Digital civil registration and legal identity systems: A joined-up approach to leave no one behind: Appendix 2
This report uses two major data sources for estimates on birth registration for the poorest 20% – USAID’s Demographic and Health Survey (DHS) or UNICEF’s Multiple Indicators Cluster Surveys (MICS). These surveys do not provide a comprehensive dataset because they:
- Are not carried out in high income countries
- Survey only households (therefore missing homeless populations, displaced populations and people in institutions)
- Do not ask about death registration
- Are only carried out periodically.
The likely result of this missing data is that global-level estimates will understate gaps between the poorest 20% of households and the rest of the population. Typically, birth registration data for high income countries indicates that rates are near 100%. Since high income countries have only a small share of the poorest 20% compared with the rest of the world, these high birth registration rates would increase the average rates of higher income populations and increase the gaps between the poorest and the rest of the population.
DI obtained the microdata for all available MICS and DHS surveys historically and sought to identify any birth registration questions. The microdata allowed DI to examine household characteristics among those who had births registered and those who did not. In particular, it allowed for disaggregation of birth registration coverage by household wealth. DHS and MICS have a series of questions about items that a household owns, such as a car, radio or television. The DHS program has developed a method to create a wealth index for each household based on the ownership of these items.
The World Bank’s PovCalNet database includes estimates of the percentage of the population below a certain poverty line. When a poverty line is selected that would include 20% of the global population, it is possible to see what percentage of a country’s population would be in the global poorest 20% of households. For instance, the most recent data on PovCalNet (from 2015) indicates that 20% of the world’s population lived below the poverty line of $2.68 per person per day (based on purchasing power parity for 2015). For 2015, 68% of the population of Benin fell below that poverty line. Assuming that PovCalNet represents the true distribution of the poorest 20%, it is possible to look at the bottom 68% of Benin’s population in a nationally representative dataset and make an assumption that they are in the global poorest 20%. In this case, this analysis does this by using the DHS/MICS wealth scores. In the case of Benin, this would mean taking the households in the bottom 68 percentile in the 2014 MICS and assuming they are in the poorest 20%. While different definitions of wellbeing would put different populations in the poorest 20%, the approach used here aims to provide indicative trends. DI uses a similar approach to estimate if a household is identified as being above the international extreme poverty line of $1.90 per person per day.
The authors of this report have made some assumptions to standardise responses from different surveys. The DHS and MICS programmes provide questionnaires that are standardised; however, with MICS there is much more variation and variable names are not as standardised (as they are with DHS). While DHS has made more effort to standardise responses, MICS typically maintains variations in variable names and responses based on context-specific adaptations. Some variables required little more modification than translating variables and responses (such as sex). On the other hand, variables such as education have highly context-specific responses. To standardise globally, this report followed the International Standards Classification of Education, as used by UNESCO and other international bodies. Education levels were categorised as ‘less than primary’, ‘primary’, ‘secondary’ or ‘higher’. In some cases, questions classified primary together with less than primary. For the purposes of this report, they were combined as ‘less than primary’. In cases where secondary was grouped with post-secondary, these were classified as secondary. Informal education was classified as a missing variable.
The MICS or DHS wealth index was used as a given. When it was not available, principle component analysis methods were replicated using dummy variables for assets. These assets were: radio, car or truck, television, electricity, computer, watch, bicycle, refrigerator and mobile phone.
The report categorised a respondent as having a birth registered if they reported their registration (civil or otherwise) or if they had a birth certificate (regardless of whether the survey enumerator saw it).
The DHS and MICS surveys are not carried out annually, with few exceptions, meaning that income distribution data and birth registration rates are interpolated. The report takes the World Bank method for providing period distributional data as correct. Surveys were then backcast or forecast within a five-year period for population estimates. If a country did not have data on birth registration within that period, that country was dropped from aggregate estimates. For China, there were no DHS or MICS data. The authors reviewed the China Family Panel Study, which has questions about the Hukou system. This system was not deemed similar enough to birth registration in other countries for China to be included in aggregate estimates. Additionally, aggregate estimates excluded high income countries which did not have surveys with birth registration questions available for use. This means that aggregate trends in this report do not include these populations.
Download this reportDownload now
There are several significant assumptions in this approach. For a discussion of these assumptions, see Development Initiatives, 2018. Coding the P20: How we developed and coded the P20 Initiative. Available at https://devinit.org/publications/coding-p20-developed-coded-p20-initiative/Return to source text