Game Theory · GT-26
Whenever one person acts on another's behalf, their interests are never perfectly aligned — and the gap is where value leaks.
A conflict that arises whenever a principal delegates decision-making authority to an agent whose interests aren't perfectly aligned with the principal's own, compounded by the principal's inability to perfectly observe or verify the agent's effort and choices — creating room for the agent to act partly in their own interest at the principal's expense.
Formal economic analysis developed by Michael Jensen and William Meckling in their hugely influential 1976 paper 'Theory of the Firm,' and independently by Stephen Ross's contemporaneous 1973 work on the economics of agency.
The Mechanism
Where principal and agent interests diverge — and what closes the gap
The distance between the top and bottom rungs is the entire agency problem — every real delegated relationship sits somewhere on this ladder, and moving up it (via better-designed incentives or monitoring) is the whole practical task of agency-cost management.
01 · AGENCY COSTS ARE THE MEASURABLE PRICE OF DELEGATION
Jensen and Meckling's core contribution
Jensen and Meckling formalized 'agency costs' as the sum of monitoring costs the principal incurs, bonding costs the agent incurs to credibly commit to good behavior, and the residual loss that remains even after both — reframing corporate governance, executive compensation, and much of modern finance around minimizing this specific quantity.
02 · IT'S EVERYWHERE ONCE YOU LOOK: MANAGERS, LAWYERS, DOCTORS, POLITICIANS
Any delegated-authority relationship qualifies
Shareholders and corporate managers, clients and lawyers, patients and doctors, voters and elected officials, and homeowners and real-estate agents are all textbook principal-agent relationships — in each, the agent has both better information and some latitude to act in ways that serve their own interest more than the principal's.
03 · THE PROBLEM IS STRUCTURALLY UNSOLVABLE, ONLY MANAGEABLE
There's no arrangement that fully eliminates it
Perfect alignment would require either perfect monitoring (usually prohibitively costly) or making the agent the full residual claimant on every outcome (which usually isn't feasible or desirable) — real institutions instead aim to reduce agency costs to an acceptable level, not eliminate them, since full elimination isn't actually achievable in most real delegated relationships.
Where It Fails / Inversion
Where it fails / inversion
Over-investing in monitoring and incentive-alignment mechanisms can itself destroy more value than the agency problem it's meant to solve — excessive performance-metric gaming, short-termism induced by aggressive incentive pay, or trust-destroying surveillance can all leave the principal worse off than a lighter-touch arrangement built on reputation and repeated interaction would have.
How To Use It
Worked example · executive stock options as an agency-cost fix
Public company shareholders (principals) can't directly observe or control every decision their CEO (agent) makes, creating room for the CEO to prioritize personal comfort, empire-building, or short-term optics over long-run shareholder value. Tying a meaningful share of CEO compensation to stock performance (equity grants, options) is a direct agency-cost fix — it converts some of the agent's payoff into the same currency as the principal's, narrowing (though never fully closing) the gap between what's good for the CEO and what's good for shareholders.
How to use it
Whenever you delegate an important decision to someone else — hiring a contractor, appointing a manager, engaging an advisor — explicitly map where their incentives diverge from yours before assuming good intentions will bridge the gap. The fix is rarely 'trust more'; it's usually a specific incentive-alignment mechanism (equity, performance pay, transparent reporting) sized to the actual value at stake.
See Also