CompanyGraph
Delegating a decision creates a gap between who acts and who bears the result. Governance is the cost of narrowing that gap, not proof that every manager is misaligned.
Agency is a relationship, not a personality diagnosis
In an agency relationship, a principal delegates a task to an agent whose information or skill makes delegation worthwhile. Shareholders and executives are one example; lenders and borrowers, boards and chief executives, customers and contractors, and limited partners and fund managers are others.
Michael Jensen and William Meckling's 1976 theory of the firm defines agency costs as monitoring expenditures by the principal, bonding expenditures by the agent, and residual loss from decisions that still diverge from the principal's best achievable outcome. The framework does not say that the principal can observe the perfect decision or that every difference in outcome is misconduct.
An owner-manager may spend less on monitoring because interests are concentrated. A public company may spend more on audit, board oversight, disclosure, compensation design, and internal controls because ownership is dispersed and the agent knows more about daily operations. Those controls consume money and management time, but they can also make delegated production possible.
What incentives can change
Compensation changes the payoff attached to an action. A bonus on revenue can encourage selling, but it can also reward unprofitable volume. A stock option can connect management wealth to equity value, but its payoff may encourage volatility or short-term share-price management. A cash-return or return-on-invested-capital target can discourage acquisitions that are useful but initially dilutive, or encourage cutting investment that would have produced later capacity.
The SEC requires U.S. public companies to disclose executive compensation and the criteria used in compensation decisions. Disclosure tells an investor what the board says it is rewarding. It does not establish that the metric caused the decision, that performance was measured well, or that the outcome was good for long-term owners.
Information makes control expensive
Managers usually know more about customers, staff, product quality, and operational bottlenecks than outside owners. A board can request reports, appoint an auditor, or hire an independent adviser, but each observation has scope and timing limits. A clean financial statement can coexist with deferred maintenance, a fragile supplier, or a project whose value depends on assumptions not yet tested.
Monitoring can also distort behaviour. If a board measures only quarterly earnings, management may defer hiring or maintenance. If it demands detailed approvals for every decision, the organization may lose speed and local knowledge. The relevant question is not whether monitoring is present but whether it reaches the decision that matters at the time it can still be changed.
Governance trades one risk for another
Boards, voting rights, debt covenants, activist investors, takeovers, clawbacks, and managerial ownership each constrain some behaviour. None is a universal solution. A lender may discipline spending that threatens repayment but prefer asset sales that reduce long-term capacity. A founder with large ownership may think long-term but also resist a succession plan. A highly independent board may challenge management effectively or lack the industry knowledge to evaluate a complex investment.
Agency language is especially dangerous around acquisitions. A large deal may reflect empire building, but it may also secure technology, remove a competitor, open distribution, or solve a capacity constraint. A recent study of CEO compensation changes after acquisitions found little support for a simple traditional agency explanation in its sample. That evidence is a reminder to test the mechanism rather than treat size or acquisition frequency as proof.
Follow the decision and the resource
The strongest agency analysis traces a decision to a physical or financial consequence. If management closes a plant, which capacity disappears and who receives the saving? If it cuts research, which future products become unavailable? If it buys a company, what integration work, debt, people, and customer commitments follow? If it changes a bonus metric, which reported number can improve without improving the underlying operation?
Money and authority determine whether correction is possible. A board may identify a problem but lack reliable operational data. A shareholder may vote against a director but have no alternative management team. A lender may have a covenant right but not the expertise or incentive to preserve the business. Agency costs are therefore not only the manager's private benefit; they include the resources spent trying to observe, constrain, and repair the divergence.
Trace the agency relationship
- Map the principal and agent. Who decides, who owns the upside, who bears the downside, and who can replace or constrain the decision-maker?
- Read the contract. What targets, vesting, clawbacks, covenants, voting rights, or termination terms change the payoff?
- Compare metric with reality. Does the rewarded measure track cash, capacity, quality, customer retention, and risk, or can it improve while those deteriorate?
- Test alternatives. Could the same action reflect customer needs, regulation, taxes, financing, or an information advantage rather than agency?
- Measure residual loss. After monitoring and incentives, what costly decisions remain unexplained, and who has authority and money to correct them?
Agency theory is most useful as an investigation discipline. It tells the investor to look at divided authority, information, contracts, and consequences. It does not supply a moral judgment in advance; the evidence must show whether the arrangement makes the underlying business more or less capable of producing the promised result.