Federalism from What Goes Wrong: Vincent Ostrom and the Liberalism of Repair

Maintaining social order in the face of conflicting values is one of liberalism’s key challenges. A focus on correcting manifest harms may provide the stability that federalism requires.

Paul Dragos AligicaContributor, Liberalism.org
Federalism from What Goes Wrong: Vincent Ostrom and the Liberalism of Repair
November 18, 2026·15 min readCivicsInstitutional Designpluralism

One of the most profound challenges confronting modern, complex, globally integrated societies is how to sustain cooperation when citizens, communities, and social groups do not possess a common vision of the good, do not inhabit a single epistemic world, and do not experience political problems in the same way. How should social order be defined and maintained in conditions of deep heterogeneity? What, under such conditions, must a heterogeneous people still share? In what measure is it possible to build functional architectures of governance when not much is shared? 

In this context, federalism is often presented as a way of organizing difference in a manner that permits a people to remain heterogeneous while sharing a political order. Federalism’s enduring problem is therefore not simply how to divide powers between national and subnational governments. It must also answer a question: What, under such conditions, must a heterogeneous people still share?

Today the question of the governance architecture and its underlying conditions has acquired new urgency. Demographic change, immigration, generational cleavages, technological transformation, and the fragmentation of the media environment have multiplied the identities, interests, experiences, and interpretive frameworks present within the polity. Diversity is not new, and heterogeneity doesn’t necessarily produce disorder. But shared institutions now operate under visibly and consequentially plural conditions, over a domain of rapidly multiplying and deepening cleavages.

We already know of two familiar answers. Both seem to be inadequate. The first is the idea that citizens must share a comprehensive creed, a common identity, or a substantive conception of the good. That answer may describe some historical, traditional communities, but it is unavailable as a general solution under the social conditions of deep and dynamic pluralism. We need to realistically relax this assumption or expectation. 

The second is the procedural approach: citizens need only share rules, rights, and institutional arrangements. This is closer to the liberal ideal, but it leaves open a difficult problem: Rules are not self-executing. They must be interpreted, trusted, monitored, contested, and repaired by people who may disagree about their ultimate purposes or their specific interpretations.

Confronted with this challenge we turn to Vincent Ostrom’s work, particularly to the ideas developed most fully in The Meaning of American Federalism: Constituting a Self-Governing Society (1991). Through it, one may find three connected moves that may help us approach this problem. First, one may show why institutional design alone cannot master the complexity of self-governing orders. Polycentric arrangements are indispensable precisely because knowledge and authority are dispersed, but they depend on habits and commitments that no blueprint can manufacture. Second, Ostrom introduces an ingenious twist: He redirects attention from the search for optimal forms toward the sources and consequences of institutional failure. Federalism should be understood not as a finished architecture but as a system that has built-in capabilities for detecting, containing, and correcting error. Third, extending Ostrom’s path, we can ask whether shared recognition of collective or public harms may offer a more attainable basis of common ground than agreement on collective, public goods. People may disagree about what is good, or a specific interpretation of the good. Yet they may still converge on what must not be tolerated or on what may be avoided, on what is destroying their capacity to coexist, and what must be repaired.

The argument is apparently modest. But it gives federalist polycentricity a very solid foundational footing. A negative or harm-centered consensus cannot replace a rich civic culture, a shared constitutional imagination, or the moral sources of covenant. It can, however, provide a threshold on which cooperation may begin. It may be the least demanding answer to the question of what a heterogeneous people must still share, and therefore a plausible starting point for renewing federalism’s promise.

Institutional Arrangements and the Covenantal Turn

Ostrom’s federalism is best understood within a broader institutional theory framework. His analysis distinguishes among operational, collective-choice, and constitutional levels of rule ordering. At the operational level, actors engage in concrete activities and respond to the rules immediately governing their circumstances. At the collective-choice level, authorized participants formulate policies, procedures, and organizational rules. At the constitutional level, they establish who may make such choices, under what conditions, and according to which principles of jurisdiction, representation, legitimacy, and accountability. These levels are nested, but they are not sealed off from one another. The institutional arrangements shape the incentives and possibilities of action below and above.

This framework, which captures very well the logic underlying federalist systems, is already a departure from simple institutional engineering. Governance systems are not machines assembled once and then left to operate. They are recursive arrangements, in which rules influence conduct, consequences generate learning and conflict, and experience feeds back into rule revision. 

Polycentricity is how Ostrom conceptualizes the institutionalization of this logic. A polycentric order contains multiple formally independent centers of decisionmaking that operate within a larger system of mutual adjustment. As Ostrom put it in The Meaning of American Federalism, “A polycentric political system is composed of many autonomous units that take account of others through cooperation, competition, conflict, and conflict resolution.” Their value is that they provide possibilities for learning, communication, trust, reciprocity, experimentation, accountability, and error correction. They allow diverse communities to adapt rules to local circumstances while preserving channels through which wider problems can be addressed.

At the same time, multiple jurisdictions may also produce evasion, duplication, jurisdictional rivalry, and failures of collective action. Social cooperation and successful collective action are not mechanically or structurally generated. 

This is where the idea of a covenant becomes important. As Ostrom notes, the word “federal” is connected to foedus, a covenant or binding agreement. In a covenantal understanding, political authority is not simply imposed from above or generated mechanically. It is constituted through relationships of mutual obligation among persons and communities. Daniel Elazar’s account of covenantal federalism, Tocqueville’s discussion of the habits of the heart, and Ostrom’s own reflections on constitutionalism converge on the idea that self-government in federal structures requires more than formal institutional arrangements. It requires a disposition to recognize others as legitimate participants, to accept limits on one’s own authority, and to sustain agreements even when immediate advantage would favor defection.

And thus, in a somewhat paradoxical way, Vincent Ostrom, one of the most rigorous analysts of rules, incentives, and administrative arrangements among the governance theorists of the second half of the twentieth century, points toward a dimension that cannot be reduced to those categories. A rule system requires interpretation and adaptive contextual compliance, what might be called a “civic ecology”: practices of association, trust, mutual recognition, and responsibility that give formal arrangements a living content. Polycentric governance is thus simultaneously an institutional technology and a cultural achievement. Federalism is not a blueprint for a social technology ready to be implemented automatically. The concepts of federalism and covenant reflect the two facets of the same complex social, cultural and institutional artifactual phenomenon. Federalism is a formal, rules-bound structure; covenantalism sustains the polity when the articulated rules don’t yield successful collective action.

This point is especially important in an age of renewed confidence in modeling, simulation, data analytics, and algorithmic administration. In such circumstances, the temptation of presuming that federalism or polycentricity can be socially engineered through technological replication is huge. The idea might take the form of a “smart” governance system in which data platforms, digitalization, and artificial intelligence are used to reproduce the rules and decision procedures of a functioning federal or polycentric order across different regions or communities. In other words, to externalize governance itself to machines. Yet technology can replicate formal arrangements; it cannot automatically reproduce the social conditions that sustain them. We know that the same institutional design may generate very different outcomes when transplanted into settings with different histories, norms, capacities, and patterns of trust. The danger is not technology itself. It is the belief that technical sophistication can substitute for judgment, legitimacy, and civic capacity. But the covenantal layer cannot be engineered in the same sense in which a procedure or administrative system can be engineered. It can only be cultivated, learned, transmitted, and renewed through human practice. 

This leaves us with a difficult question: If the conditions of self-government cannot be designed directly, how should institutional analysis and design orient itself? Ostrom’s answer offers an ingenious twist: He invites us to shift our attention away from the search for the optimal form and functioning, and to focus on problems, anomalies, and how governance arrangements go wrong. In the standard approach, optimal design and performance occupy the foreground, while dysfunctions, anomalies, and failures appear in the background, as deviations from an ideal model. In Ostrom’s alternative perspective, the relationship is reversed: optimality recedes into the background as a regulative aspiration, while the foreground is occupied by the identification, diagnosis, prevention, and correction of institutional failures.

From Optimal Design to the Sources of Failure

In his essay “Where to Begin?”, Ostrom argued that our approach to federalism should focus more on dysfunctions and potential sources of failure than on ideal models or blueprints of government as such. This is a truly consequential reorientation. Instead of asking which institutional arrangement is optimal in the abstract, we ask how arrangements fail under particular conditions, how failure is recognized, who bears its costs, and whether the system contains mechanisms for correction.

The failure perspective is consistent with the Madisonian and public-choice traditions. It begins from the assumption that rulers, administrators, citizens, and organized interests are fallible and responsive to incentives. It does not expect good outcomes to follow automatically from good intentions. And if one takes that approach seriously, as Ostrom surely did, that changes the meaning of institutional architecture.

The error-correcting element of the system becomes pivotal: Where does information about failure arise? Can it travel across jurisdictional boundaries? Are those who experience the consequences able to contest decisions? Can unsuccessful policies be revised without destabilizing the entire order? Is there an exit option, an alternative provider, or a competing center capable of demonstrating that another path is possible? This is a fresh way of thinking about federalism.

The contrast between monocentric and polycentric structures is now seen in a new light. In a monocentric system, there are fewer independent sites at which error can be exposed or resisted. In a polycentric system, however, many failures may remain local, at least initially. Alternative centers can experiment with different rules; citizens may compare outcomes; specialized knowledge can be mobilized close to the problem; and competing authorities may expose one another’s mistakes. Redundancy can be a source of resilience. Polycentricity does not eliminate failure, but it changes its scale, visibility, and correctability. It asks less of our knowledge in advance. But it requires attention to warning signs, distributional effects, institutional bottlenecks, and recurrent breakdowns, directing reform toward monitoring, graduated responses, and learning.

At this point we get an even better understanding of why formal rules alone cannot determine institutional performance and why the covenantal facet matters. The same federal structure may produce cooperation in one episode and failure in another. What varies is not necessarily the constitutional text or design but the incentives facing participants, the knowledge processes, the trust among relevant centers, the presence of conflict-resolution practices, and the willingness to treat disagreement as a problem. A federal governance system is resilient not because it prevents all conflict or dysfunctions but because it offers means to address conflict and dysfunctions contextually, while preserving the possibility of error correction.

The “failure-focused” perspective is therefore the basis of a truly constructive approach. However, once failure becomes the organizing concept, a further question emerges naturally, as an enticing conjecture: Can a shared understanding of failure supply the common ground that an agreement on the good cannot?

Harm as a Focal Point for Overlapping Consensus

Rawls’s notion of “overlapping consensus” exemplifies the typical conceptualization of the problem of common ground and political legitimacy under reasonable pluralism. Yet it is evident that even shared values like freedom, equality, and security can generate deep dissensus and conflict once their meaning and ranking are contested. How should one stabilize the common ground on which to erect the governance systems of federalism?

This is where we invoke the thesis that agreement about harms may be easier to reach than agreement about goods. This approach assumes an asymmetry: People who disagree about the best form of family, education, religion, economic organization, or national purpose may still agree that certain things are not acceptable. Such agreement is thinner and more negative than a shared conception of justice. Precisely for that reason, it may be more available under deep pluralism.

The argument has affinities with Judith Shklar’s liberalism of fear, which places cruelty and abuse of power at the center of liberal concern; with Karl Popper’s priority of reducing suffering over maximizing happiness through comprehensive planning; and with John Gray’s account of modus vivendi among incompatible ways of life. Also one may turn to Thomas Schelling for a strategic insight: coordination need not require agreement on ultimate purposes, but may arise around “salient points” recognized as workable or mutually intelligible.

The extension of these ideas through Vincent Ostrom’s polycentric theory is straightforward and significant: Harms are experienced locally. They may appear as failures of public services, arbitrary administrative decisions, unsafe environments, exclusionary practices, or jurisdictional breakdowns. A polycentric order gives individuals who have had those experiences multiple arenas in which they can name, compare, contest, and address the harms they have suffered. The idea is not even to create a single central (“overlapping consensus”) definition of every harm. Instead it is to sustain a distributed capacity to discover harms and to prevent them from being concealed by administrative aggregation or political power.

This is where Ostrom’s design logic becomes truly relevant. Monitoring, accessible conflict resolution arenas, graduated sanctions, recognition of the right to organize, and nested arrangements are not merely instruments for producing efficiency. They are also ways of handling institutional failure without immediately resorting to centralized command. A harm-focused federalism thus asks whether individuals and communities can make grievances visible, whether the responsible center can be held accountable, and whether remedies can be proportionate to the problem. It preserves the distinction between serious harms that threaten the conditions of coexistence and contested claims that should remain open to local experimentation and political disagreement.

The resulting modus vivendi is not an agreement on the good life. It is an agreement to prevent certain forms of domination and breakdown, while leaving room for divergent projects. This is a liberalism of repair. It treats peaceable coexistence in conditions of deep social heterogeneity as a continuing accomplishment of institutions and citizens, who work out under broadly defined covenantal assumptions their problems and disagreements.

Let us note that the harm perspective also returns us to covenant. Jointly recognizing a harm is a thin form of mutual recognition. It says that the suffering, the costs incurred, or the institutional injury experienced by others cannot simply be dismissed because their comprehensive beliefs differ from ours. Joint repair creates repeated encounters in which actors learn that cooperation is possible, that commitments can be honored, and that one’s own vulnerability may depend on the protection of another’s autonomy. Over time, these practices may cultivate the trust and reciprocity that cannot be legislated directly. Covenant is not simply declared into existence; it must be continuously built through the experience of keeping promises, fixing problems, and repairing breaches.

The minimum common ground would therefore be neither just a common creed nor a mere procedural compliance. The most basic foundation would be in a shared commitment to detect, acknowledge, and remedy serious harms. That commitment remains an open, ongoingly renewable, and contestable proposition in its application. It is a covenantal, contextual commitment to the exercise of judgment. That commitment gives future specific judgments a direction compatible with pluralism: before asking the state or the polity to realize an elusive and contestable ideal, ask whether the proposed arrangement exposes some people to coercion, operational failure, exclusion, and the destruction of the institutional conditions of self-government. Agreement on that much may be sufficient common ground for building a working governance system in conditions of deep heterogeneity and rapid social change.

Conclusion

We are now in a position to gain a clearer understanding of Vincent Ostrom’s logic and where, following it through to its implications, it will lead us in our approach to federalism and polycentric governance systems. We have seen that Vincent Ostrom’s contribution can be reconstructed through three interconnected insights. 

First, institutional design is necessary, but it cannot fully master the complexity of self-governing orders defined by deep heterogeneity and rapid social change. Federal systems depend on knowledge, trust, reciprocity, and civic habits that no institutional blueprint can fully generate or sustain. Covenantalism and federalism are, in this sense, two facets of the same underlying reality.

Second, our approach to federalism should be based less on the search for optimal forms than on a focus on sources of failure. The relevant question is not which architecture promises perfection, but which arrangements reveal error, contain breakdown, and make correction possible. 

Third, this failure-first orientation suggests a possible answer to the question of what a heterogeneous people must still share. They may not share a comprehensive vision of the good. They may nevertheless share a commitment to notice and correct what threatens their capacity to live together as free and self-governing people. Recognition of serious harm is a modest foundation. It is also a demanding one.

A failure-first agenda would reorder institutional priorities. The first task is simply to be prepared to identify where failures are concentrated, how they travel across levels, and who bears their costs. The second task would be structural or constitutional, and it should be aiming at preserving exit, contestation, local experimentation, and the ability of affected actors to participate in rule revision. The third task would be organizational: We must strengthen the venues in which harms can be named and resolved before they become systemic crises. The fourth task may be broadly defined as “cultural” and “civic”: protect associations, congregations, civic groups, professional communities, and local and “third sector” institutions in which habits of mutual recognition, collaboration, and coexistence are learned.

This agenda also clarifies what one should decline even if it comes under the label of “liberalism.” One should decline both the liberalism of comprehensive design and the liberalism of passive proceduralism. The former overestimates what can be known and what can be legitimately and efficiently imposed from a center. The latter underestimates the cultural and relational conditions, the habits of hearts and minds that make procedures usable and legitimate. 

In this sense, Vincent Ostrom’s federalism may be understood as a working approach to the constitutional organization of coexistence under conditions of persistent disagreement, limited knowledge, and institutional fallibility. Its covenantal character does not imply the existence of a homogeneous political community. Rather, it points to the practices and commitments through which heterogeneous people may continue to constitute themselves as a community of self-governing citizens. Federalism becomes not the institutional expression of a unity already achieved, but one of the principal means through which a limited, revisable, and workable common ground can be ongoingly discovered, maintained, and renewed.

Bibliographical References

Aligica, P.D. (2013) Institutional Diversity and Political Economy: The Ostroms and Beyond. Oxford: Oxford University Press.

Aligica, P.D. and Boettke, P.J. (2009) Challenging Institutional Analysis and Development: The Bloomington School. London: Routledge.

Buchanan, J.M. (1990) The Economics and the Ethics of Constitutional Order. Ann Arbor: University of Michigan Press.

Elazar, D.J. (1987) Exploring Federalism. Tuscaloosa: University of Alabama Press.

Gray, J. (2000) Two Faces of Liberalism. Cambridge: Polity Press.

Hayek, F.A. (1974) “The pretence of knowledge,” Nobel Memorial Lecture, 11 December. Available at: https://www.nobelprize.org/prizes/economic-sciences/1974/hayek/lecture/.

Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press.

Ostrom, V. (1980) ‘Hobbes, covenant, and constitution’, Publius, 10(4), pp. 83–100.

Ostrom, V. (1991) The Meaning of American Federalism: Constituting a Self-Governing Society. San Francisco: ICS Press.

Ostrom, V. (1995) ‘Where to begin?’, Publius: The Journal of Federalism, 25(2), pp. 45–60.

Ostrom, V. (1997) The Meaning of Democracy and the Vulnerability of Democracies: A Response to Tocqueville’s Challenge. Ann Arbor: University of Michigan Press.

Ostrom, V. (2008) The Political Theory of a Compound Republic: Designing the American Experiment. 3rd edn. Lanham, MD: Lexington Books.

Popper, K.R. (1945) The Open Society and Its Enemies. London: Routledge.

Rawls, J. (1993) Political Liberalism. New York: Columbia University Press.

Schelling, T.C. (1960) The Strategy of Conflict. Cambridge, MA: Harvard University Press.

Shklar, J.N. (1989) ‘The liberalism of fear’, in Rosenblum, N.L. (ed.) Liberalism and the Moral Life. Cambridge, MA: Harvard University Press, pp. 21–38.

Tocqueville, A. de (2000) Democracy in America. Translated by H.C. Mansfield and D. Winthrop. Chicago: University of Chicago Press.


Paul Dragos Aligica

Contributor, Liberalism.org

Dragos Paul Aligica is KPMG Professor of Governance at the University of Bucharest, a Senior Research Fellow in the F. A. Hayek Program for Advanced Study in Philosophy, Politics and Economics at the Mercatus Center at George Mason University, and a Senior Nonresident Scholar with The Center for Governance and Markets at the University of Pittsburgh.

Full profile →

More in this theme

View more