Apples to apples — PovcalNet introduces a new comparability indicator

This blog does not refer to the latest information on comparability. More information about the comparability database and how to use it is available in PIP’s Methodological Handbook. Comparability is also indicated in the main output on the PIP website, the PIP Stata command and the PIP API.

As part of the World Bank’s September 2019 global poverty update published a few weeks ago, metadata on the comparability of poverty estimates within countries over time was added to the PovcalNet website. This metadata has been compiled by the Country Poverty Economists in the World Bank’s Poverty and Equity Global Practice. In this blog, we illustrate how to use this metadata through the R package povcalnetR and the Stata command povcalnet.

As countries frequently improve household surveys and measurement methodologies, strict comparability of poverty estimates over time is often limited. Strictly comparable poverty estimates within a country require a consistent production process, including the sampling frame, questionnaires, the methodological construction of the welfare aggregates and poverty lines, a consistent deflation of prices in time and space, among many other considerations. Thus, the assessment of comparability is country-dependent and relies on the knowledge of the Country Poverty Economist and Regional Poverty Teams of the Poverty Global Practice, as well as close dialogue with national data producers with intimate knowledge of the survey design and methodology. Within a country, we assume comparability of poverty estimates over time unless there is a known change to survey methodology, measurement or data structure. More details on the comparability metadata database can be found in Atamanov et al. (2019) (section 4). The database can be downloaded in csv.

In this blog, we look at changes in inequality for Argentina, Ghana and Thailand since 1990. For these three countries, the comparability database provides the following information (showing only selected years):

Source: Comparability database (csv), showing selected years only.

Within each country, the comparability variable starts with the value zero (0) for the oldest comparable series. When comparability is broken, the value changes to one (1) for the year of the break and it goes on until the comparability is broken again in a subsequent year. In the table above, Thailand starts with a 1 in 1998, since there was a break in comparability in 1990. In 2000, there is another break and the survey remains comparable until 2013, when there is another break.

PovcalNet reports several measures of inequality, such as the Gini index, the mean log deviation and the income or consumption shares of the decile groups. The following command line loads all available inequality measures in the survey years for Argentina, Ghana and Thailand:

povcalnet, country(arg gha tha) year(all) clear

This database can be easily merged with the comparability metadata (replication code for R and Stata). We first look at the trend in the Gini index for the three countries since 1990, separating the comparable spells for each country. In Argentina, inequality rose sharply until 2002, by almost 10 points. After 2003, the situation reversed with a strong and continued declined in inequality. Argentina and Thailand started from similar levels of the Gini index in the 1990s, but Thailand saw a largely downward trend. In contrast, inequality widened in Ghana between the early-1990s and mid-2000s, with a plateauing in recent years.

PovcalNet also reports the income or consumption shares of the decile groups, which can be used to analyze distributional changes in more detail. In the next figure, we have drawn (anonymous) growth incidence curves (GIC) for these three countries, looking at the longest comparable spell for each country. The GIC shows the growth rate of mean income between the initial and final year (varies across countries) for a given quantile group (e.g. the bottom 10%). A downward-sloping GIC is associated with a pro-poor or equalizing pattern of growth.

Between 2003 and 2017, Argentina showed strong income growth, which was pro-poor. While the incomes of the bottom 10% grew at close to 8% per year, the top 10% grew at less than 2%. Over roughly the same period, consumption growth in Thailand was also strong and pro-poor, but gains were more evenly distributed than in Argentina. Beginning in an earlier period in Ghana, growth was pro-rich with the top 10% growing at a rate close to 4% and the bottom 10% growing at only 2%. These patterns in the growth incidence curves are consistent with the trend in the Gini index observed in the first Figure. In Argentina, the Gini index decreased strongly from 51.2 in 2003 to 41.2 in 2017. Thailand saw a smaller decline from 42.8 in 2000 to 37.8 in 2013. In contrast, the Gini index in Ghana increased from 38.4 in 1991 to 43.5 in 2016.