How To Design Governance For Seven Billion Connected People

In 1648, the Peace of Westphalia ended the Thirty Years' War and, more consequentially, articulated a principle that would organize international relations for the next four centuries: state sovereignty. Each state was supreme within its own territory; no external authority had the right to intervene in its internal affairs. This principle solved the immediate problem — the religious wars that had devastated central Europe — by removing religious questions from the domain of international legitimacy. What happened inside a state's borders was not subject to external judgment.

The Westphalian order worked, in the sense that it provided a framework stable enough to sustain international relations for centuries. It also had a built-in design flaw that is now actively producing civilizational risk: it has no mechanism for addressing problems that cross borders. If the most important governance challenges of a given era are all internal to states — maintaining order, managing economies, preventing civil war — the Westphalian order is adequate. If the most important governance challenges are transnational — climate change, pandemic disease, financial contagion, AI development — the Westphalian order is structurally incompetent.

We are in the second situation. Almost every problem on the list of civilization-threatening challenges is transnational. And we are trying to address them with an international governance architecture designed on Westphalian principles for a world of fifty-some sovereign states, now applied to a world of nearly two hundred sovereign states, eight billion people, and communication infrastructure that makes the globe effectively one information space.

The Legitimacy Problem

Any serious discussion of global governance runs immediately into the legitimacy problem: who decides, on whose behalf, by what right?

National governments derive their legitimacy from some combination of consent (democratic elections, however imperfect), tradition (monarchies and established states), and effectiveness (states that deliver security and services). International institutions derive their legitimacy from... the states that created them. They are at least two steps removed from any direct democratic accountability. No one votes for the WTO's Dispute Settlement Body or the IMF's Executive Board. Their decisions have enormous consequences for populations that have no formal channel of influence over them.

This legitimacy deficit is not just a procedural problem. It creates substantive governance failures. Institutions that lack democratic accountability tend to make decisions that serve the interests of the powerful actors that created them — the wealthy states, the financial sector, the large corporations that have resources to participate in technical standard-setting processes. The outcomes of international governance reflect this systematically: intellectual property rules favor pharmaceutical companies over global health needs; financial regulations favor financial centers over developing economies; trade rules favor export-oriented industries over labor and environmental standards.

Solving the legitimacy problem does not require creating a world parliament, though that is one proposal that periodically gets serious academic attention. It requires creating mechanisms for meaningful participation by affected populations in the decisions that affect them. This is not impossible; it is difficult. The European Parliament represents an imperfect but real attempt to create democratic accountability in a supranational institution. The Aarhus Convention gives citizens the right to participate in environmental decision-making in their countries. The internet governance processes at ICANN and the Internet Governance Forum include multi-stakeholder models that give civil society organizations some formal role, however constrained.

None of these are sufficient. All of them represent genuine partial progress toward the principle that people should have some say in the decisions that shape their lives, even when those decisions are made at scales larger than the nation-state.

The Polycentric Alternative

The political scientist Elinor Ostrom, who won the Nobel Prize in Economics in 2009 for her work on the governance of common-pool resources, developed a concept that is directly applicable to global governance: polycentric governance.

Ostrom's research showed that the management of common resources — fisheries, water systems, forests, grazing lands — was not well handled by either centralized state control or pure market mechanisms. Both the "government must manage this" and the "privatize it" approaches missed the actual solution, which was community-based governance systems with specific design characteristics that enabled effective self-management without either the inefficiency of central planning or the tragedy of unregulated commons.

The design characteristics that Ostrom identified — clear boundaries, rules matched to local conditions, collective-choice arrangements, effective monitoring, graduated sanctions, conflict-resolution mechanisms, and recognition by external authorities — were not specific to any particular resource or community. They were general principles of effective commons governance that appeared wherever effective commons governance appeared.

Polycentric governance means multiple, overlapping governance authorities rather than a single hierarchical authority. Cities and states make some decisions; national governments make others; regional bodies (the EU, ASEAN, the African Union) make others; global institutions make still others. No single level is supreme; each level handles the problems it is best positioned to handle; there are mechanisms for coordination between levels when issues cross them.

This is actually how governance already works in most developed countries, most of the time. Federal states like the United States, Germany, and India have constitutional arrangements that distribute authority across multiple levels. The problem is extending this logic to the global level, where the equivalent of a constitutional arrangement — something that specifies which global institutions have authority over what, under what conditions, with what accountability mechanisms — does not exist.

Design Principles for Global Governance

Drawing on Ostrom's work, on the comparative study of international institutions, and on the emerging scholarship on complex adaptive systems, it is possible to identify design principles for global governance that would be more adequate than what we currently have.

Subsidiarity. Decisions should be made at the lowest level that can make them effectively. Climate change cannot be managed at the local level; atmospheric commons require global coordination. But most governance decisions can and should be made much closer to the people they affect. The principle of subsidiarity — enshrined in the European Union's founding treaties — means that the global level should only address what the national and subnational levels genuinely cannot.

In practice, this means that global institutions should set minimum standards and coordinate on cross-border issues while leaving implementation to national and local actors who understand local conditions and are accountable to local populations. The Montreal Protocol worked partly because it set global targets while leaving the specific mechanisms for achieving those targets to national governments that understood their own industrial economies.

Differentiated obligations. Countries and communities have different histories, different capacities, and different contributions to the problems being addressed. Governance that assigns identical obligations to all parties — regardless of their capacity to fulfill those obligations or their historical responsibility for creating the problem — will not secure the cooperation of the parties most disadvantaged by uniform treatment.

The Montreal Protocol's differential treatment — industrialized countries faced immediate phase-out schedules while developing countries had a ten-year delay — was not a concession to unfairness but a recognition that legitimacy and effectiveness required addressing the legitimate concerns of developing countries. The Paris Agreement's "common but differentiated responsibilities" is an attempt at the same logic applied to climate change. The attempt is imperfect but the principle is correct.

Adaptive mechanisms. Any governance architecture for a complex system needs built-in mechanisms for learning and adaptation. The problems of the 21st century are not fully known; the technologies that will address or exacerbate them are not fully developed; the consequences of governance choices will not be fully visible for decades. Governance that locks in specific rules without mechanisms for revision will be wrong in predictable ways.

The Montreal Protocol's design included automatic review processes and the capacity to tighten restrictions as new scientific evidence emerged. This adaptability is part of why it succeeded — it was not hostage to the specific scientific understanding of 1987. The Framework Convention on Climate Change is structured similarly: a framework convention that sets principles and processes, with specific commitments to be negotiated and updated in subsequent protocols and agreements.

Meaningful accountability. The gap between governance decisions and the people affected by them corrodes legitimacy over time. Closing this gap at the global level does not require creating global democratic institutions immediately, though the long-term trajectory should move in that direction. In the interim, it requires transparency (decisions and the processes by which they were made should be publicly accessible), participation (affected communities should have formal channels for input, not just informal lobbying access), and review (commitments should be subject to independent assessment and public reporting).

The Aarhus Convention on environmental governance, which guarantees citizens the right to information, participation, and access to justice on environmental decisions, is a model for this at the regional level. The UN's Sustainable Development Goals' review processes — including voluntary national reviews and civil society participation — are an imperfect but real attempt at this at the global level.

The Technology Question

The governance of global digital infrastructure is one of the most urgent and least solved problems in global governance. The internet was built largely in the United States by American institutions, with American values about free speech, commercial freedom, and limited regulation embedded in its architecture and governance structures. As the internet became global, these embedded values became sources of conflict.

The domain name system is governed by ICANN, a nominally private organization operating under a contract with the US Department of Commerce that was only fully severed in 2016. The core internet protocols are managed by the Internet Engineering Task Force, a nominally open process that operates largely through consensus among technically sophisticated participants — a group that is not representative of the global internet user population. Social media platforms are primarily US companies operating under US law but serving global users under terms of service that reflect US legal and cultural assumptions.

The responses to this arrangement include the EU's regulatory assertiveness (GDPR, the Digital Markets Act, the Digital Services Act), China's "internet sovereignty" model (domestic internet governance independent of global norms), and various attempts at multi-stakeholder internet governance processes. None of these constitutes a satisfactory answer to the question of how the digital infrastructure that now mediates most global communication should be governed, by whom, and on whose behalf.

The emergence of artificial intelligence as a transformative technology has made this question more urgent. AI systems developed primarily by a small number of US and Chinese companies are being deployed globally with implications that their developers are not fully equipped to assess and that affected populations have no meaningful voice in shaping. The governance gaps in AI development are visible and are being partially addressed by national regulatory efforts (the EU's AI Act, the US executive orders, China's algorithmic regulations) and international coordination efforts (the OECD AI principles, the G7 Hiroshima process). These are early-stage responses to a governance challenge that will take decades to adequately address.

The Practical Path

Global governance adequate to the challenges of the 21st century will not be built in a single moment of institutional creativity. It will accumulate through a series of specific agreements, institutions, and norms — some successful, some failed, some reformed after initial failure — over decades. This is how the current, inadequate international order was built, and it is how a better one will be built.

The practical path involves several things happening simultaneously. Existing international institutions need to be reformed — their voting structures, accountability mechanisms, and mandate scope all require updating for a world different from the one in which they were designed. New institutions and agreements need to be negotiated for problems that current institutions cannot address, including AI governance, pandemic preparedness, and space governance. Regional institutions need to be strengthened as an intermediate level between the national and global. Civil society organizations need to develop the capacity to participate in global governance processes, which requires resources, expertise, and access that most civil society organizations in most countries do not have.

None of this is sufficient. All of it is necessary. The alternative — attempting to address 21st-century problems with 20th-century governance architecture — is not working. The evidence is visible in the gap between the scale of the challenges and the adequacy of the responses. Climate change commitments fall short of what science requires. Pandemic preparedness remains grossly underfunded relative to risk. AI development proceeds without governance frameworks adequate to manage its risks.

The design of governance for a connected civilization of eight billion people is the central political challenge of this century. It is not primarily a technical problem — the design principles are not unknown. It is primarily a political problem: gathering the will to create institutions that require powerful actors to accept constraints, and creating accountability mechanisms that give the least powerful voices in global governance genuine influence. These are not easy things to do. They have not been done adequately before. They need to be done now, with less time than we have spent on easier challenges. Whether they will be is genuinely uncertain.