Advancing responsibility in the digital public infrastructure (DPI) space
Reflections from the 2024 Responsible Finance Forum
The push to expand digital public infrastructure (DPI) has been gaining momentum, with many governments, donors, and international organizations enthused by its potential to advance financial and social inclusion. This growing interest brought DPI to the Responsible Finance Forum 2024 agenda. Leading up to the event, the Center for Financial Inclusion published a background paper outlining current debates in the space. At the event itself, participants heard from Mayara Yano about the ingredients that produced the enormously popular Pix interoperable payment service in Brazil. They also heard a cautionary tale from my colleague, Nanjira Sambuli, and me about what happens with DPI when the governments pursuing them are not public value maximizers. Buoyed by international momentum, the Government of Kenya is pursuing a digital ID framed as DPI. The new ID offers little marginal value beyond the current system, imposes new costs on citizens, fails to address existing shortcomings of the existing ID system, and introduces new risks around data protection and misuse.
The contrasting cases of Brazil’s focused, intentional, and successful DPI, and Kenya’s high-risk DPI journey raise important questions about the conditions under which governments can be successful facilitators of these kinds of systems in ways that generate citizen and user value. Certainly, proponents of DPI have long acknowledged the risks such systems entail (for example, here, here, and here). However, the approach to mitigation has thus far been focused on identifying soft sets of vague principles or technical safeguards recommended for countries to self-impose to limit expected harms and data misuse. While helpful conversations, such approaches seem to take for granted that all government facilitators are both committed to delivering high quality systems that maximize public value and that accountability environments are sufficient to manage risk by ensuring these types of principles and safeguards are implemented seriously. It was enlightening to consider Brazil’s selective pursuit of DPI to fit its needs—focusing on payments and open banking—alongside Kenya’s apparent interest in adopting the “package” of DPI pillars (digital ID, payments, and data exchange) regardless of where there is a strong need and opportunity to deliver value. That raises questions about the scope of DPI and the merits of pitching it as a package with essential elements relevant regardless of context.
As I reflect on the presentations and conversations at the 2024 RFF, I recognize four themes for deeper research and conversation needed to move us towards responsible DPI. Many of these themes rhyme with longstanding conversations and debates from financial inclusion, inviting us to remember all that we’ve already learned, even as we move into this new era.
Theme 1: Centering outcomes over methods and tools.
Over many decades, financial inclusion practitioners have been taught the hard lesson that there is no technical silver bullet for financial inclusion, and more importantly, financial health. The sector moved from a focus on microloans to bank accounts and savings, to mobile and digital payments, to digital credit and e-commerce in waves only to learn in each successive wave that user financial needs must take precedent over any kind of product. We learned over and over again that leading with the favored tool of the era was a recipe for short sightedness, wasted resources, and costly mistakes that sometimes harmed users.
One of the most important success factors of Brazil’s Pix project is the Central Bank’s insistence on user centricity. They pursued this project to deliver low cost, ubiquitous payments to all Brazilians and organized themselves to deliver value from day one. Before the technical design of the product, they conducted detailed user research including payment diaries to understand user behaviors and pain points. They analyzed other payment channels to understand their value proposition better. This all fed into Pix’s technical design, but it didn’t stop there. The team relentlessly engages with its members who offer Pix through their channels, with users of the system, and with data on usage and challenges to constantly iterate to address emergent challenges, shortcomings, and risks. For the Central Bank, DPI is not the goal; the goals are consumer value and financial health.
Pushing for speed in DPI adoption and advancing DPI as a package of pillars of interventions that appear to be context-blind can jeopardize this kind of attention to outcomes. I worry that initiatives explicitly encouraging speed and scale—like “50 in 5” create incentives to prioritize the adoption of the “DPI package” over user value. I fear they create accountabilities to the tool and to the international community rather than to citizens and their outcomes. Where risks are substantial, might slow, thoughtful policy be more responsible than “moving fast and breaking things”?
Theme 2: The urgent need to deepen understanding.
A second theme – again with echoes from financial inclusion experience – is the urgent need to deepen our understanding of DPI as a sector. In financial inclusion, to understand the value of digital payments, we needed to understand the landscape of informal finance and remittances that preceded it. It was this expansive view that led to value adding innovations. An important step in deepening our understanding of DPI is broadening our scope for what kinds of initiatives fall within the landscape of interventions we care about. Definitional debates about what constitutes DPI are important, but narrowing the frame of what “counts” as worthy of our attention is risky. It may lead us to look at only those cases that fit our normative ideals, and leave out those that fall short. It can exclude non-DPI government-led digitalization experiences that offer rich lessons about conditions required for states to build citizen/user centric digital infrastructure and the fuller nature of emergent risk. I would argue that we need to broaden our analytical view to include state-led digitalization of all sorts to deepen our understanding of what delivers success and value. Within that wider landscape, we can identify DPI is a subset to recognize what truly sets it apart in theory, practice, and utility.
It also means expanding the set of case studies for our consideration and reflection beyond the few big wins. Our understanding of all of these cases could be enriched by rigorous research that seriously interrogate processes, value, and citizen outcomes quantitatively and qualitatively. Systemically considering cases based on the same criteria could be enormously useful in guiding where resources should be focused in ways that will really matter for the outcomes we care about.
And, we need to take an expansive, long-term view on these impacts, including not just uptake, usage, and costs of proximate usage, but also general equilibrium effects. What do DPIs lead to in terms of competition? Are they making it easier or more difficult for smaller businesses to thrive? Where they are the foundation for new systems of taxation and credit, what are the distributional effects of these new systems?
Theme 3: Making wise choices about where and how to support governments.
Nanjira Sambuli and I presented our hypothesis that high-value DPI is a function of two local political conditions: states themselves being fully committed to high value implementation and a strong accountability environment holding them accountable to delivering value and managing user risks. If true beyond our admittedly small set of current cases, this means that the international community of promotors and funders should think quite carefully about which contexts are truly ready to implement responsible DPI.
In my view, this may be one of the most important ways to advance responsibility in DPI. DPI – even under the best of circumstances – entrusts new power to states. How will that power be used in the short and long term? I do not think we fully know. We know in some places, digitalization has enabled increased surveillance and interference with the lives of those out of favor with government. But that is not the only risk. There is also the risk of interruption to critical service delivery, intentionally and unintentionally. There are risks of data breaches where governments are not able or willing to properly safeguard data. DPI can make citizens beholden to and dependent on infrastructure that can be disabled during periods of unrest or protest against the state. And I imagine there are a range of other risks from this power transference that we cannot yet fully know.
Theme 4: Mechanisms for strengthening genuine accountability.
Linked to Theme 3, we need to be strengthening meaningful mechanisms for accountability for both states and for the community of DPI promotors. Where donors are funding, supporting, or even legitimating this new endowment of power to states, existing political processes for negotiating state power can be circumvented. States don’t need to make the same case to their citizens about the value and safeguards for these initiatives when they are not subject to ordinary legislative and budgeting processes. The choice of international actors to thus empower states over their citizens in this way is an incredibly risky endeavor that calls for heightened levels of responsibility. In my view, that should entail very careful consideration of where to intervene, high levels of transparency locally and internationally to the nature and levels of donor support given to government initiatives, and an insistence on supporting initiatives only when they adhere to local laws and are subject to appropriate legislative and political processes. What incentives and accountability measures should be in place to ensure responsibility among DPI promotors?
Some international donors—like the World Bank—put in place transparency and safeguarding measures in the past, including conducting and publishing social and environmental risk assessments. Many of these processes grew out of harms generated from World Bank-funded physical infrastructure interventions. But, as Nanjira and I point out, such risk assessments applied in the context of digital infrastructure are mismatched to the nature of real harms emerging in practice. Our old tools are in desperate need of revamping to account for new landscapes of risks that we do not yet fully understand.
State actors also respond to incentives and accountability structures. In Brazil, it was the Central Bank’s incentive to make digital payments ubiquitous and low cost that led to the rise of Pix. The Pix team is accountable to the Central Bank for delivering on this mandate, and public data on usage –and also internal data on fraud and other problems—keeps them accountable for what they are achieving and where they fall short. But this kind of accountability—something that goes beyond having the right technical safeguards in place—is not present everywhere. What should it look like? At a minimum, I would suggest it entails the capture and publishing of regular public data on DPI performance and user value to be used for learning, but also as a tool civil society can use to hold implementing agencies to account, where political conditions allow. Conversations around metrics –what they should capture and how—have already begun (for example Vora and Dolan, 2022), but more work is needed here to practically implement feedback loops and start establishing exemplary cases.
A final thought
Emergent DPI successes, like Pix, are indeed very exciting for what they can accomplish for inclusion in a very short time. At the same time, the DPI space is new territory, where the risks – particularly because of the necessity of state involvement—are enormous and uncharted. After articulating these views at the RFF, a number of attendees thanked me for raising these concerns and commented that such work was “brave.” It should not require courage to beg for caution around this new space. Truly pursuing responsibility in this new domain demands an openness to evidence, debate, and independent inquiry. While we are in new territory, we can also be guided by the wisdom of decades of pursuing responsibility in financial inclusion. That wisdom calls for us to be guided by outcomes, broaden our view of the landscape, and to be guided by deep, rich, and rigorous evidence.
Recommended reading:
· Gelb, Alan H., and Anna Diofasi Metz. Identification Revolution: Can Digital ID Be Harnessed for Development? Washington D.C: Center for Global Development, 2018. https://www.cgdev.org/publication/identification-revolution-can-digital-id-be-harnessed-development
· Solon, Olivia. “How Surveillance Gets Weaponized Against Critics in Uganda.” Bloomberg.Com, June 4, 2024. https://www.bloomberg.com/news/features/2024-06-04/uganda-yoweri-museveni-s-critics-targeted-via-biometric-id-system.
· Vekatesan, Jayshree, Edoardo Totolo, and Loretta Michaels. “Responsible DPI for Improving Outcomes Beyond Inclusion.” Washington, D.C.: Center for Financial Inclusion, July 2024. https://www.centerforfinancialinclusion.org/wp-content/uploads/2024/07/Responsible-DPI-for-Improving-Outcomes-Beyond-Inclusion_jul1.pdf
· Zollmann, Julie, Nanjira Sambuli, Catherine Wanjala. “Citizen Experiences with DPI: Kenya’s Digital ID Transition.” Washington, D.C.: Center for Financial Inclusion, August 2024. https://www.centerforfinancialinclusion.org/citizen-experiences-with-dpi-kenyas-digital-id-transition/