NextBillion - An Initiative of the William Davidson Institute at the University of Michigan

In recent months, there has been an outpouring of media coverage and analysis on the topic of digital public infrastructure (DPI), including articles, ideas, podcasts and convenings.

This deluge of attention is hardly surprising. The 77th session of the United Nations General Assembly, which formally came to a close on September 5, emphasized the need for “country-led digital cooperation to build safe, trusted, and inclusive DPI” under its theme of “transformative solutions to interlocking challenges.” And for the past year, India, as the current president of the G20, has placed DPI at the center of the group’s agenda. The Indian government has also leveraged digital platforms and technologies to promote economic progress, innovation and financial inclusion, showcasing the vast potential of DPI for emerging markets.

In an effort to harness this momentum, key stakeholders in this emerging field — including some of this article’s authors — have made more efforts to define what DPI actually means. In a paper written last January, David was able to identify just one substantive definition of DPI at that time — that offered by the Digital Public Goods Alliance’s GovStack working group. Six months later, there are now several more, including this definition from Co-Develop, the independent specialist agency founded in 2021 to promote DPI, and most notably, the recently released draft definition from the G20 working group on DPI. That definition defines DPI as “a set of shared digital systems which is interoperable, inclusive, secure and trusted, built on open standards and specifications governed by enabling rules that respect fundamental rights to deliver and provide access to public and/or private services at societal scale to drive innovation, trust and competition.”

This heightened definitional discourse is encouraging, especially considering the global prominence of those involved. Yet because of the difficulty of reaching a global consensus on a comprehensive definition, some, like Carnegie India, have argued for abandoning this quest in favor of setting out common principles guiding DPI. However, a principles-based approach does not avoid the need for sufficient clarity entirely: It still requires clarity over what the principles cover — and what they don’t.

Even in the absence of full definitional consensus, it seems appropriate to pause and ask: Is DPI, as it is now being framed, likely to support a positive digital development agenda? This question looks beyond the sterile quest for definitional perfection to consider the likely outcomes of the emerging understanding of DPI. We believe DPI can be a helpful lens to provide focus for digital development. However, to realize its potential, stakeholders must look beyond narrow conceptions based on government-centered use cases and existing digital systems to envision next-generation applications. But to start to consider those, we need first to assess the emerging understanding of DPI.

Converging on a Shared Understanding of DPI

The two DPI definitions discussed above, from the G20 and Co-Develop, have several features in common: They say that DPI must be built on shared “building blocks” of some sort; it must be interoperable, secure and trusted; and it must ultimately focus on delivering digital services at scale.

Neither the G20 nor Co-Develop definitions require that a DPI be owned or operated by a state: Rather, DPIs must benefit the broader society — i.e., they must have a public purpose regardless of whether they have public ownership or management. This is certainly the case in digital payment systems, which are often owned and operated by utilities established by private sector players, though they are usually subject to regulatory oversight by the state. This focus on public purpose leaves a wider space for DPI forms to evolve, allowing for private sector and civil society innovation. However, both definitions are also clear that the DPI space is not a laissez faire free-for-all: It requires “enabling rules … which respect fundamental rights” (as the G20 put it) and “public institutions which guarantee” oversight of the DPIs in question (as in Co-Develop’s definition).

So far, so good, then: Not only is there evidence of convergence, but both the G20 and Co-Develop definitions appear broad enough to allow breathing room for further evolution of the approach. But are they too broad, allowing for too many things to fit under the DPI umbrella? One specific concern is that their notion of infrastructure as building blocks remains vague: “a set of digital systems” (G20) or “digital capabilities” (Co-Develop). This does not rule out software packages or even codes or protocols, introducing the risk that DPI comes to be seen merely as a specialized subset of govtech solutions. This association would limit DPI to providing solutions for how governments interface with citizens, rather than promoting a wider, whole-of-society approach to digital transformation. As envisioned by Pramod Varma, the chief architect of India’s Aadhaar digital identity program and other elements of its much-lauded “digital stack,” DPI allows and even encourages other parts of society — including the private sector — to build use cases on top of it.

Consider an example to highlight the value of this broader definition of DPI: Many Ministries of Finance are implementing or upgrading their systems for tax collection. These govtech solutions are intended to benefit both government and taxpayers, but they are typically not designed to also benefit third parties. However, if these new taxation systems also introduced features that extended beyond government, such as allowing for third party verification of taxable income, this could benefit borrowers and the lenders to which they authorize access to their verified record of taxable income. This vision for DPI sees it not just as a digital service, but rather as a foundation on which other services may also be built.

Systems for digital payments, digital identity and data exchange are increasingly accepted as three self-evident categories of foundational DPI. However, we see little value in simply renaming existing systems as DPI, unless they are also designed with this essential characteristic in mind: to interoperate with other systems rather than to provide standalone solutions. Varma has dubbed this approach as “DPI thinking,” in which DPI refers to the “system of systems,” rather than the individual pieces. Under this framing, these individual building blocks would not be labelled as DPIs, but rather would be assessed as potentially “DPI-ready” — i.e., ready to serve as building blocks in a next-generation approach to digital public infrastructure.

Envisioning the Next Generation of DPI

While much activity is currently focused on building up foundational infrastructure in areas like payments or digital identity, the most promising emerging examples of next-generation DPI go beyond these often-siloed spaces to envision other changes that may be genuinely transformational when different digital systems start to interoperate.

One interesting new example comes from the Bank for International Settlements (BIS), which supports central banks’ efforts to maintain monetary and financial stability, and hosts several international standard-setting bodies in the financial sector. In February 2023, its General Manager Agustin Carstens set out a vision of a global financial infrastructure in which central bank digital currencies interoperate with bank-issued digital currencies and registries of financial or even nonfinancial claims through a unified public ledger. This vision was further developed in a chapter of the June 2023 BIS annual report entitled “Blueprint for the future monetary system.” While the BIS does not attach the term “DPI” to this blueprint, we nonetheless believe that it exemplifies a next-generation DPI approach. This is because it combines payment systems for the transfer of digital money — both central bank-issued and privately-issued digital currencies — with the use of distributed digital ledgers of financial or real claims to effect simultaneous transfer of these assets (with payment using smart contracts).

The blueprint is at a conceptual stage and, as the BIS acknowledges, it is ambitious. To achieve it would require solving the governance challenge of aligning public-private collaboration across its various components, since digital currencies would still be privately issued, and blockchain-based asset registers may also be privately operated. Smart contracts can help execute the payment and transfer instructions simultaneously, but the system would still require rules and standards to be developed and enforced. However, the existence of these challenges doesn’t mean the proposal is unworkable: The quest for public-private governance solutions is not new, even in digital systems. Payment systems have long wrestled with different configurations of public oversight and public or private ownership and operation, demonstrating that several stable public-private partnership configurations can exist in DPI.

The BIS blueprint provides a foretaste of the true promise of the next generation of DPI to “drive innovation, trust and competition,” as the G20 definition sets as the end goal. Although increasing numbers of countries are now exploring different elements of this blueprint, such as central bank digital currencies, none have yet configured these pieces to deliver the next-generation digital infrastructure the BIS envisions. But the fact that a body with the stature of the BIS is actively promoting this vision highlights the vast potential of DPI approaches. In addition to the BIS, DPI thinking is starting to permeate other areas, like climate change, in which public and private collaboration can unlock next-generation solutions.

On a country level, the emerging example of how India has deployed a DPI approach also hints at the wide-ranging potential that can be realized when the different layers of DPI start to interoperate as intended. India’s progress in creating DPIs around identity, data and payments on a national scale has also been based on public-private partnerships, enabling the design and operation of the various layers of this digital infrastructure, and supported by the country’s deep reservoir of highly skilled tech talent. Many Global South countries will also need to adopt a flexible public-private approach, since the complexity of next-generation DPI solutions will mean that siloed government agencies will struggle to build the capacity to design and operate them alone — especially in the face of high demand for scarce digital skills.

A flexible approach means that governments may decide they can and should operate some solutions using open-source software, while licensing others from the private sector. They may adopt shared DPI-as-a-service models from countries like India and Estonia, or explore shared infrastructure models that sit atop existing trading bloc agreements between participating states. The shared infrastructure approach could proceed along the lines of the global shared services centers widely used by inter-governmental organizations and the private sector alike, with governments overseeing them as utility-type solutions. However, there will be an acute public interest in the stability, coverage and efficiency of these utilities — and most nations today lack the regulatory mindset and capacity to effectively oversee emerging digital utilities of this nature, even if they have precedents from the physical infrastructure sector. Addressing this gap will require the DPI approach to be inculcated more broadly in the public and private sectors, moving beyond the narrow regulatory domains, like payments, ID or data, where most DPI solutions currently focus.

Next-generation DPI will also require an ongoing openness to innovation: There are entire classes of DPIs yet to be conceptualized and designed, addressing the multitude of problems faced at the local, national and international levels. The evolving DPI definitions we’ve discussed in this article don’t (and shouldn’t) preclude this shift — in fact, they point towards the need for it. These definitions will have fulfilled their purpose if they lay a foundation that encourages experimentation with new and emerging forms of next-generation DPI, rather than simply re-labeling existing approaches. If both the public and private sector can work together to build this foundation, they — and the communities they serve — will reap the benefits for decades to come.

David Porteous is the founder of Integral: Governance Solutions; Priya Vora is the CEO of Digital Impact Alliance; Ravi Shankar Chaturvedi is the Managing Director of Digital Planet; Peter Rabley is the co-founder and Managing Partner of PLACE.

Photo courtesy of Marc-Olivier Jodoin.