Distribution of Colombian population’s socio-economical conditions according to Sisbén IV. A- Extreme poverty. B- Moderate poverty. C- Vulnerable population. D- Population not affected by poverty. The blue dashed line represents the limit up to which Ingr
While the score criteria were made public, people were completely left in the dark as to what these scores meant in the first place and because some of the databases used were completely unknown to the population affected, people were not able to access or rectify any outdated or inaccurate information used to make a decision on their situation. After the government published the list of beneficiaries and after numerous citizens reporting being excluded, it was concluded there were nearly 17,000 records with inconsistencies. Because the system is a “black box” it was – and is – impossible to know how many people were unfairly excluded in the crosschecks between different databases.
In Myanmar, the government similarly disbursed payments to households in poverty without disclosing the eligibility criteria or the relevant factors relied upon to calculate the amount received by any recipient household. Myanmar’s Covid-19 Economic Recovery Plan (known as CERP), launched in April 2020, announced the introduction cash transfers of “an appropriate amount (…) to most vulnerable and affected households” through mobile cash transfers. However, no clarity was ever provided as to how the most vulnerable and affected were identified, and the basis on which the “appropriate amount” was calculated.
Other countries have taken similar routes, holding onto the idea of a shiny innovative solution while leaving their populations in the dark about how these algorithms make decisions.
In Nigeria, the Federal Government’s COVID-19 Cash Transfer Project which aimed to lift “the urban poor” affected by the pandemic out of poverty was the first strategy to be developed and tested in the Sub-Saharan Africa region. According to official statements this solution relies on methods such as satellite remote sensing technology, machine learning and big data analysis in order to identify potential beneficiaries. To start, the relevance of these methods to identifying those in need was not explained. However, more worryingly, nothing was made public regarding the criteria utilised to select who this benefit was targeting, which makes it impossible to scrutinise and audit. The pride in being the first country in the region to deploy such a solution should come accompanied by additional safeguards and processes to ensure the solution is adequate and inclusive before it is innovative.
It is also worth highlighting two other countries where beneficiaries were left with no clarity on how their living situations were concretely assessed. In Mozambique, the government decided to identify priority areas by using the MultiDimensional Poverty Index mapping. This index combines social and economic indicators and relies on the data gathered in the most recent census as well as high resolution satellite imaginary of urban poverty maps. In this case, the vulnerability criteria used by the National Institute of Social Action to identify potential beneficiaries was made public but still an independent observer stated that only 61% of beneficiaries knew why they had been enrolled. Similarly, the observer highlighted ongoing complaints about the lack of information on the program including who’s eligible, how much they’re entitled to and with what periodicity the benefit is to be distributed. Similarly, a welfare benefit introduced by the Peruvian government not only relied on unclear eligibility criteria, but has been shown to exclude many. Within the “Stay at Home” program, the Peruvian government identified eligible people through the National Household Register, which categorises households into different socioeconomic bands. Among other factors, households categorised by the Register as being in poverty are deemed eligible to the benefit. However, in a similar fashion to Colombia’s case, the criteria used to define poverty are not publicly available, making it difficult for people to assert their entitlement to the benefit and leading to individuals being deemed not eligible despite being in a situation of poverty.
Data-intensive methodologies
After defining the eligibility criteria, governments must ensure that they correctly identify vulnerable populations. For some governments the way to take this task forward was to collect and make use of as much data as possible.
