Public Digital Infrastructure or Digital Public Goods? What’s the difference?

Note: These are remarks made at a discussion organized by the AAPTI Institute on Digital Public Goods discussion on 21st Jan, 2021

I spent around 15 years in technology journalism, and another 15 years in Public Policy related work. The policy work includes working in the open source software community, and with iSPIRT to help some of India’s largest banks implement India Stack components like eKYC, UPI and others. So my perspective is that of someone who has worked to explain complex policy matters in simple language. We are at a juncture where we need to take the complex ideas embedded in DPGs and make them accessible to policy makers and the common man and woman.

I am a geek at heart, so the first time I heard Pramod Varma talk about iSPIRT building India Stack in 2015 as an infrastructure for cashless, presenceless and paperless transactions, my first thought was, “Oh, yes. That makes complete sense, for so many reasons.”

The first is that, in the digital realm, governments around the world had completely abdicated their role of building infrastructure. In a world that is becoming increasingly digital, unless governments have a Public Digital Infrastructure (PDI) strategy, their ability to govern will decay and fade. To put it bluntly, an analog government will not be able to govern a digital society.

Secondly, if governments want to achieve public policy goals like financial inclusion, health inclusion etc, their ability to do so by depending purely on the private sector is limited. It is foolish to expect the algorithms written by the private sector to maximise social welfare over profits. Without digital infrastructure, the ability of governments to set the rules of the playground and balance social welfare, innovation and the growth of private enterprises will always be limited to issuing policy diktats, and changes in law. The problem with this approach is that policies and law will always lag in a rapidly changing technology landscape. We are now in an era where software code and the legal code have to work in tandem with each other to achieve policy goals, and this is where PDIs come in.

Building horizontal services like identity, payments and data as PDIs brings the government back into the infrastructure game in the digital era. It also brings with it a host of concerns around concentration of data in the hands of the government, which is a serious concern especially with authoritarian regimes. This is because, unlike the private sector’s digital infrastructure, PDIs combined with the coercive power of the state (army, police, tax authorities and others) could result in a nightmarish human rights scenario. PDIs should therefore be carefully designed and deployed with checks and balances to counter state overreach.

There is also a compelling case for why the private sector needs the government to invest in digital infrastructure. In the case of identity systems, the private sector clearly has no incentive in building a non-rivalrous, non-excludable identity system. The result is that, in some countries, the primary form of ID is a Bank ID, which is issued to those who are profitable customers for the bank. Everyone else is excluded, which drives up the cost of lending, and financial inclusion for those excluded.

The private sector, especially the Big Tech digital platforms, are two decades ahead of governments in terms of building digital infrastructure. Governments therefore have a huge amount of catching up to do, and it starts with explaining the benefits of building this infrastructure to policy makers. Economists might debate the meaning of the word “public” or the term “public goods,” but if you explain the benefits to smart politicians, they will instantly get it. For them, there is no ambiguity around the word, “public” and the utility of “public goods” in achieving policy goals like financial inclusion.

I have thought about these two terms, “Digital Public Goods” and “Digital Public Infrastructure.” As someone who has spent two decades in the open source community, the term “Digital Public Goods” has an obvious and immediate appeal to me. To me, the term “Digital Public Infrastructure” seems to be too techno-centric because it puts “digital” first. I prefer the term “Public Digital Infrastructure” because it puts the public first.

This Public Digital Infrastructure is ideally built on top of Digital Public Goods, though it could also be built on top of proprietary software. Indeed, some of the largest e-gov projects in India like Aadhaar, UPI, GST etc have been built on top of open source software, but the application software that powers these projects is proprietary, closed source software. Also, services like UPI payments that are built on top of OSS cannot be non-excludable because of regulatory and other reasons and cannot be classified as DPG.

Thus PDI is a governance objective, while DPG is the means to an end. The two terms are closely interrelated but distinct from each other.