During the last six months, I have focused much of my time on governance in early-stage companies. Together with my colleagues in the new Agile Governance Platform, we have made the case why governance at the early stage needs to be different from that of large, listed companies. For example, in a recent Integral article, I argue that governance for firms at this stage of development should be redeveloped from the ground up, rather than imposed top-down by stripping down the bulging Governance Codes applicable to later stage entities. In the course of considering these issues, I have been reminded that digital public infrastructures (DPIs) begin their lives as startups too. DPIs may look a bit different from private tech startups: they may be less funding-constrained, and they may not have the pressures of commercial venture capital funding behind them. However, they too must pass through an early stage of experimentation before they can reach national scale and organizational maturity. In that early stage, they too have to wrestle with questions like how to make effective decisions in the face of complexity; and when to prioritize which policies and procedures to adopt. To assume that DPIs are ‘born adult’ is to miss out on addressing the complex, contested issues of early-stage governance which characterize the growth path of DPIs as they seek to achieve society-wide scale and influence. 

So what might governance by design look like for new DPIs? And how does the answer differ, if at all, from early stage companies in general? To get to some answers, let’s first look through the lens of two recent publications which address the governance of DPIs more broadly.

Governance by deliberation

In 2021, a team of Harvard Kennedy School researchers led by Jeff Behrens reported on ethical considerations for the design and deployment of foundational DPIs in developing economies (although their analysis and findings seem hardly restricted to developing economies alone). 

Their main proposal is that governance of DPI should be ‘deliberative by design’--that is, the mechanisms for structured consultation with all stakeholders should be built in from the design phase onwards. They argue: “Because digital public infrastructure is both experimental and fundamental to the lives of residents, deliberative governance should be integrated as deeply as is possible at all stages of the design, deployment, and stewardship of DPGs. Deliberation should be a permanent, rather than temporary, feature of the design and deployment of all digital public infrastructure.” 



Although the HKS researchers do not state this explicitly, the notion of deliberative governance can be juxtaposed with the seemingly hasty governance of the private startup world which ‘moves fast and breaks things.’ This is a key difference between DPIs and private startups: the latter are generally not required to consider broader public stakeholder needs; and most don’t in the focused rush to achieve product-market fit. 

To be clear, speed alone is never a principle of good governance at any stage. But in my opinion agility in governance is also an important principle for DPIs in their early stages–at least as ‘agile’ has come to be understood in the world of software development. There, agile refers to the ability to break up a complex project into smaller tasks which can be developed, tested, and reassessed if need be, without impairing the overall outcome. Agile development was a response to long complex planning processes for software projects which often resulted in failure, notwithstanding all the consideration–either because the market had moved on, rendering the project obsolete or simply because the best planning struggles in the face of complexity. Similarly, deliberation alone will not actually deliver complex digital infrastructure–on its own, deliberation may risk bogging down in a quagmire of contestation and complexity, creating stranded digital assets. Deliberation needs to be linked with clear and effective decision-making protocols and leadership in cultures which consider risk. DPIs need 

Lessons from governance of retail payments systems

To consider what effective DPI provision might look like, let us turn our attention to one of the most developed classes of DPI today–retail payment systems. From small beginnings, many have advanced well beyond startup stages to become indispensable infrastructure of the modern economy. In recent years, there have also been a number of startup payment systems. Ownership models for retail payment systems vary from wholly private (think Visa or Mastercard) to publicly owned and operated, with public-private options in between. There is no one size fits all for governance models, according to the team of World Bank Group authors who considered these issues in a recent report. 

These authors don’t explicitly mention ‘governance by deliberation’ although some of their recommendations–about acknowledging, defining, and addressing public interests and stakeholder involvement is key, for example–align with the approach which HKS researchers propose. But what else can we learn from how payment systems are governed as an active, even booming, class of DPI? 

There are at least two pointers in this report. First, the authors point to the distinction between external governance, that is, regulatory oversight, and internal governance–how internal bodies like boards make decisions on behalf of participants or members. Payment systems usually have both in some form. External governance can be helpful in creating external accountability where needed while allowing internal governance to wrestle with complex issues on a more hands-on basis. 

Second, the World Bank authors stress the need to allow for evolution in governance. Evolution can be linked to good practices like undertaking regular board effectiveness reviews. While the HKS researchers would likely not dispute this, the World Bank authors are more explicit about the need for these infrastructures to recognize and respond to changing market conditions: even if some payment systems function as domestic monopoly infrastructures, technology innovations such as cryptocurrencies or central bank digital currencies may threaten their existence. Recognizing these threats may help to keep DPIs agile. 



Proportionality in governance

While proportionality has already become a widely accepted principle undergirding financial regulation, the World Bank authors include proportionality among areas for further research in governance of payment systems. According to Fernando Restoy of BIS, regulatory proportionality means that the regulatory approach should be tailored to a firm’s size, systemic importance, complexity, and risk profile. How to apply this principle to DPI? By definition, DPI has from the outset the potential to be systemically important; indeed that potential scale motivates the argument for governance by deliberation to take into account the possible effects on stakeholder groups. However, to paraphrase the title of the famous book by philosopher Alaisdair McIntyre, ‘Whose proportionality? Which deliberations?’

Proportionality involves balancing tradeoffs. Someone, or more specifically, some group of people needs to make those calls. Deliberation alone does not resolve tradeoffs or conflicts, though it may highlight them. Deliberation needs to be proportional too since it is not cost-free in time or resources. To apply this, I would argue that there is a pressing need to come up with playbooks for new DPIs written after considering cases with too much or too little deliberation. As one example of too little deliberation, I described in an earlier Article the case of the UK’s Open Banking Implementation Entity as a type of new DPI which arguably moved too fast to achieve remarkable outcomes but with too little deliberation about its internal structures, resulting in an external investigation and changes in governance announced in 2021. But my hunch is that too much deliberation may be as problematic for advancing societal goals, even though the consequences may be different. 

Alongside deliberation, proportionality should therefore be a foundational principle of governance by design for DPIs. Rather than simply accepting that everything to do with a new DPI may be systemically important, the application of proportionality should enable internal and external governance structures of DPIs to parse a DPI project into dimensions requiring different levels of intensity of deliberation. This approach can allow for greater agility in early-stage DPIs which will reduce the risk of a future landscape cluttered with digital white elephants.