How should a researcher determine country-wide inequality? What can control for varying levels of household wealth in a regression analysis? How can wealth even be measured if income data is not available?
These are the questions that we (myself and four other authors) try to address in a paper that we just presented at MEA (Midwest Economic Association). We all work on the IPUMS-I project that gathers, harmonizes and distributes international census data to make it publicly available for research. Actually, next to the UN, its the largest database of international census data. This data is wonderful to have available, but one thing it lacks is any measurement of wealth. Only a few countries even have income data, but especially in developing countries, this data is subject to so much fluctuation and recall bias.
So, we create an asset-based measurement of socio-economic status. It takes into consideration things like dwelling characteristics (what is the material of your walls?), asset ownership (do you own a cell phone?) and access to utilities (does your household have electricity?). The wealth index then ranks households according to their answers to these types of questions.
A lot of the paper is about how we evaluate the index to confirm its validity as an accurate representation of socio-economic status. While this topic may not seem groundbreaking, the important addition here is that the census wealth index is actually a public good. The index will eventually be available in the (free) IPUMS-I data and could serve many a researcher using international census data seeking some measurement of wealth.
It's still a preliminary draft, but here is the paper.
These are the questions that we (myself and four other authors) try to address in a paper that we just presented at MEA (Midwest Economic Association). We all work on the IPUMS-I project that gathers, harmonizes and distributes international census data to make it publicly available for research. Actually, next to the UN, its the largest database of international census data. This data is wonderful to have available, but one thing it lacks is any measurement of wealth. Only a few countries even have income data, but especially in developing countries, this data is subject to so much fluctuation and recall bias.
So, we create an asset-based measurement of socio-economic status. It takes into consideration things like dwelling characteristics (what is the material of your walls?), asset ownership (do you own a cell phone?) and access to utilities (does your household have electricity?). The wealth index then ranks households according to their answers to these types of questions.
A lot of the paper is about how we evaluate the index to confirm its validity as an accurate representation of socio-economic status. While this topic may not seem groundbreaking, the important addition here is that the census wealth index is actually a public good. The index will eventually be available in the (free) IPUMS-I data and could serve many a researcher using international census data seeking some measurement of wealth.
It's still a preliminary draft, but here is the paper.
RSS Feed