When Paul DiMaggio introduced the term institutional entrepreneurship in 1988, he was addressing a gap in institutional theory. The framework explained stability well. Organizations converged, practices diffused, and routines hardened into taken-for-granted rules. What it could not easily explain was change. DiMaggio's answer was that new institutions arise when organized actors with enough resources see in them a chance to realize interests they value highly.

The concept restored agency to the theory, but it also exposed a puzzle that later scholars called the paradox of embedded agency. If institutions shape actors' preferences, cognitions, and identities, how can those same actors imagine alternatives, let alone build them? The literature has often answered with success stories: the reformer who remade a profession, or the founder who created a new organizational form. These narratives suffer from survivorship bias. For every arrangement that took hold, many more attempts dissolved, were absorbed by incumbents, or lived on as marginal curiosities.

These failures are not random. They follow a structural logic that can be studied comparatively. Three obstacles recur across sectors, cultures, and historical periods: the legitimacy paradox, the resource mobilization problem, and the collective action problem. They interlock, and they share a root. Existing institutions are not a neutral backdrop against which entrepreneurs act. They are the environment from which entrepreneurs must draw almost everything they need, including credibility, capital, and cooperation.

Mark Suchman's widely used definition treats legitimacy as a generalized perception that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions. The decisive phrase is the final clause. Legitimacy is conferred within an existing system of meaning, and a genuinely new institutional form, by definition, sits partly outside that system.

Howard Aldrich and Marlene Fiol separated cognitive legitimacy, meaning taken-for-grantedness, from sociopolitical legitimacy, meaning endorsement by key stakeholders and authorities. Arthur Stinchcombe had earlier described the liability of newness that young organizations face. Novel institutional forms carry this liability twice over, because both the organization and the category it belongs to are unfamiliar. Ezra Zuckerman's research on securities analysts showed that firms which fit no established category suffer an illegitimacy discount. Evaluators struggle to price what they cannot classify.

The result is a closed loop. A new arrangement succeeds only if it is adopted. It is adopted only if it is seen as legitimate. It is seen as legitimate only once its success has been demonstrated. Regulators wait for evidence, funders wait for precedent, and potential adopters wait for each other. Each of these postures is individually reasonable, and together they produce stasis.

The usual escape is to borrow legitimacy. Andrew Hargadon and Yellowlees Douglas showed how Edison presented electric lighting through what they called robust design. He modeled his system on the gas industry it would displace. Lines ran underground like gas mains, consumption was metered, central stations echoed gasworks, and early bulbs were made roughly as bright as the gas jets people already knew. The novelty arrived wearing familiar clothing, which made it legible to investors, municipal authorities, and households.

Borrowing has a cost, and DiMaggio and Powell's account of mimetic isomorphism anticipates it. An entrepreneur who gains acceptance by resembling the incumbent may find that the resemblance stops being a disguise and becomes part of the new form itself. The drift of parts of microfinance toward conventional commercial lending is one illustration. Many entrepreneurs therefore fail in one of two directions. Some are too novel to be understood, and others become too familiar to transform anything.

Takeaway

Novelty and legitimacy pull against each other, so compromise cannot be avoided. The strategic question is which features of a new arrangement are essential and which can safely wear borrowed clothing.

Resource mobilization theory, developed by John McCarthy and Mayer Zald, made a useful observation: grievances are nearly universal, but access to the means of acting on them is not. Institutional entrepreneurs need money, but they also need network position, formal authority, technical expertise, and what might be called discursive standing, which is the recognized right to define problems and propose solutions.

The difficulty is that existing institutions distribute these resources according to their own logics. Capital moves through intermediaries whose evaluation criteria were built around established forms. Professions certify expertise. Regulatory agencies grant access, and they already have long-standing clients. An entrepreneur seeking resources is in effect asking the beneficiaries of the current order to finance its replacement.

This creates a center–periphery dilemma. Research on American radio broadcasting by Huseyin Leblebici and colleagues found that innovations often came from fringe players, who had the least to lose. Peripheral actors have the motivation but lack the resources. Central actors have the resources but are deeply embedded in the existing arrangements. When elites do become entrepreneurs, as Royston Greenwood and Roy Suddaby documented with the large accounting firms that pushed multidisciplinary practice, they tend to do so after hitting the limits of their field. The changes they pursue usually extend their dominance rather than redistribute it.

Resources that do arrive come with conditions. Raghu Garud and colleagues studied Sun Microsystems' sponsorship of Java and found that other firms would commit investment only if Sun gave up meaningful control over the standard. The entrepreneur needed partners, and the partners reshaped the project. Philip Selznick's earlier study of the Tennessee Valley Authority described the same dynamic as co-optation. Absorbing local elites secured the agency's survival, and the price was a change in its mission.

Entrepreneurs are therefore often caught between starvation and capture. Those who manage to escape tend to share two strategies. First, they draw resources from many different sources, so that no single patron can set the terms. Second, they seek out actors whom the status quo already disadvantages, such as dissatisfied incumbents, organizations in adjacent fields, or newcomers locked out of existing advantages. For these actors, backing change is less a sacrifice than an investment.

Takeaway

Every resource carries the fingerprints of the arrangement that produced it. Before accepting one, ask what obligations and assumptions arrive with it, as well as whether you can get it at all.

No institution belongs to one actor, so no single actor can change one. Mancur Olson's The Logic of Collective Action explains why this matters. Large groups that share diffuse benefits tend to underprovide public goods, while small groups with concentrated stakes organize effectively. Institutional change usually fits this pattern. It offers diffuse, uncertain gains to many actors and concentrated, immediate losses to a few. This asymmetry is systematic, which means incumbents begin every contest with an organizational advantage.

Coordination problems make the situation worse. The benefits of many new arrangements depend on widespread adoption because of network effects, complementary practices, and shared standards. For any single actor, waiting is rational. The American metric effort shows how this plays out. The Metric Conversion Act of 1975 declared the metric system preferred but made conversion voluntary, and the U.S. Metric Board was dissolved in 1982. Each firm faced switching costs on its own, while the benefits required near-universal participation.

Coalitions also have to unite actors who want different things from the same change. Neil Fligstein argues that institutional entrepreneurship depends on social skill, the ability to induce cooperation by offering others identities and frames in which they can see their own interests. Ambiguity becomes a resource in this work. John Padgett and Christopher Ansell's analysis of Cosimo de' Medici describes robust action: conduct that is multivocal enough for different audiences to read different purposes into it.

Ambiguity has limits, however. Implementation forces abstract commitments into concrete allocations of budgets, authority, and credit, and coalitions often fracture at that point. In Implementation, Jeffrey Pressman and Aaron Wildavsky traced how a well-funded federal program in Oakland stalled as the number of required clearances multiplied. Each participant had agreed to the goal. Few had agreed to the specific burdens the goal required of them.

Timing compounds these problems. The costs of change come early and are certain, while the benefits come later and are speculative. Incumbents defending a position can sustain their effort, but challenger coalitions decay as attention moves elsewhere. Change therefore tends to come in bursts rather than gradually. It often follows crises that align expectations, or it is driven by what Olson called a privileged group, in which one member gains enough to bear the costs alone.

Takeaway

Institutional change depends less on convincing people that an alternative is better than on solving two problems: nobody wants to move first, and everyone wants something slightly different from the move.

The three obstacles share a common source. Institutional entrepreneurs must draw legitimacy, resources, and cooperation from the very order they are trying to change. The paradox of embedded agency is therefore more than a theoretical puzzle. It is the practical condition under which every reformer works.

This has consequences for institutional design. James C. Scott warned against grand schemes imposed on complex social orders without metis, the practical knowledge that keeps those orders functioning. Successful entrepreneurs rarely replace institutions outright. More often they graft. Wolfgang Streeck and Kathleen Thelen describe processes of layering and conversion, in which new elements are attached to old structures and gradually redirect their purpose.

For those who fund, govern, or attempt reform, high failure rates are not a judgment on the quality of ideas. They reflect structural position. The most useful points of intervention lie in the infrastructure of change: the categories that confer legitimacy, the channels through which resources flow, and the devices that let actors move together rather than wait for one another.