling is owned or rented, 2) access to electricity, 3) access to
piped water, 4) connection to public sewage, 5) whether the dwelling has a
toilet or a bathroom, 6) finished floors, 7) cement, brick or concrete walls or
roof, 8) number of persons per room, and 9) years of schooling of the
household head. A summary measure is produced to assess overall household
wealth through the aggregation of these nine indicators. This measure is
defined as a linear combination of the indicator variables by applying

235 | I S I W S C 2 0 1 9