Jensen on agency theory and corporate governance

Jensen’s influential paper with William Meckling (Jensen and Meckling, Reference Jensen and Meckling1976) laid the foundation for what has become known as agency theory. Their central contribution was to integrate property rights theory, principal–agent analysis, and corporate finance into a unified account of ownership structure and control. They famously defined the firm as a ‘nexus of a set of contracting relationships among individuals’ (Jensen and Meckling, Reference Jensen and Meckling1976: 311). On this view, the firm is not a unified actor with intentions, purposes, or moral agency. It is a legal and economic construct that facilitates coordination among contracting parties: owners and managers, employers and employees, firms and suppliers. Analytical attention is therefore directed away from questions such as ‘What is the objective of the firm?’ or ‘Does the firm have a social responsibility?’ and towards the problem of how individuals with potentially divergent objectives interact within contractual and organisational arrangements. Only individuals act; the firm itself does not.

Although Jensen and Meckling’s tools are often described as ‘neoclassical’ (Madden and Stevens, Reference Madden and Stevens2024: 3), their approach is better understood as institutional. As Gindis (Reference Gindis2020: 970) notes, they explicitly adopt the comparative institutional reasoning advocated by Coase (Reference Coase1964) and Demsetz (Reference Demsetz1969), building on the insights of Alchian and Demsetz (Reference Alchian and Demsetz1972) to analyse alternative ownership and control structures. Their objective was not to construct a psychologically rich account of behaviour, but to develop what Coase (Reference Coase, Williamson and Winter1993: 52) described as a ‘manageable’ theory; one in which organisational arrangements could be analysed using standard economic reasoning while remaining attentive to transaction costs, contractual incompleteness, and the allocation of decision rights. Agency theory, in this sense, is a framework for analysing how institutional arrangements structure incentives and constrain opportunism under conditions of dispersed knowledge and costly coordination.

A central implication of this institutional perspective is that the firm can be analysed as a set of agency relationships. An agency relationship arises whenever one party, the principal, delegates decision-making authority to another, the agent. In a world characterised by specialisation, dispersed knowledge, information asymmetries, and positive transaction costs, such delegation is not only ubiquitous but efficient. As Coase (Reference Coase1937) argued, firms emerge precisely where reliance on the price mechanism becomes too costly for coordination. Once firms are understood as institutional responses to these costs, and as collections of individuals acting under incomplete contracts, it becomes analytically natural to take self-interest as a foundational assumption. Following Alchian and Demsetz (Reference Alchian and Demsetz1972), Jensen and Meckling construct governance arrangements that are robust to predictable frictions, including self-interest. The purpose is not to assume the worst about individual motives, but to design contractual, ownership, and incentive structures that remain workable even when objectives diverge. Agency theory, on this account, provides a general framework for understanding the internal organisation of the firm as an economic response to delegated decision-making under conditions of incomplete information and costly enforcement.

A central concept formalised by Jensen and Meckling is that of agency costs. These arise whenever an agent is delegated decision-making authority but retains discretion to pursue objectives that may diverge from those of the principal. In the corporate context, managers may rationally engage in activities that generate private benefits, such as perquisite consumption, empire-building, or reduced effort, that do not maximise firm value. Agency costs are therefore not pathological or exceptional; they are a predictable consequence of delegated decision-making under incomplete contracts and asymmetric information.

Jensen and Meckling decompose agency costs into three analytically distinct components. First, monitoring expenditures are costs incurred by the principal to observe, evaluate, and constrain the agent’s behaviour; these include auditing, reporting requirements, performance evaluation, and oversight mechanisms. Second, bonding expenditures are costs borne by the agent to credibly signal alignment with the principal’s interests, such as contractual commitments, performance guarantees, or restrictions on discretion. Third, residual loss captures the expected value lost when, even after monitoring and bonding, the agent’s actions fail to maximise the principal’s objective. Unlike monitoring and bonding, which may involve observable expenditures, residual loss represents foregone surplus; gains from trade that do not materialise because incentive conflicts are imperfectly resolved.

While analytically separable, Jensen and Meckling emphasise that all three components of agency costs are ultimately reflected in firm value. In competitive capital markets, expected monitoring costs, bonding costs, and residual losses are capitalised into financing terms, ownership structures, and valuations. This creates systematic pressure for firms to adopt governance arrangements that economise on total agency costs rather than eliminate any single component in isolation. The importance of this framework lies not in identifying opportunism per se, but in specifying how alternative institutional arrangements trade-off monitoring, bonding, and residual loss under different organisational conditions. It is this general framework, rather than any particular governance prescription, that underpins the enduring influence of agency theory.

The analytical framework developed by Jensen and Meckling also generated a set of concrete implications for corporate governance. If agency costs are reflected in firm value, then governance mechanisms that constrain managerial discretion or align incentives should emerge endogenously. Ownership concentration, incentive-based compensation, debt financing, and contractual restrictions on discretion can all be understood as institutional responses to agency problems rather than as normative prescriptions. Agency theory thus shifted attention away from abstract debates about managerial purpose and towards the comparative analysis of governance arrangements that trade-off monitoring, bonding, and residual loss under different organisational conditions.

Jensen extended the agency framework to concrete governance problems. In his analysis of free cash flow, he argued that mature firms with limited growth opportunities face systematic incentive conflicts when internally generated funds exceed profitable investment opportunities (Jensen, Reference Jensen1986). Leverage and active markets for corporate control were presented as institutional mechanisms that constrain managerial discretion and reduce residual loss. Related arguments had been made earlier by Easterbrook (Reference Easterbrook1984), who emphasised dividend policy as a means of disciplining management. The same logic informed Jensen’s work on executive compensation; Jensen and Murphy (Reference Jensen and Murphy1990) contended that weak pay–performance sensitivity reflected misaligned incentives and advocated stronger equity-based incentives. In each case, agency theory framed governance innovations as endogenous responses to predictable agency conflicts rather than as moral reform.

Despite its analytical influence, agency theory has been subject to sustained criticism, particularly from management scholars and business ethicists (Fourcade and Khurana, Reference Fourcade and Khurana2017; Ghoshal, Reference Ghoshal2005; Ghoshal and Moran, Reference Ghoshal and Moran1996; Heath, Reference Heath2009; Khurana, Reference Khurana2007; see also McCloskey, Reference McCloskey2017). The core objection is not to its internal logic, but to its behavioural foundations and educational consequences. Critics argue that by modelling agents as self-interested and opportunistic, agency theory crowds out alternative accounts of managerial motivation grounded in professionalism, trust, or ethical commitment. On this view, the problem is not that agency theory is analytically false, but that it promotes an unduly narrow conception of managerial behaviour when treated as a general account of organisational life.

Rakesh Khurana (Reference Khurana2007) offers one of the most influential statements of this critique. He argues that the ascendancy of agency theory within business schools altered how managers understood their own role; shifting from a conception of managerial professionalism grounded in stewardship to one defined by contractual obligation and incentive alignment. This shift, Khurana suggests, had institutional consequences. Agency theory became embedded not merely as an analytical framework, but as a normative lens through which organisational problems were interpreted and addressed. Similar concerns are raised by Ghoshal (Reference Ghoshal2005), who contends that agency-theoretic models fostered cynicism by treating opportunism as the default condition rather than as a pathology to be constrained.

More polemical versions of this argument attribute a broader moral and institutional transformation to agency theory. McDonald (Reference McDonald2017), for example, characterises Jensen’s work as legitimating opportunism by reducing integrity to calculative self-interest, while McCloskey (Reference McCloskey2017) criticises agency theory for displacing ethical reasoning with instrumental rationality. In these accounts, agency theory is said not merely to explain corporate behaviour, but to have reshaped it by providing intellectual justification for short-termism, financialisation, and the erosion of managerial responsibility. Whether or not these claims are persuasive, they frame the context in which Jensen’s later concern with integrity was received and interpreted, often as an implicit admission that the earlier framework was ethically deficient.

These criticisms, however, conflate two distinct questions. The first concerns whether agency theory provides a useful positive framework for understanding governance problems within firms. The second concerns whether that framework ought to be treated as a comprehensive account of managerial purpose or ethical conduct. Jensen and Meckling’s contribution is directed squarely at the former. Agency theory does not purport to describe human motivation in all its richness, nor does it claim that managers ought to behave opportunistically. It offers a structured way of analysing how incentive conflicts arise under delegation, and how alternative institutional arrangements mitigate the resulting costs. When critics treat agency theory as a moral doctrine, they attribute to it claims it does not make.

From a methodological perspective, however, objections to the behavioural assumptions of agency theory are either irrelevant (Friedman, Reference Friedman1953) or, if relevant, uncontroversial (Coase, Reference Coase and Coase1981). Friedman’s argument is that economic theories should be judged by their explanatory and predictive power rather than by the descriptive realism of their assumptions. On this view, assuming self-interested behaviour is unobjectionable so long as it yields reliable insights into organisational structure and performance. Coase’s position is complementary rather than opposed. He did not object to abstraction per se, but to abstraction that lacked institutional content. His concern was not that economists assumed rationality or self-interest, but that they ignored transaction costs, legal constraints, and the allocation of decision rights. Jensen and Meckling’s framework responds directly to this concern by placing contracting, incentives, and governance at the centre of analysis. Their framework is entirely consistent with Williamson’s notion of pragmatic methodology: ‘first, keep it simple; second, get it right; third, make it plausible; fourth, make predictions and engage in empirical testing to see how the theory corresponds with real world experience’ (Hodgson and Gindis, Reference Hodgson and Gindis2007: 374).

This methodological stance is also consistent with Jensen’s own clarification of value maximisation. As he emphasised, maximising long-term firm value does not imply indifference to non-shareholder constituencies. As Jensen (Reference Jensen2001: 309) put it, ‘it is obvious that we cannot maximise the long-term market value of an organisation if we ignore or mistreat any important constituency. We cannot create value without good relationships with customers, employees, financial backers, suppliers, regulators, communities, and so on’. Read in this way, value maximisation functions as a performance criterion rather than as a moral doctrine, and agency theory remains a positive framework for analysing governance rather than a prescriptive account of managerial virtue.

Nor did agency theory introduce concerns about managerial discretion or non-value-maximising behaviour where none previously existed. Economists had long recognised that managers might pursue objectives other than profit maximisation, including private benefits and discretionary goals, well before Jensen and Meckling formalised these issues within an agency-cost framework (Alchian, Reference Alchian1965; Baumol, Reference Baumol1959; Friedman, Reference Friedman and Leube1970; Marris, Reference Marris1963; Williamson, Reference Williamson1964). What their contribution supplied was not the discovery of agency problems, but a tractable analytical structure for examining how alternative institutional arrangements respond to them.

Seen in this light, agency theory satisfies Coase’s criterion for a ‘good’ theory; it serves as a base for thinking about firms and helps organise inquiry into real institutional arrangements. It explains why particular ownership structures, incentive schemes, and control mechanisms emerge under different conditions, without claiming that these arrangements are ethically complete or socially sufficient. The common charge that agency theory ‘changed the world’ by promoting greed or short-termism, therefore rests on a category error. The theory clarifies existing governance problems and makes sense of observed practices; it does not prescribe a moral vision of the firm. If agency theory has been normatively overextended in business education or managerial practice, which reflects misapplication rather than a defect in the analytical framework itself.

Agency theory provides a powerful account of governance under incomplete contracting, explaining the emergence of ownership structures, incentive pay, and control as responses to agency costs. Its analytical focus, however, leaves under-specified how cooperation is sustained where monitoring is limited, contracts are incomplete, and enforcement relies on informal mechanisms.

In such settings, discretion is unavoidable and residual loss persists even under optimal formal governance. While agency theory can describe these conditions, it does not specify the mechanism through which informal constraints operate or substitute for formal control. It is at this margin that Jensen’s later contribution should be located; namely, an attempt to explain how agents voluntarily constrain behaviour when formal enforcement is costly or infeasible.

Jensen, human behaviour, and the turn to integrity

Jensen’s view of the public corporation evolved markedly over the course of his career, but this evolution should be understood as a reorientation of emphasis rather than a repudiation of agency theory. In the late 1970s and early 1980s, Jensen and Meckling defended the corporation as a highly productive organisational form that had emerged through market competition to economise on coordination costs, while warning that its survival was threatened by political and legal interference that weakened property rights (Jensen and Meckling, Reference Jensen and Meckling1978, Reference Jensen and Meckling1983). By the end of that decade, however, Jensen’s critique shifted inward. In The Eclipse of the Public Corporation, he argued that the internal control systems of large firms had failed and that boards of directors were often too passive and too slow to discipline management effectively (Jensen, Reference Jensen1989). This diagnosis reached its most forceful statement in his 1993 Journal of Finance Presidential Address, where he concluded that failures of internal control constituted ‘the most important problem in the modern corporation’ (Jensen, Reference Jensen1993: 850). What began as a defence of the corporate form against external threats had become a concern with its internal governance capacity.

At the same time, Jensen devoted increasing attention to the behavioural foundations of agency theory. Together with Meckling, he developed the REMM model of human behaviour, which characterised individuals as resourceful, evaluative, and maximising, while allowing for non-pecuniary motivations such as honour, affiliation, and concern for others (Jensen and Meckling, Reference Jensen and Meckling1994). REMM was explicitly intended to be broader than the narrow homo economicus assumption, while remaining compatible with rational choice. Jensen further suggested that REMM would be integrated with a Pain Avoidance Model (PAM) into a more comprehensive ‘dualistic’ theory of behaviour, though this programme was never completed. In the republication of the Jensen and Meckling (Reference Jensen and Meckling1994) paper, Jensen noted that this work was still at an early stage and therefore not included (Jensen, Reference Jensen1998: 4). Had this line of inquiry been pursued, cooperation and trust may have entered agency theory through richer accounts of preferences or cognition.

This, however, is not what Jensen’s later work on integrity does. The integrity project does not extend the behavioural model of the agent. It abandons that route altogether. Integrity operates through self-imposed constraint, not through altered preferences or bounded rationality. Agents remain fully rational and self-interested, but they bind themselves ex ante to rules that preclude opportunistic recalculation once commitments are made. The distinction is not merely terminological. A richer preference model of the REMM-PAM kind would have modified the agent’s utility function directly, so that non-pecuniary motivations such as honour or concern for others enter as arguments alongside material payoffs; the agent would then trade off opportunism against these intrinsic costs continuously, at every decision point. Integrity, by contrast, operates by contracting the feasible action set prior to the action stage; it does not alter what the agent values but removes certain actions from consideration altogether.

In this sense, integrity should be understood not as the fulfilment of the REMM–PAM agenda, but as an alternative response to the same underlying problem; how cooperation can be sustained when contracts are incomplete and enforcement is costly. Where REMM–PAM sought to explain cooperation by modelling behaviour differently, integrity seeks to explain it by altering the institutional environment within which behaviour takes place. Rationality is not softened; as I argue below, it is relocated from individual choice at the point of commitment to the choice of rules.

Seen in this light, Jensen’s claim that his later work on integrity was a ‘logical extension’ of the Jensen–Meckling framework becomes intelligible only at the level of governance. In a 2009 interview, Jensen stated that ‘it is a straight line from where I started [and] it turns out to be a logical extension of the Jensen-Meckling agency framework, although it wouldn’t appear that way to most people’ (Walkling, Reference Walkling2011). In Jensen’s own mind, this continuity lay in the shared concern with reducing waste arising from misaligned incentives. Yet the continuity does not run through behavioural enrichment. Integrity does not modify agency theory’s assumptions about motivation; it specifies a mechanism through which agents voluntarily constrain their future actions in order to reduce expected losses from opportunism under incomplete contracting.

A different interpretation of Jensen’s later work is offered by Dierksmeier (Reference Dierksmeier2020), who reads the introduction of integrity as part of a broader shift from a ‘mechanistic’ conception of economic behaviour grounded in self-interest to a more ‘humanistic’ approach that incorporates personal values and purpose. On this account, Jensen’s continued commitment to a positivist (Friedmanite) framework sits uneasily with this shift and may constrain its implications for management and business education. This is a plausible reading of the trajectory of Jensen’s work. The present analysis suggests another plausible reading by accepting Jensen’s own characterisation of the integrity project as a ‘logical extension’ of the Jensen–Meckling framework at face value.

Evidence of Jensen’s emerging concern with integrity appears in the early 2000s. In Jensen (Reference Jensen2003) and Jensen (Reference Jensen2005), integrity is used to diagnose organisational failure. In these papers, Jensen describes how internal corporate systems create incentives for managers to distort information, manipulate earnings, and engage in strategic misrepresentation. Behaviour of this kind is labelled ‘out-of-integrity’ not because it violates moral norms, but because it undermines organisational effectiveness and destroys value. Budgeting systems that reward beating targets, for example, encourage managers to game both the setting and the achievement of those targets, impairing coordination and decision-making (Jensen, Reference Jensen2003). Similarly, in his analysis of overvalued equity, Jensen argues that managers who manipulate earnings are ‘lying and making poor decisions that destroy value’ (Jensen, Reference Jensen2005: 8). In this usage, integrity is treated as a practical condition for organisational performance rather than as an individual virtue.

This orientation is made explicit in Jensen’s subsequent collaboration with Erhard and Zaffron. Integrity is defined as ‘the state of being whole, complete, unbroken, unimpaired, sound, in perfect condition’ (Erhard and Jensen, Reference Erhard and Jensen2017: 9). Jensen claims this definition to be explicitly non-normative; integrity is not treated as a moral virtue, but as a condition that applies equally to objects, systems, and human agents. Its significance lies in what they term ‘workability’; the capacity of a system to function effectively. Maximum workability is a necessary condition for maximum performance, and in the economic domain performance is ultimately understood in terms of long-run value maximisation (Erhard and Jensen, Reference Erhard and Jensen2017: 24–25).

For individuals and organisations, however, integrity is given a more precise operational content. It is a matter of one’s word, ‘nothing more and nothing less’ (Erhard et al. Reference Erhard, Jensen, Zaffron, Orlitzky and Monga2018: 17). An agent’s word includes not only explicit promises but also what is implicitly committed through roles, agreements, expectations, and accepted standards of conduct. To be ‘in-integrity’ is for this word to be whole and complete; that is, for commitments to be kept, or, when they cannot be kept, to be acknowledged promptly together with actions taken to address the resulting consequences. Conversely, integrity is reduced when commitments are broken, ignored, or left unaddressed. In this sense, integrity operates by constraining behaviour. Actions that would otherwise be privately advantageous become unavailable, or costly, because they would require the agent to depart from their word. Jensen summarises this relationship in what he terms the ‘law of integrity’; as integrity declines, workability declines, and as workability declines, the opportunity for performance correspondingly diminishes (Erhard and Jensen, Reference Erhard and Jensen2017).

Jensen’s conception of integrity can be located within the standard problem of moral hazard under incomplete contracting, but it operates through a distinct mechanism. Williamson (Reference Williamson1993) characterises opportunism as self-interest seeking with guile, encompassing not only misrepresentation and concealment but also the exploitation of contractual gaps and the violation of the spirit of agreements. In this framework, contractual incompleteness creates discretion, and moral hazard manifests in the range of opportunistic actions available within those gaps; governance structures are therefore required to mitigate its misuse. Hart (Reference Hart1995) formalises the same underlying problem by emphasising that ownership confers residual control rights; when contracts are silent, the holder of these rights determines how assets are used, and moral hazard arises in the exercise of that discretion. Both approaches treat opportunism as a behavioural possibility that must be constrained through external institutional arrangements, whether through contractual safeguards or the allocation of authority.

Jensen’s notion of integrity operates on a different margin of the same problem. Rather than reallocating control rights or strengthening contractual enforcement, integrity restricts how discretion is exercised and, in doing so, alters the agent’s response to moral hazard. To be in-integrity is to honour one’s word, or to renegotiate it and repair any resulting breach, thereby eliminating the scope for opportunistic recalculation ex post (Erhard and Jensen, Reference Erhard and Jensen2017). In this sense, integrity functions as a voluntary contraction of the agent’s feasible action set. Where Williamson analyses how moral hazard expands behaviour within the gaps of incomplete contracts, and Hart examines who controls those gaps, integrity removes the incentive to exploit them by binding the agent to a rule governing action. If interpreted as a bonding cost, as I argue below, integrity increases the private cost of opportunism and provides a non-contractual commitment device that operates within, rather than outside, the standard agency framework.

Jensen’s definition, however, introduces an additional ambiguity. Integrity is defined as being ‘whole, complete, unbroken’, which suggests a binary condition; one is either in-integrity or not. Yet throughout his later work Jensen repeatedly treats integrity as varying in degree across individuals and organisations. Firms are described as having higher or lower levels of integrity; organisational systems are said to operate with more or less integrity; and performance effects are discussed as marginal improvements associated with increases in integrity rather than as discrete shifts. Integrity is thus simultaneously presented as a binary condition at the level of commitment and as a continuous variable at the level of organisational outcomes. Jensen does not resolve this tension explicitly, but instead moves between the two conceptions as the argument requires.

This unresolved duality has contributed to confusion about the integrity project. If integrity is strictly binary, it is difficult to see how it can be incorporated into comparative institutional analysis or treated as a scalable governance mechanism. If it is continuous, it becomes unclear how this is compatible with the insistence that integrity requires the suspension of cost–benefit analysis at the moment of action. The integrity project thus appears to oscillate between a categorical rule-following conception and a marginal, performance-based conception, without specifying how the two relate.

Applied to organisational settings, integrity is operationalised as reliability in commitments; saying what one will do and then doing what one said. When individuals or organisations systematically break promises, distort information, or withhold relevant facts, trust erodes and coordination becomes more costly. The resulting losses appear not as explicit sanctions, but as hidden frictions: time spent monitoring, duplicated effort, failed negotiations, and foregone opportunities. In this sense, integrity is presented by Jensen as economically consequential. Subsequent empirical research is broadly consistent with this interpretation. In an explicit test of Jensen’s theory, Chen et al. (Reference Chen, Xia and Zhang2021) find that firms perceived to lack integrity face higher costs of debt and equity financing, while firms viewed as more reliable exhibit lower default risk and stronger accounting quality.

Despite Jensen’s insistence that he was advancing a positive theory, this formulation generates an apparent tension. Erhard and Jensen argue that integrity requires agents not to apply cost–benefit analysis at the moment of honouring commitments. Individuals must apply cost–benefit reasoning when giving their word, but once a commitment is made, it must be honoured regardless of subsequent incentives; treating integrity as a matter of continuous calculation, they contend, guarantees untrustworthiness (Erhard and Jensen, Reference Erhard and Jensen2017). This position appears to conflict directly with the standard economic assumption that agents continuously optimise, and it has led many critics to interpret the integrity project as covertly normative or behavioural. Lemann (Reference Lemann2019) and McCloskey (Reference McCloskey2017), for example, interpret Jensen’s integrity claims as an attempt to reform character rather than institutions. Indeed, Demsetz (Reference Demsetz1969) identifies exhortations that ‘people could be different’ as a form of nirvana reasoning.

This apparent conflict has generated much of the confusion surrounding Jensen’s late work. The tension dissolves once integrity is understood not as a change in preferences, but as voluntary self-binding chosen ex ante. Interpreted as an endogenous bonding cost, integrity does not suspend optimisation; it relocates it to the choice of rules. On this account, integrity functions as an institutional mechanism within the Jensen–Meckling framework rather than as an ethical supplement to it.

Integrity as a bonding cost

Integrity can be understood as a form of voluntary bonding. In Jensen and Meckling’s (Reference Jensen and Meckling1976) framework, agency costs comprise monitoring expenditures, bonding expenditures, and residual loss. While monitoring and incentive alignment have received extensive attention, bonding remains under-specified; Jensen and Meckling describe it only as resources expended by the agent to assure the principal that harmful actions will not be taken (1976: 308). Integrity fills this gap. It operates as a self-imposed constraint that limits opportunistic behaviour and substitutes, at the margin, for external enforcement. Pagano and Vatiero (Reference Pagano and Vatiero2015) argue that institutions are themselves costly and that substitution among them is constrained by complementarities and path dependence; integrity, on that interpretation, operates as one such costly substitute.

Seen in this way, integrity is not an ethical add-on to agency theory, but a particular way of bearing bonding costs. An agent who internalises integrity commits to honouring representations and obligations even when opportunistic deviation would be privately advantageous. This commitment is costly. It involves foregone gains from opportunism, the effort required to maintain consistency under pressure, and the willingness to absorb losses that could otherwise be shifted onto the principal. These are bonding costs borne by the agent rather than imposed contractually.

Framing integrity as a form of bonding, however, raises a further problem that agency theory does not resolve. If integrity involves real and ongoing costs in the form of foregone opportunism and voluntary loss-bearing, then it cannot be treated as a marginal choice recalculated at each decision point. The relevant question is therefore not whether integrity is costly, but how and when a rational agent would choose to bind themselves to such a constraint in the first place, given uncertainty about their future positions, incentives, and opportunities. It is at this point that James M. Buchanan’s constitutional framework becomes analytically useful.

What Buchanan adds, beyond this general endorsement of self-binding, is a precise account of when and how such constraints are chosen. In Buchanan’s constitutional political economy, individuals distinguish between choices over rules and choices within rules (Brennan and Buchanan, Reference Brennan and Buchanan1985; Buchanan, Reference Buchanan1991). At the constitutional stage, agents select general constraints on their own future behaviour under conditions of uncertainty about their future positions, incentives, and opportunities (Brennan and Buchanan, Reference Brennan and Buchanan1985). At the post-constitutional stage, agents act within those constraints, taking them as given rather than as objects of ongoing optimisation. This distinction is central to understanding integrity as an economically rational form of self-binding.

Integrity, on this view, is chosen at the constitutional stage rather than at the level of particular transactions (Buchanan, Reference Buchanan1994). An agent does not decide whether to honour their word on a case-by-case basis; instead, they choose to operate under a rule that precludes opportunistic recalculation once commitments have been made. This explains why integrity can be rational without being continuously optimised. The economic rationality of integrity lies in the ex-ante choice to adopt a binding constraint, not in the ex-post suspension of cost–benefit analysis when honouring a commitment. Once the rule is chosen, deviation is no longer treated as an option to be weighed against alternatives; it is treated as a violation.

This constitutional interpretation resolves the apparent tension in Jensen’s account between positive analysis and behavioural commitment. Integrity does not require agents to abandon rationality in particular choices. Rather, rationality operates at an earlier (constitutional) stage in which agents choose to bind their own future discretion. In Buchanan’s terms, integrity functions as a self-imposed rule that restricts the opportunity set available at the post-constitutional stage in order to improve outcomes across a wide range of future interactions. The cost of integrity is borne precisely because it forecloses opportunistic responses that would otherwise appear locally advantageous.

This perspective also clarifies why integrity cannot be fully contractualised. Constitutional rules, by their nature, precede and structure contractual relations rather than being reducible to them. Integrity operates prior to, and independently of, formal contracting because it governs how representations, promises, and commitments are made in the first place. Attempting to specify integrity contractually converts it from a self-binding rule into an object of enforcement, thereby undermining the very mechanism through which it economises on governance costs. In Buchanan’s framework, integrity belongs to the category of rules that make contracting possible, not to the category of terms that contracts can efficiently specify.

Seen in this light, integrity is neither a moral residue nor a behavioural anomaly within agency theory. It is a constitutional constraint voluntarily adopted by agents who recognise that opportunism, while locally profitable, is collectively destructive in environments characterised by repeated interaction and incomplete contracting. By choosing integrity ex ante, agents alter the structure of their future incentives and expand the range of cooperative outcomes available to them. This is precisely the role Buchanan assigns to ethical rules within economic order.

To make the mechanism explicit, consider a stylised principal–agent setting with hidden action. A principal delegates a task to an agent whose choice is not directly observable. The agent selects a level of opportunistic behaviour u ≥ 0. Opportunism is defined as actions that increase measured performance but reduce true firm value. This captures familiar behaviours such as earnings manipulation, misreporting, or effort misallocation that inflate observable metrics while imposing real costs on the principal.

Measured performance is given by:

$y=u+\varepsilon$

where ϵ is a mean-zero random variable capturing exogenous noise in performance measurement. The linear specification isolates the incentive effect of opportunism without introducing additional curvature; it implies that opportunistic actions are indistinguishable from genuine performance improvements in the observed signal.

The agent is compensated according to a standard linear contract

$ w=\omega +by$

where ω is a fixed payment and b > 0 is the incentive intensity. The parameter b determines the marginal return to measured performance and therefore the strength of incentives to engage in opportunistic behaviour.

Opportunism imposes a real cost on the principal, captured by a loss function D(u), with D′(u) > 0. This cost represents the reduction in true firm value associated with opportunistic actions. The precise functional form of D(u)is not required for the agent’s optimisation problem, but it underpins the interpretation of opportunism as socially costly and gives rise to residual loss in the Jensen–Meckling decomposition.

The agent is risk-neutral and chooses uto maximise expected utility

$ U=\mathbb{E}[w]-C(u;I)$

where C(u; I) is the private cost of opportunistic behaviour. The expectation operator reflects uncertainty in ϵ, but since ϵ is mean-zero, expected compensation simplifies to $\mathbb{E}[w]$ = ω + bu.

The cost function is specified as

$ C\left(u;I\right)={1 \over 2}\left(c+\kappa \left(1+\lambda I\right)\right)u^{2}$

with parameters defined as follows:

c > 0: baseline marginal cost of opportunism; this captures effort costs, risk of detection, or internal discipline independent of institutional constraints;

κ > 0: scaling parameter governing the strength of informal constraints;

λ > 0: sensitivity of the marginal cost of opportunism to integrity;

I ∈ [0,1]: integrity parameter.

The quadratic form ensures convexity, so that marginal costs increase in u, and delivers a unique interior solution. More importantly, the multiplicative structure c + κ(1+λI) allows integrity to enter as a cost shifter. Following Fischer and Huddart (Reference Fischer and Huddart2008), informal constraints are represented as factors that increase the marginal cost of opportunistic behaviour rather than as elements of the utility function or contractual terms. Integrity is therefore not a preference for honesty, nor an externally imposed rule, but a parameter that alters the agent’s opportunity set.

The integrity parameter I is interpreted as a constitutionally chosen self-binding constraint. It is selected ex ante, at the contracting stage, and taken as given at the action stage. Higher values of I increase the marginal private cost of opportunism through reputational concerns, internal discipline, or anticipated informal sanctions. The restriction I ∈ [0,1] is a normalisation that allows integrity to be interpreted as the strength of an informal institutional constraint.

Substituting expected wages into the utility function, the agent’s problem becomes

$ \underset{u\geq 0}{\max\; }\omega +bu-{1 \over 2}\left(c+\kappa \left(1+\lambda I\right)\right)u^{2}$

The first-order condition is

$ b-(c+\kappa (1+\lambda I))u=0$

which yields the equilibrium level of opportunism

$ u^{*}={b \over c+\kappa (1+\lambda I)}$

This expression establishes the central comparative static. Opportunism increases in incentive intensity b and decreases in the marginal cost parameters c, κ, and λ, as well as in the integrity parameter I. In particular,

$ {\partial u^{*} \over \partial I}\lt 0$

so higher integrity reduces opportunistic behaviour.

Integrity raises the private marginal cost of opportunism without altering preferences, contracts, or monitoring technology. As a result, it reduces equilibrium opportunism and, by implication, reduces residual loss. From the perspective of the principal, integrity therefore substitutes, at the margin, for monitoring expenditures. From the perspective of the agent, it represents a bonding cost; a self-imposed constraint that entails foregone gains from opportunistic behaviour.

The formalisation highlights two general implications that will guide the discussion that follows. First, integrity operates by shifting the private cost of opportunism rather than by altering preferences or contractual terms. As a result, integrity and monitoring are substitutes at the margin; increases in I reduce equilibrium opportunism u* and economise on external enforcement. Second, because I functions through decentralised sanctions rather than formal verification, its effectiveness depends on the institutional environment that sustains those sanctions. These implications motivate the analysis of integrity as an informal institution and of the conditions under which integrity-based bonding can or cannot substitute for formal governance.

At first glance, integrity appears to be a binary concept; promises are either honoured or breached, representations are either truthful or deceptive. In this sense, the intuition that integrity is not continuous is correct at the level of rule compliance. However, the model operates at a different level of analysis. What varies is not the rule itself, but the expected cost and likelihood of breach. These depend on the strength and reliability of decentralised enforcement mechanisms, which vary continuously with context. Integrity is therefore binary as a rule, but continuous as an institution.

This distinction is captured by Voigt’s (Reference Voigt2018) taxonomy of institutions. The integrity parameter I should not be interpreted as a moral preference for honesty, nor as a purely cultural disposition. It represents an informal institutional constraint that lies between custom and formal private rule. Integrity is more than a convention because deviation can be privately profitable and compliance is not automatically self-enforcing; yet it is less than a formal rule because adherence is not guaranteed by courts or hierarchical authority. Integrity is sustained instead by decentralised sanctions such as reputation loss, exclusion from future exchange, and the withdrawal of discretionary autonomy. This institutional location is a boundary condition for the model. I captures the strength of an informal constraint that is neither mere habit nor formal enforcement.

With this interpretation in place, integrity can be incorporated directly into the agency cost framework. Let total agency cost be composed of monitoring costs M, bonding costs B, and residual loss L. Integrity affects all three:

$ T\left(I\right)=M\left(I\right)+B\left(I\right)+L\left(I\right)$

with M′(I) < 0, B′(I) > 0, and L′(I) < 0. Greater integrity substitutes internal self-binding for external enforcement. From the principal’s perspective, integrity reduces the need for monitoring; from the agent’s perspective, it replaces external discipline with internal constraint. The choice between integrity and monitoring is strategic. Agents trade-off the private cost of self-binding against the burdens imposed by monitoring regimes, while principals trade-off monitoring expenditures against residual loss.

Interpreting integrity as a bonding cost clarifies Jensen’s intellectual trajectory. His late emphasis on integrity does not abandon agency theory’s core assumptions; it identifies a mechanism through which agents can credibly commit to cooperative behaviour when contracts are incomplete and enforcement is costly. Integrity supplies an internal correction to the agency framework, not by rejecting its logic, but by specifying how informal institutional constraints can economise on agency costs.

This raises a further institutional question. If integrity is an informal constraint located between custom and formal private rule, how does it interact with other informal institutions, and under what conditions does it complement or conflict with formal governance structures? The next section addresses this question by examining integrity within a broader taxonomy of informal institutions.

Integrity as an Informal Institution and its Limits

Having reconstructed integrity as a voluntary bonding mechanism within the agency framework, the next step is to locate it within the broader institutional environment in which governance takes place. A useful starting point is the distinction between formal and informal institutions (Hodgson, Reference Hodgson2025). Formal institutions rely on explicit rules and centralised enforcement; informal institutions operate through decentralised mechanisms such as reputation, reciprocity, and exclusion. Integrity does not fit neatly at either pole (Voigt, Reference Voigt2018). It is intentionally adopted by agents and organisations but enforced through decentralised responses rather than third-party coercion.

These institutional features determine the effectiveness of the integrity parameter I in the formal model. Integrity raises the expected private cost of opportunism through informal sanctions such as reputational loss, diminished discretion, exclusion from future exchange, or the withdrawal of trust. Where such sanctions are credible and observable, increases in I translate into higher marginal costs of opportunistic action and lower equilibrium levels of opportunism u*. Where enforcement is weak or noisy, the same nominal commitment has little effect. Integrity is therefore binary as a behavioural rule, but continuous as an institutional constraint, because its effectiveness depends on the strength of decentralised enforcement.

This interpretation distinguishes integrity from culture, norms, and preference-based models. Culture refers to broad patterns of shared meaning; norms describe behavioural regularities sustained by mutual expectations. A related literature incorporates social norms into agency models by treating norm compliance as part of the agent’s utility function (for example, Stevens, Reference Stevens2018; Stevens and Thevaranjan, Reference Stevens and Thevaranjan2010). In those models, informal constraints operate through changes in preferences, so that opportunistic behaviour adjusts smoothly in response to incentives. By contrast, integrity is modelled here as a self-imposed constraint on the feasible set of actions, chosen prior to action in response to governance problems. It operates on a different margin. Once adopted, it restricts admissible actions rather than altering marginal trade-offs. When integrity binds, marginal changes in incentive intensity do not induce continuous adjustments in opportunistic behaviour but affect outcomes only at the boundary where the constraint ceases to bind. At the level of reduced-form equilibrium outcomes, the two formulations may be difficult to distinguish empirically; a sufficiently steep preference cost of opportunism in a utility-based model can approximate the corner solution produced by a binding feasible-set constraint. The interpretations diverge, however, in the conditions under which each mechanism fails. Preference-based accounts depend on the agent’s motivational endowment remaining stable across contexts; if preferences shift under pressure or vary across individuals, the informal constraint weakens continuously and without a clear threshold. A feasible-set constraint, by contrast, fails at a determinate point, namely when the decentralised sanctions that sustain it, reputation loss, exclusion, and the withdrawal of discretion, can no longer be credibly applied. This distinction matters for the scaling argument developed below; the conditions under which integrity-based bonding attenuates are institutional rather than psychological, and are therefore amenable to organisational analysis in a way that preference instability is not.

This interpretation is consistent with Buchanan’s distinction between choices within rules and choices over rules. Integrity operates at the level of rule selection; it is chosen ex ante as a commitment that governs subsequent behaviour. Its relationship with formal governance is therefore conditional. Integrity can reduce the marginal return to monitoring and support delegation, while formal rules can stabilise expectations and protect those who self-bind. However, the effectiveness of integrity depends on institutional conditions. Because enforcement is decentralised, it relies on observability, repeated interaction, and credible reputational consequences.

These conditions weaken as organisational scale increases. Specialisation and layering reduce observability; reputational signals become noisier; and individual actions have diminishing effects on collective beliefs. As a result, the mapping between I and u* attenuates. Higher declared or constitutionally chosen levels of integrity no longer produce proportionate reductions in opportunism. Principals must then either tolerate higher residual loss or reintroduce formal monitoring. Attempts to mandate integrity at scale require verification through reporting and oversight, transforming it from a self-imposed bonding mechanism into a compliance requirement.

This outcome reflects the institutional limits of informal bonding rather than a failure of integrity as such. Integrity functions effectively only within bounded domains where decentralised sanctions are sufficiently strong to sustain its cost-shifting effect. Beyond those domains, it either collapses into rhetoric or is absorbed into formal control. Its advantage, reducing opportunism without additional monitoring, is lost. Effective governance therefore requires alignment between informal bonding and formal control, and that alignment becomes increasingly fragile as organisational scale increases.