What do we know about the state of DPI in the world? Preliminary insights from the DPI Map
By David Eaves and Krisstina Rao
Last month (June of 2024), we launched the DPI Map — the first comprehensive view of the state of digital identity, digital payments and data exchanges across the world.
This work started — in part — because there was little understanding of how many countries had adopted Digital Public Infrastructure (DPI) around the world. Given the growing interest by policymakers, multilateral funders, public officials and public interests advocates, we hypothesized that a “map” of DPI could help create a baseline understanding of what is happening, and more importantly, help serve as a foundation for both future research and community on this topic.
And while early successes suggest that that hypothesis has been validated we’ve been so busy with outreach that we’ve not had much chance to share any emerging insights from the initial DPI map we’ve published. This blog post seeks to change that!
Caveats
Before we begin sharing the exciting insights we’ve gleaned so far, there are two important pieces of information to share.
First, to create an objective list of digital public infrastructures (DPI) around the world, we had to develop a framework that defines the attributes of digital public infrastructure and a set of indicators to assess if these attributes exist. These insights are thus both made possible by and limited by, our methodology (which you can read more about here).
Second, the DPI map is, to date, only partially complete. We’ve only focused on 132 countries — predominantly in the global south — to date with the hope of completing the rest of the world before the year is out. There are also many indicators for which we could find no evidence either way, so the data is not perfect. So these insights are preliminary. But still exciting!
Insight #1 There is a great deal more DPI than anyone thought
Our current assessment is that there are at least 46 identity systems, 38 payment systems and 46 data exchanges on the planet that could be considered DPI. Our conversations with stakeholders before this project’s launch suggested people believed there were a handful of DPIs in the world. Our work indicates that even with a conservative definition of DPI, there is a lot more than the 3–4 DPIs from Brazil, India and Estonia that are typically (and rightfully) showcased as success stories. This has several implications. One is that there are more opportunities to study successful (and unsuccessful) DPIs than previously thought. And that there are likely many more approaches to governance, technical implementation and use cases that researchers and policymakers could be seeking to understand.
Insight #2: DPI has been around for some time and emerged in many places
Not only is there more DPI than widely believed. We found that the phenomenon has a longer history than was previously thought, with some deployments going back as far as 20+ years. Most notably, the first iteration of Estonia’s X-road system was launched in 2001. X-road has also been adopted by several countries, an example of governments learning from each other and adopting not only a technology but also processes.
But not all countries that engage in data sharing use X-Road. The same is true for digital IDs and payments. There has been significant parallel evolution– with many governments developing digital public infrastructure concurrently, without calling it DPI or sharing the output or means of production with others.
The narrative around DPIis sometimes focuses on a single country’s success. The reality is that this is a domain of parallel evolution across many governments. Sometimes these governments borrowed explicitly (sharing code) or implicitly (adopting approaches) from each other, but there appear also to have been completely independent projects that emerged in isolation. This suggests a general acknowledgement of the important role of infrastructures such as identity, payments, and data exchange is common to public service delivery– by national governments.
Insight #3: DPI is a Global South/Emerging Market Phenomena
While we need to wait until our map is fully complete, there is a strong indication that DPI is a Global South phenomenon. This is not an example where a technology and process emerged in an economic core and then migrated to the periphery. Quite the opposite. It is the Global South that is taking a leadership role in building, deploying and demanding DPI. Interviews from other research we have done suggest three core reasons:
- A development imperative: to accelerate improving government services and formalisation of the economy
- A sovereignty imperative: limit or exclude foreign actors ability to control core functions for operating a digital era society/economy
- A regulatory imperative: limit or make public the rents that are made possible by common infrastructure
However, for all three of these a DPI approach has the additional benefit of commoditizing, and thus making cheaper, core components needed for operating a modern state. This has an additional benefit for resource-constrained governments.
For example, In 2022, the global government IT spend was $693 billion, of which most ($612 billion) was accounted for by governments of North America, Western Europe, China, mature Asia and Japan. That left the rest of the governments (almost all Global South) with a minuscule $81 billion budget– or a per capita spend of $14 per person (as compared to a spend of $252 in mature economies).
These budgets are highly disparate as it is, but become even more curious because they do not reflect the quality of digital services in these countries. Not only did the DPI Map find evidence of way more DPI deployments in the Global South than we previously thought, but it also found emerging data on the impact (in terms of % of the population reached, in the case of digital ID enrolment, or number of participating financial institutions, in the case of payment systems) of these governments in terms of scale and effectiveness. One anecdotal example came during the COVID-19 pandemic when governments like India and Togo were able to leverage DPI to deliver direct benefits more rapidly and effectively than many mature economy governments. It begs the question: how can governments spend on building digital capabilities in a way that offers the most public value?
Insight #4: Evidence points to emerging proto-DPIs.
One exciting set of data to emerge from this mapping project is the capacity to identify a non-trivial number of “proto-DPIs.” The mapping exercise laid bare the fact that to define DPI, we would need greater consensus on the definitions of digital identity, digital payments and data exchanges, whose scope remains blurry. Studying DPI deployments, however, did help us draw patterns between how the ecosystem would like DPI to be designed (we called these attributes, see more here) and how the aspirational attributes show evidence of being realised in real-world deployments (called indicators). The evidence for proto-DPIs emerges in different ways, explored below.
Insight #5: The lack of a standard definition around DPI remains a challenge
Governments have variable definitions of digital identity, digital payments and data exchange. This means that while a country may claim to have a digital identity, it may not be identical in its technical and governance features to another country’s digital identity. For context, of the 132 countries mapped, 75 claimed to have digital IDs. Most (94% Data on digital authentication was only available for 50 of the 132 digital identities studied) of these identities facilitated digital authentication for individuals to access follow-on services. However, fewer were supported by effective governance, with 51 of the 75 countries having data protection acts and 26 having a defined legal or policy framework specific to the identity. In some ways, consensus around a baseline of features of DPI will lay rest to the definitional issues that DPI, and digital identity as DPI, face. But this issue is parallelly fuelled by a separate insight.
Insight #6: A Diversity of Approaches
Building on insight #2, the parallel evolution of many DPIs also means that governments have adopted different ways to build DPI and at different paces. This is evident in the large proportion (42%) of digital payment systems that were built either completely privately or in association with private entities. Most of these systems have a certain degree of public-interest governance, either through regulator-mandated oversight or self-governing practices that make transparent how they govern for privacy, protection and safety of users. The map on data exchanges also highlights the emergence of ‘proto-DPIs’ among data exchanges. Data exchanges are a global story with projects being proposed or implemented on every continent we researched. What we call data exchanges today in the DPI ecosystem have somehow become synonymous with X-Road. But we found that governments not only use different base architectures (X-Road, but also WSO, UXP) but also variable models (API gateways, enterprise service buses or consent-manager-based data sharing) to exchange data between governments and individuals– all with variable measures of technical functionalities and governance.
Finally, it’s worth noting that our understanding of the state of DPI is subject to the state of evidence on DPI out there. The DPI Map’s methodology is limited to publicly reported sources of information, which needs to further be accessible to our team. Specifically, in several places, the indicator we try to measure is weak because the supportive evidence is the best version of measurable evidence out there. It could be that the DPI in question has way more features than the Map reports, but weak reporting constrains the Map’s ability to study these features.
What does this mean for us?
We know more about the state of DPI now, especially in the Global South. It is of value to not only investigate these deployments as stories of success and document their practices but also learn from their context of what works. Initiatives like the 50-in-5 champions country practices from the Global South already, and it is imperative for us to document ‘what works’ (and what doesn’t) in DPI from these early champions.
We also know way more about the governance of DPI. We know that the governance of the DPI, both through its design and its use, matters as much as, if not more than its technical architecture. The Map has also systematically investigated both– achievable and aspirational — governance practices. When combined with multi-stakeholder initiatives like the UN DPI Safeguards initiative, this will hopefully create a playbook for governments interested in building DPI that is safe, inclusive and publicly accountable.
While the DPI Map and the Measuring DPI Framework do not compare one deployment with another, we recognise the need for better reporting on certain metrics that might promote countries to better learn from each others’ documented practices. The DPI Map does not intend to serve as an index for global DPI but aims to create a baseline of evidence that generates consensus on the attributes and indicators of DPI.
This article is part of a series of blogs that document insights from the DPI Map, a project of the UCL Institute for Innovation and Public Purpose, which is supported by Co-Develop. To stay in touch with insights like these, subscribe to the newsletter.
