Agency Problem
The conflict of interest that arises when an agent (manager, advisor, lawyer, broker) is hired to act in a principal's (owner, client) interest but has their own interests that may diverge. Munger considered the agency problem one of the most pervasive and underappreciated forces in business and finance.
“For instance, economists, speaking from the employer's point of view, have long had a name for the natural results of incentive-caused bias: "agency cost." As the name implies, the economists have typically known that, just as grain is always lost to rats, employers always lose to employees who improperly think of themselves first.”
“For instance, economists, speaking from the employer's point of view, have long had a name for the natural results of incentive-caused bias: "agency cost." As the name implies, the economists have typically known that, just as grain is always lost to rats, employers always lose to employees who improperly think of themselves first.”
Concept Analysis
Definition & Origins
The Agency Problem is the structural conflict of interest that arises when one party (the agent) is authorized to act on behalf of another (the principal), but the agent's personal interests diverge from the principal's interests. In corporate finance, the canonical agency problem is between shareholders (principals) and management (agents): shareholders want management to maximize long-term enterprise value; management may prefer to maximize their own compensation, minimize personal risk, build empires, or simply enjoy organizational comfort. Because shareholders cannot perfectly monitor management behavior, and because management controls the information shareholders receive, agents can systematically pursue their own interests at shareholders' expense.
Munger absorbed the agency problem framework from academic finance but applied it with characteristic breadth and specificity — extending it beyond the shareholder-management relationship to every principal-agent structure he encountered: auditor-client, analyst-investor, investment banker-issuing company, lawyer-client, physician-patient. Wherever one party acts on behalf of another and their incentives diverge, the agency problem operates.
Core Ideas
The information asymmetry driver. Agency problems are most severe when the principal cannot observe or evaluate the agent's performance directly. A company's board of directors cannot directly observe whether management is making optimal capital allocation decisions, or whether reported earnings accurately reflect economic reality. A patient cannot directly evaluate whether a physician's treatment recommendation is medically optimal or driven by billing incentives. The agency problem intensifies as information asymmetry increases.
The incentive misalignment taxonomy. Munger identified several specific forms of incentive misalignment that characterize agency relationships:
-
Empire building: management that derives status, compensation, and job security from organizational size will make acquisitions and capital investments that increase size without improving returns. The acquisition of General Re by Berkshire in 1998 revealed that General Re's management had built a derivatives book that served their interests (complex, high-revenue, impressive) rather than shareholders' interests (transparent, low-risk, profitable).
-
Risk asymmetry: agents who share in upside but are protected from downside will systematically take more risk than principals would approve. Financial institution employees compensated on annual bonus structures captured the upside of risk-taking in good years while leaving the downside — losses and potential bankruptcy — to shareholders and, ultimately, taxpayers.
-
Short-termism: agents compensated on short-term metrics (quarterly earnings, annual return targets) will systematically underinvest in long-term value creation (brand building, research, employee development) when those investments require short-term sacrifice.
-
Information control: agents control what information principals receive, which allows them to present their performance selectively, emphasize successes, and manage perceptions of failures. The mechanism is Munger's Incentive-Caused Bias operating on information reporting: management whose compensation depends on appearing successful will unconsciously shape information to make their performance appear favorable.
The board failure mode. Munger was specifically critical of corporate board structures that create an agency problem at the governance level. When directors are nominated by management, compensated primarily through the shares of the company they nominally oversee, and rely on management for the information they use to evaluate management performance, the board operates as an agent of management rather than of shareholders. The independence required to fulfill the board's principal-agent role is structurally compromised.
Practical Application
Executive compensation design. Munger's consistent prescription for mitigating the management-shareholder agency problem was compensation design that aligns long-term incentives. Compensation heavily weighted toward long-duration equity (options with multi-year vesting, restricted stock with 5-10 year holding requirements) aligns management's wealth with shareholder wealth at the horizon relevant to durable value creation. Compensation heavily weighted toward annual cash bonuses aligns management's wealth with short-term earnings, which can be manipulated.
The Berkshire model as agency problem solution. Munger cited Berkshire's unique structure — decentralized operating subsidiaries, extremely long holding periods, minimal management turnover, and Buffett and Munger's own large equity ownership — as an unusually effective mitigation of the agency problem. Management of Berkshire subsidiaries has high autonomy (which attracts quality people) combined with clear accountability for long-term results. The owners (Buffett and Munger) are also the managers of capital allocation, eliminating the shareholder-management agency problem at the holding company level.
Auditor independence. Munger was consistently critical of the accounting profession's agency problem: audit firms are paid by the companies they audit, and the same firms that provide audit services frequently also provide lucrative consulting services to the same clients. This creates an incentive to validate management's accounting choices rather than challenge them. The Enron scandal, Munger argued, was partly an agency problem in the audit relationship: Arthur Andersen's financial dependence on Enron's consulting fees made genuine independence structurally impossible.
Common Misconceptions
Misconception 1: Agency problems are solved by good intentions. The agency problem is structural, not motivational. An agent with excellent intentions but misaligned incentives will systematically produce outcomes that serve their interests over the principal's, often without recognizing it. Incentive-Caused Bias operates unconsciously.
Misconception 2: Disclosure solves the agency problem. Disclosure requirements reduce information asymmetry but do not eliminate incentive misalignment. An agent who discloses their conflicts of interest is still operating with those conflicts.
Munger's Own Words
"Talking about what the economists call agency problems, if you're managing your own affairs you're going to be pretty efficient because taking care of your own property. If you're working for somebody else, the truth of the matter is you care more about yourself and your future and your family than you care about the telephone company you're working for. So, capitalism is efficient when the people who are making the decisions, they're doing it about their own property instead of just as hired employees of some say, state owned enterprise." — Charlie Munger, Daily Journal Annual Meeting (2022)
Thought Evolution
Case Study: General Re — The Derivatives Book the Agents Built
Berkshire's 1998 acquisition of General Re is Munger's most personal case study in the agency problem, because he and Buffett were the principals who paid for it. General Re was by reputation one of the most conservative institutions in insurance. What Berkshire discovered after the acquisition was a derivatives operation — Gen Re Securities — whose book of long-dated, illiquid contracts had been assembled over years by managers whose incentives ran toward complexity, volume, and reported revenue rather than toward transparent, low-risk profit.
The structural reading, which is Munger's: nobody in the derivatives unit set out to damage shareholders. The traders and managers were compensated on metrics that rewarded writing business and marking positions that would not be settled for decades — a design in which optimistic assumptions accrue to current bonuses while the verification arrives on someone else's watch. The agents were behaving exactly as the incentive structure instructed; the principals, inside and outside the company, could not see the exposure because the agents controlled the information about it. That is the textbook agency problem — information asymmetry plus incentive misalignment — operating inside an institution everyone believed was proof against it.
The cleanup took years. Buffett's 2002 shareholder letter reported that the derivatives book was being unwound at real cost, and used the episode to issue his famous warning about derivatives as financial weapons of mass destruction. Munger's summary lesson was the sharper one for governance: when an agent's compensation depends on numbers that only the agent can measure, the numbers will drift toward the compensation. The board's duty is not to find honest agents and trust them; it is to refuse structures in which even honest agents are paid to be optimistic.
Legacy & Influence
The agency problem is the concept through which Munger's institutional pessimism becomes practical. Academically, Jensen and Meckling formalized it in 1976 and a generation of corporate-finance research followed; Munger's contribution was to insist that the problem is not a technical footnote about monitoring costs but the central, permanent defect of all delegated judgment — in corporations, in finance, in professions, and in government. His extension of the framework beyond the shareholder-manager pair to auditors, analysts, consultants, lawyers, and physicians remains the broadest applied version of agency theory in circulation.
The concept's fingerprints are on most post-crisis governance reform. Deferred and clawback-able compensation, say-on-pay votes, restrictions on auditors selling consulting to audit clients, fee-only financial advice, and the fiduciary-standard debate are all, in form, attempts to close the specific divergences Munger spent decades cataloguing. His verdict on their sufficiency was characteristically bleak: structural fixes help at the margin, but the problem is regenerated wherever one party controls information and another supplies the money — which is to say, everywhere.
For investors, the agency problem's legacy is a screening discipline now standard in the quality-investing tradition: prefer owner-operators, distrust serial acquirers, read compensation plans before annual reports, and treat intermediary enthusiasm as a cost, not a signal. Berkshire itself stands as the doctrine's proof of concept — a structure deliberately engineered so that the people allocating capital own the consequences of the allocation. The seamless web of deserved trust, Munger argued, is not a moral luxury; it is the only known design in which the agency problem shrinks to manageable size.
Related Concepts
Mentioned In
Source: Poor Charlie's Almanack, The Wit and Wisdom of Charles T. Munger; Jensen & Meckling (1976)