A Conversation with Cina Lawson, Togolese Minister of Digital Economy and Digital Transformation
CEGA hosted Cina Lawson, Minister of Digital Economy and Digital Transformation in Togo, for an in-person talk at UC Berkeley on October 13, as well as a keynote talk at CEGA’s Evidence to Action (E2A) symposium on November 4. The Minister joined us virtually from the COP26 summit in Glasgow — you can watch her remarks here.
During her time in office, Minister Lawson has prioritized Togo’s digital transformation and helped to reduce the digital divide for Togolese people. When COVID-19 began, she quickly spearheaded Togo’s flagship digital cash transfer program, “Novissi,” delivering cash transfers to over 920,000 of the poorest Togolese citizens. Today, Novissi is hailed as one of the most successful COVID-19 social protection programs in the world.
Since early 2020, CEGA Faculty Co-Director Joshua Blumenstock and his team have been working with Minister Lawson and the Government of Togo to guide the expansion of Novissi using machine learning and non-traditional data sources to target the country’s most vulnerable citizens. This blog is a compilation of remarks delivered by Minister Lawson during her visit to UC Berkeley, her keynote at E2A 2021, and email communication.
Read more about this exciting collaboration here and CEGA’s Targeting Aid Better initiative here.
How did your partnership with CEGA start?
At the start of the pandemic, the Togolese government started to lock down districts based on COVID-19 case counts. With 80% of citizens living in extreme poverty (less than $1.90 a day), we needed to find alternative measures of support, in addition to the electricity and water subsidies we were already providing. We determined that unconditional cash transfers made the most sense and launched the Novissi program, a 100% digital universal basic income program.
When the program started, the cash was given to informal workers, who were identified by leveraging a voter ID database that had information on the voter’s profession, correct?
Indeed, soon after the launch, the Government of Togo (GoT) started thinking about how to target the poorest individuals and prioritize them for cash transfers using artificial intelligence and alternative data such as call detail records (CDR). This was in anticipation of a worst-case scenario where the whole country or at least multiple localities need to be placed under curfew and locked down due to the pandemic.
We wanted to help the most vulnerable Togolese, but we had no idea who they were. We contacted Ester Duflo (MIT, French Development Agency) for help and she connected us with Professor Joshua Blumenstock and CEGA. Josh and the CEGA team first helped us determine the poorest districts in Togo using satellite imagery. Once we knew where the poorest people lived, we then needed to know who they were. To do this, we used cell phone metadata and machine learning to identify individuals earning less than $1.25 per day and living in the poorest areas.
How did working with Josh [Blumenstock] and his team help you improve the targeting for Novissi?
In Togo, there is no social registry or up-to-date census data, which makes it challenging to target the most vulnerable for social protection benefits. Josh’s research and related approach to targeting changed this. Using AI wasn’t perfect, but it was proven to be more efficient than our initial approach. So Togo became a country using AI to fight poverty. In November 2020, we were able to test this AI-powered targeting approach for the first time in real life: the American non-profit GiveDirectly agreed to give Novissi ten million US dollars to distribute in cash transfers to 140,000 people who would be identified exclusively with this pioneering approach.
After this massive success, what’s next for Togo?
Now that we know our targeting and digital infrastructure works, we have plans to scale this approach across government programs. First, we will digitize all social protection programs and Government to Person (G2P) and expand mobile payment platforms, building on the Novissi model, which leverages CEGA research.
Next, we provide digital identification for all. In 2021, we will implement “E-ID Togo’’ for all citizens aged five and up, to better target social protections and revolutionize service delivery in Togo across a person’s lifespan. Finally, we will harness data-driven approaches to policy-making.
Through our collaboration with Josh and CEGA, we have learned that big data analysis and ML can unlock fresh insights into poverty reduction. We need to transition to data-driven policy-making in Africa. And Togo can serve as a model for other countries.
What can Africa learn from Togo?
You need to be able to support your citizens in crisis situations. You need to make people understand that they belong together, that they are part of a nation state. One way for them to understand that is for you to support them when they need assistance.
Togo is just one example where Novissi and data-driven policy worked. Right now [Josh’s] algorithm is trained on Togo data, but if you look at the world’s poorest, they’re in India and elsewhere in Africa. If we could replicate the same approach [in other countries] as the one we did in Togo, we would have a very large sample size and get a better understanding of poverty.
How can data science and researchers help?
I think it’s really important for African countries to have their local teams of data scientists. I’m imagining a team of data scientists that reside in each country, and work on digital transformation and machine learning projects across the different ministries, on a case-by-case basis. It’s important that centers such as CEGA be in Africa.
