> For the complete documentation index, see [llms.txt](https://laurence-wilse-samson.gitbook.io/textbooks/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://laurence-wilse-samson.gitbook.io/textbooks/financial-economics-claims-prices-holders/part-v-firms-as-issuers-of-claims/chapter_24_governance.md).

# Chapter 24: Financial Contracting and Corporate Governance

*Part V: Firms as Issuers of Claims — Financial Economics: Claims, Prices, and Holders*

***

## Opening Episode: Three Seats at Exxon

On 26 May 2021, at Exxon Mobil's annual meeting, shareholders elected directors nominated by an investment firm that had existed for five months. Engine No. 1 had launched in December 2020 and had built a position in Exxon worth roughly fifty million dollars — against a company then capitalized at something close to a quarter of a trillion. That is about two hundredths of one percent of the shares: one share in five thousand. Two of its four nominees were declared elected on the day; a third emerged the following week, when Exxon filed the estimated vote totals and Alexander Karsner was shown ahead of the last company nominee by some fourteen million votes out of more than a billion cast for each. The fourth, Anders Runevad, finished last of the sixteen and was not elected.

Nothing about the campaign's substance was exotic. Engine No. 1 argued that Exxon had spent a decade committing capital to projects that had not earned their cost of capital, that its returns had trailed its closest peers, that it had reported a loss for 2020 and had been removed from the Dow Jones Industrial Average in August of that year after more than nine decades in the index, and that its board contained no director with operating experience in energy outside oil and gas. The remedy proposed was four directors and a capital-allocation review. It was, in other words, an ordinary argument about investment discipline of exactly the kind Chapter 22 §22.1 formalizes.

What was extraordinary was the arithmetic of who decided. A holder of one share in five thousand cannot elect anybody. The three largest asset managers held, between them, something closer to one share in five, and they vote very nearly all of what they hold. BlackRock published a vote bulletin explaining that it had supported three of the four dissident nominees; State Street and Vanguard each disclosed support for part of the slate; several of the largest public pension funds announced their support in advance; and both major proxy advisers recommended for some of the dissidents. Engine No. 1 did not win a proxy contest against Exxon's shareholder base. It persuaded a small number of stewardship teams, and the shareholder base followed.

Hold that structure in mind, because it is the chapter's subject in miniature. Chapter 21 §21.4 established that a security is an allocation of cash-flow rights *and* control rights across states of the world. Governance is what the holders of residual control do with it — and the questions worth asking are which contracts assign that control, in which states, and to whom. This chapter takes those in order: debt contracts, which move control on a trigger; the agency problem that makes control worth having; the internal apparatus shareholders use to exercise it; and, last, what happens to all of it when the residual control rights of most of corporate America are concentrated in three ballots.

***

## 24.1 Covenants: Control That Moves on a Trigger

A loan agreement is the cleanest instance in finance of Chapter 21 §21.4's proposition. The lender's cash-flow right is fixed and senior. Its control right is zero while the borrower performs and substantial the moment the borrower does not. **Covenants** are the schedule that specifies when the switch flips.

The taxonomy that matters is not affirmative versus negative but **maintenance** versus **incurrence**.

**Table 24.1: Two kinds of covenant**

|                     | Maintenance covenant                                                                   | Incurrence covenant                                                |
| ------------------- | -------------------------------------------------------------------------------------- | ------------------------------------------------------------------ |
| Tested              | Every quarter, automatically                                                           | Only when the borrower takes a specified action                    |
| Typical test        | Net debt to EBITDA below a stated multiple; EBITDA to interest above a stated multiple | The same ratios, but evaluated *pro forma* for the proposed action |
| Triggered by        | Deterioration in the business                                                          | A new borrowing, an acquisition, a dividend, an asset sale         |
| Control transferred | Continuously contingent on performance                                                 | Contingent on a discretionary act                                  |
| Home                | Bank and pro-rata loans; middle-market direct lending                                  | High-yield bond indentures; cov-lite institutional loans           |

*Source: Author's summary of standard credit-agreement and indenture terms.*

The difference is the whole of the economics. A maintenance covenant is a state-contingent transfer of authority: the borrower's business worsens, a ratio crosses a line, and control passes to the lender *without the borrower having done anything and without any payment having been missed*. An incurrence covenant transfers nothing on deterioration alone. It restrains a set of actions and is silent about everything else.

What the lender receives on a maintenance breach is not repayment. It is a bargaining position. The technical default gives the lender the right to accelerate, which it almost never exercises, because accelerating a loan to a firm that is still generating cash converts a performing asset into a bankruptcy claim. What the lender does instead is sell a waiver. The price is a fee, a higher margin, tighter future thresholds, additional collateral, and — the part that matters for this chapter — restrictions on what the borrower may do with its money: caps on capital expenditure, bans on acquisitions and dividends, mandatory asset sales, sometimes a required change of management or of financial officer.

The empirical literature on this is among the most convincing in corporate finance, and its design is the reason. Firms just below a covenant threshold and firms just above it are similar in everything except that one has tripped and one has not, so the discontinuity identifies the effect of the control transfer rather than of the underlying distress. The results, associated with Chava and Roberts and with Nini, Smith and Sufi, are consistent across specifications: capital expenditure falls sharply and immediately after a violation, leverage falls, payouts stop, and chief executive turnover rises — with the response concentrated where the lender's monitoring incentive is strongest, and larger where information problems between borrower and lender are worse. Figure 24.1 draws the two regimes side by side. Read against §24.2, this is striking: the discipline conventionally attributed to shareholders is being applied by a creditor with no cash-flow interest in the upside at all, acting on a right that vested automatically when an accounting ratio crossed a line.

![Figure 24.1: Control follows performance](https://846781005-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3EupdX99vVBoNySDtmxb%2Fuploads%2Fgit-blob-dd686230d63e17fc70b4c888f1fadddeedd1055a%2Ffig_24_01_control_follows_performance.png?alt=media)

**Figure 24.1: Control follows performance.** The same firm, deteriorating from left to right, under two documentation regimes. Under a maintenance covenant the control right moves twice: once when a quarterly accounting test is failed — with no payment missed and no act by the borrower — and again at a payment default, when it moves to the court. Under incurrence covenants only, the first transfer does not happen: authority stays with the borrower through the whole middle of the axis and moves only at an actual default. The band between the two rules is what the cov-lite migration of §24.1 gave back to borrowers, and it is precisely the region in which the empirical literature finds the transfer doing its work — capital expenditure cut, payouts stopped, leverage reduced, chief executives replaced. Thresholds are illustrative levels of the kind Table 24.1 describes, not any particular credit agreement's, and the horizontal axis is an ordering of states rather than a measured distribution. *Source: Author's construction from Table 24.1.*

### The cov-lite migration

Which makes the last two decades' change in loan documentation a repricing of that control right rather than a technicality. **Cov-lite** loans carry incurrence covenants only. They were a small share of the US institutional leveraged loan market before 2007, vanished in the crisis, and then became the standard.

Two supply-side facts explain the migration, and both are holder facts. The buyer of an institutional term loan is no longer a bank that will sit in a waiver negotiation but a collateralized loan obligation, a loan mutual fund, or a separate account — holders who want a floating-rate asset and have neither the mandate nor the staff to run a workout, which is Chapter 19 §19.4's account of nonbank inventory — the securitization vehicle that can buy only what its own tranche buyers will fund — applied to a credit instrument. And the borrowers themselves increasingly have the alternative of a private credit fund (Chapter 18 §18.5) that will write a bilateral agreement on its own terms, so documentation competes. The consequence for pricing is Chapter 10's: a loan whose control right no longer vests on deterioration is a different security, its recovery distribution is worse, and the spread should say so.

What the covenant actually did was force an early conversation. Removing it does not remove the distress; it removes the meeting. The evidence from the 2020 and 2022-2023 cycles is that cov-lite structures postpone the moment of contact until liquidity is nearly exhausted, at which point the borrower's remaining flexibility — moving collateral into an unrestricted subsidiary, raising priming debt inside the existing agreement's baskets — is exercised against the lenders rather than negotiated with them. Control did not disappear. It stayed with the borrower for longer, and the fights moved from the covenant to the definitions.

### Bankruptcy: the last transfer

The final step in the sequence is the one Chapter 10 prices. In a US reorganization — Chapter 11 of the Bankruptcy Code, not of this book — the automatic stay suspends creditors' individual remedies, the debtor's management ordinarily stays in place with an exclusive period to propose a plan, and claims vote by class. Formally this is absolute priority: senior creditors are paid in full before junior ones receive anything, and equity is last. In practice the plan is a negotiation conducted in the shadow of that rule, and deviations from absolute priority in favor of junior claims and equity are common — bought, in effect, with the holdup value of the junior classes' consent and their ability to delay. Chapter 10 §10.1's recovery rate is the empirical residue of that bargaining, which is why it is a distribution rather than a number, and why the identity of the holders on each rung is part of what determines it.

***

## 24.2 The Agency Costs of Equity

Chapter 21 §21.5 named the apparatus and deferred it here. Jensen and Meckling's construction is the simplest device in the subject and the most useful.

Begin with a firm entirely owned by the person who runs it. She chooses how much of the firm's resources to divert into consumption on the job — an aircraft, an office, a headquarters building, a favored division kept alive, a diversifying acquisition that makes her position safer. Write that spending $$F$$, in millions of dollars a year, and let the private benefit it yields her be

$$
B(F) = bF - \tfrac{1}{2}cF^2
$$

with $$b = 1.4$$ and $$c = 0.02$$. A sole owner spends until the marginal private benefit equals the marginal cost, which she bears in full: $$b - cF = 1$$, so $$F = 20$$. That is the efficient level, and it is not zero — a corner office is a real benefit and the firm should buy some. Total surplus from perquisite spending, $$-F + B(F)$$, is then four.

Now let her sell a fraction $$1-\alpha$$ of the equity to outside investors, keeping $$\alpha$$. She still receives the whole private benefit but now bears only $$\alpha$$ of the cost, so she spends until $$b - cF = \alpha$$:

$$
F^{\ast}(\alpha) = \frac{b-\alpha}{c}
$$

At $$\alpha = 0.4$$ that is fifty, two and a half times the efficient level. The firm's market value, gross value less perquisite spending, falls from four hundred and eighty to four hundred and fifty.

Here is the step that makes the model an argument rather than an observation. Outside investors are not fooled. They can see $$\alpha$$, they know $$F^{\ast}(\alpha)$$, and they will pay only $$(1-\alpha) \times 450 = 270$$. The owner-manager therefore ends up with two hundred and seventy in cash, one hundred and eighty in retained shares, and forty-five of private benefit: four hundred and ninety-five in total, against five hundred and four had she kept the whole firm. **The nine million dollars of value destroyed is borne entirely by the person doing the destroying, at the moment she sells.** Agency costs are not a transfer from outside shareholders to insiders. They are a deadweight loss capitalized into the price at which the entrepreneur can sell, which is why she has every incentive to bond herself against them — with a board, with an auditor, with a covenant, with a compensation contract — up to the point where the last dollar of bonding buys a dollar of value. Jensen and Meckling's accounting is exactly this: agency costs are monitoring expenditures by the principal, plus bonding expenditures by the agent, plus the residual loss that survives both.

The residual loss rises as ownership is diluted: sixteen at $$\alpha = 0.2$$, about twenty-three at $$\alpha = 0.05$$. Which states the problem of the public corporation in one line. Diffuse ownership is efficient for risk-sharing and for liquidity, and it is precisely the ownership structure in which the manager's stake is too small to align her.

**Free cash flow.** Jensen's 1986 refinement identifies where the problem bites hardest. The dangerous firm is not the one losing money; it is the one generating cash in excess of the projects that clear its cost of capital. A manager who returns that cash shrinks the firm she runs, forgoes the growth on which her pay and status depend, and must come back to the market — and be examined — the next time she needs capital. A manager who retains it need not. So free cash flow gets spent: on acquisitions of the kind Chapter 22 §22.5 shows the market marking down at announcement, on diversification, on capacity in mature industries. Both remedies are claim-design remedies, which is why the argument belongs to this part of the book. Payout commits the cash to holders, and Chapter 23 §23.5 shows why repurchases and dividends are not equivalent commitments. Debt commits it harder, because a missed coupon is a default and a suspended dividend is a press release — which is the sense in which leverage is a bonding technology and not only a tax shield, and the sense in which §24.1's covenant is the fine print of the bond.

![Figure 24.2: Risk shifting](https://846781005-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3EupdX99vVBoNySDtmxb%2Fuploads%2Fgit-blob-ed25fba42051fae6fd2e3995c62cc02f93db9dba%2Ffig_24_02_risk_shifting.png?alt=media)

**Figure 24.2: Risk shifting.** Why the control right of Figure 24.1 is worth having. Panel (a): with debt outstanding, equity is a call on the firm struck at the face value and the debt is what is left — Chapter 8 §8.1's payoff diagram, read as a capital structure, and Chapter 10 §10.2's structural model in one line. Panel (b): two projects with the *same* expected firm value of 100 and different variance. Because a call is worth more the more volatile its underlying, the switch from the safer project to the riskier one raises the equity from 9.95 to 21.67 and lowers the debt from 90.05 to 78.33. Nothing is created: 11.72 moves across the boundary between the claims, and the firm is worth exactly what it was. This is the conflict a covenant is written against, and it is the reason a lender with no interest in the upside nonetheless wants a say in what the borrower does. Valued without discounting, so the two claims sum to the firm's value exactly; the volatility parameters are chosen to make the transfer legible and are not estimates. *Source: Author's calculation. Equity valued as a call on the firm struck at the face value of the debt.*

**Whose risk preferences?** The third conflict is about the variance rather than the mean, and Figure 24.2 prices the version of it that runs between the claims rather than between the people. A diversified shareholder cares about a project's contribution to systematic risk and is indifferent to its idiosyncratic component, which is Chapter 4's result. A chief executive has her human capital, her reputation, and typically a large undiversified equity stake in one firm, and she cannot hedge any of it. She is therefore too conservative about idiosyncratic risk in ways the shareholder would not sanction — declining a positive-net-present-value project that might fail visibly, or diversifying the firm to smooth its earnings when shareholders could have diversified more cheaply themselves. The apparatus built to correct that, in §24.3, corrects it with instruments whose payoff is convex in the stock price, and convexity has the opposite failure mode. A manager holding many out-of-the-money options prefers a riskier firm than shareholders do. There is no contract that gets the mean and the variance right at once with a single instrument, and the practice of compensation design is the search for a mix that is wrong in both directions by a tolerable amount.

***

## 24.3 The Governance Toolkit

Four internal instruments, in the order a firm encounters them.

### Boards

The board is the institution to which the shareholders' residual control right is delegated. Since the listing-standard reforms that followed Sarbanes-Oxley in 2002, majority-independent boards and fully independent audit, compensation, and nominating committees have been effectively universal among large US listed firms, so independence has almost no cross-sectional variation left to explain anything with.

The evidence on independence is worth stating precisely, because it is routinely overstated in both directions. Studies relating the *fraction* of independent directors to firm performance — Bhagat and Black's is the standard reference — find essentially nothing on average. That is the right finding and it is not an argument for insider boards, for two reasons. First, board composition is chosen, and chosen partly in response to past performance, so the cross-sectional correlation is not a treatment effect; Hermalin and Weisbach's bargaining model of board formation is the canonical statement of why. Second, running a company is a continuous activity that independent directors, who meet several times a year and know less than management, are badly placed to improve. What independent boards demonstrably do better is the discrete, adversarial, verifiable decision: replacing a chief executive after poor performance, evaluating a takeover bid, responding to an accounting failure. Independence is insurance against a small number of large events, and average performance regressions are not the place to look for it.

Board *structure* has clearer evidence than board composition. A staggered board — directors elected in classes, so that only a third of the seats are contested each year — is associated with lower firm value in the entrenchment literature, and it is the provision that makes a poison pill decisive against a hostile bidder. Chapter 22 §22.6 develops that, along with the governance-index results of Gompers, Ishii and Metrick and the honest footnote about their out-of-sample performance, and this chapter does not restate it. The mechanism worth carrying forward is that a staggered board raises the cost of the one instrument shareholders hold unconditionally, which is the vote.

### Executive compensation

The measure is **pay-performance sensitivity**: the change in the manager's wealth per one thousand dollars of change in shareholder wealth. Jensen and Murphy's 1990 estimate for the median large-firm chief executive was about three dollars and twenty-five cents per thousand, of which only about thirty cents came from salary and bonus — a number they presented as evidence that chief executives were paid like bureaucrats rather than like owners, and that the resulting incentive was too weak to matter.

The response was the options era. Grants of at-the-money stock options went from a supplement to the largest component of large-cap chief-executive pay through the 1990s, and measured sensitivity rose by roughly an order of magnitude. Three features of the instrument drove its adoption and three distortions followed from the same features. Options were, until the accounting change that took effect in the mid-2000s, capable of being granted without a charge to reported earnings, which made high-powered incentives free in the metric the board was itself judged by. Their payoff is convex, which magnifies sensitivity per dollar of grant-date cost — and, per §24.2, encourages exactly the risk-taking a diversified shareholder may or may not want. And their value depends on the grant-date price, which turned the date itself into a decision: the backdating literature, beginning with Lie's observation that grant dates fell too often at local price minima to be chance, established that a nontrivial number of firms had been choosing dates retrospectively. Two further distortions are structural rather than abusive. Standard options are not indexed to the market or the industry, so a manager is paid for luck — Bertrand and Mullainathan's finding that oil-company executives are rewarded for the oil price is the cleanest demonstration — and a fall in the stock destroys the incentive entirely, which is what makes repricing after a decline both common and self-defeating. The post-2006 drift toward restricted stock and performance-vesting shares is a partial answer to the convexity and luck problems and no answer at all to the last one. Say-on-pay, the advisory shareholder vote mandated in the United States by the Dodd-Frank Act of 2010, adds a low-cost annual signal that boards attend to more than its non-binding status implies.

### Blockholders: voice and exit

A shareholder large enough to internalize a meaningful share of the gain from monitoring has an incentive that a diffuse holder lacks, which is Shleifer and Vishny's argument for concentrated ownership and the reason blockholders appear in every account of what makes governance work. A block confers two distinct instruments.

**Voice** is the direct one: engagement, board representation, a public campaign, a vote against the slate. It is costly, and the cost is borne by the blockholder while the benefit accrues to every holder in proportion to holdings — the free-rider problem that makes voice systematically undersupplied.

**Exit** is the indirect one and is more interesting theoretically. A blockholder who sells drives the price down; if the manager's wealth is tied to the price, the credible threat to sell disciplines her *without any exercise of control at all*. The mechanism requires that informed selling actually move the price, which is Chapter 7's account of how private information enters prices; the more informative the price, the sharper the threat. Two implications follow and both are testable. Governance by exit is strongest where the blockholder is informed and the stock is liquid enough that trading conveys the information without destroying the position — so liquidity is a governance input, not merely a trading cost. And an owner who *cannot* sell has no exit threat, which is the structural fact §24.4 is about.

### Hedge-fund activism

The modern instrument combines the two. An activist fund buys a stake — commonly between one and ten percent, disclosed on a Schedule 13D once it crosses five percent in the United States — and campaigns publicly for a change in payout, capital structure, strategy, or management.

The evidence, in the form Brav, Jiang, Partnoy and Thomas established, is that target stock prices rise on the order of seven percent around the disclosure, that the gain is not reversed over the following year, and that targets subsequently increase payouts, raise leverage, sell divisions, replace chief executives more often, and show improved operating profitability. The natural objection is that the announcement return simply prices the probability of a takeover and the operating "improvement" is asset stripping deferred. Bebchuk, Brav and Jiang tested the myopia claim directly by following targets for five years after an intervention and found no evidence of the predicted decline in operating performance — the sharpest available answer to the short-termism critique.

It is not a settled question, and the chapter should not pretend it is. The critique advanced by Coffee and Palia and others is partly about identification — activists select targets whose performance was already mean-reverting, and the comparison group is not obvious — and partly about scope, since portfolio-level averages can be consistent with severe harm in particular cases, and since the studies measure accounting profitability rather than the investment in intangibles that a myopia story predicts would be cut. Note also what the practice has become: most campaigns at large firms now end in a negotiated settlement granting board seats without a vote, which means the measured "activism event" is increasingly a bargaining outcome whose threat point is the vote rather than the vote itself.

**Table 24.3: The internal toolkit and what the evidence supports**

| Instrument          | What it transfers                                     | What the evidence supports                                                                                                                   |
| ------------------- | ----------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
| Independent board   | Monitoring and dismissal authority                    | Little on average performance; real effects on chief-executive turnover, bid evaluation, and crisis response                                 |
| Staggered board     | Removes annual accountability for two thirds of seats | Associated with lower firm value; decisive as a takeover defense (Ch 22 §22.6)                                                               |
| Equity-based pay    | Aligns the manager's wealth with the price            | Sensitivity rose by roughly an order of magnitude after 1990; convexity, luck, and grant-date discretion are the documented costs            |
| Blockholder voice   | Direct engagement and board seats                     | Undersupplied because the monitoring cost is private and the gain is shared                                                                  |
| Blockholder exit    | Discipline through the price                          | Requires informative prices and liquidity; unavailable to an index holder                                                                    |
| Hedge-fund activism | Both, plus a credible proxy threat                    | Positive and persistent announcement returns, operating improvements; the myopia critique is not supported over five years but is not closed |

*Source: Author's summary of the literature cited in §24.3.*

And the external instrument: the takeover market disciplines every management team whether or not a bid arrives, and Chapter 22 §22.6 develops it — the 1980s evidence, the poison pill and the staggered board that blunted it, and the four reasons hostile activity has never returned to its earlier levels. What matters here is the sequencing: the internal toolkit of this section grew into the space the defenses created, and activism is the instrument that most resembles a takeover threat without a takeover.

***

## 24.4 Who Holds the Vote, and What Their Constraints Do to Its Price

Every instrument in §24.3 assumes a holder who has decided that this company is worth thinking about. Chapter 17 §17.7 describes a holder who has decided the opposite by construction, and who nonetheless owns roughly a fifth of it. This section is the governance reading of that fact.

Start from Chapter 2's Table 2.5. The funds column holds roughly fifteen trillion dollars of a sixty-trillion-dollar US equity market, and the three largest managers account for a large share of it. Their holdings are formal residual control rights: votes for directors, on mergers, on say-on-pay, on shareholder proposals. And index funds vote very nearly everything they hold, while retail holders mostly do not vote at all — so the Big Three's share of *votes cast* is materially larger than their share of *shares outstanding*. Problem 4 works the arithmetic; with plausible turnout assumptions a twenty percent ownership stake becomes something above a quarter of the ballots, and the number of parties a challenger must persuade to reach a majority collapses.

Chapter 17 §17.7 frames the stewardship problem as one of incentives: an index manager competing on fee captures only its ownership fraction of any governance gain, so serious company-by-company engagement is privately unprofitable, and each of the Big Three runs a stewardship team of a few dozen people against tens of thousands of meetings a year. The governance consequence is not that they fail to vote. It is that they must vote by **policy** — a published, general, ex ante rule, applied at scale, and heavily informed by two proxy advisers whose own concentration is a separate concern.

Governing by policy has a property that the incentive framing does not immediately reveal: it makes control *addressable*. A published policy is a public document with an author, and persuading its author is a fixed cost that does not scale with the size of the target. This is what the opening episode is about. Engine No. 1 put the cost of its Exxon campaign at approximately thirty million dollars in its own definitive proxy statement — an absurd sum relative to its own stake and a trivial one relative to the outcome — rational only because the persuadable population was small, identifiable, and had committed in advance to criteria against which a case could be made. The capacity constraint that makes index funds weak monitors of any particular firm is the same constraint that makes them a concentrated and reachable audience. A bottleneck is a weakness from one side and a lever from the other.

Three developments follow, and each is unresolved.

**Pass-through voting.** Since 2022 the largest managers have offered some fund investors the ability to direct how their shares are voted — BlackRock's Voting Choice, Vanguard's pilot programs, State Street's equivalent — initially to institutional clients and then, in limited form, to retail holders in some funds. The stated rationale is precisely the concentration problem. Whether it dissolves the bottleneck depends on a question Chapter 14 has already answered elsewhere: what the default is. If the default remains the manager's own policy and uptake is low, the arrangement redistributes accountability without redistributing votes. Uptake so far has been modest, and it has been concentrated among institutional clients that had a policy of their own already.

**Common ownership.** If diversified owners' preferences transmit into product-market conduct, they transmit through the same ballots and the same policies. Chapter 17 §17.7 states the case and the objections; the honest summary is that the correlation between common ownership and market power is real in some datasets and fragile in others, and the mechanism is unestablished. The governance point to add here is narrow and worth keeping separate from the antitrust claim: the mechanisms §24.3 documents — board seats, engagement, compensation design — are precisely the ones an index manager does *not* use, which is why the common-ownership literature has had to propose channels (industry-relative compensation, managerial passivity) that sit outside the toolkit.

**Mandate-driven voting.** Chapter 16 §16.6 treats mandates as holder constraints that move prices. They also move votes, and the voting record is the cleaner demonstration. Support from the largest managers for environmental and social shareholder proposals rose sharply through the 2021 season — the season of the opening episode — and fell back substantially from 2022, as the managers narrowed the range of proposals they judged consistent with a financial mandate under changed political and legal pressure. The shares did not change. The companies did not change. The mandate did. A control right whose exercise is a function of the holder's mandate rather than of the firm's condition is a control right that belongs, analytically, in Part IV.

Which resolves the episode. Engine No. 1 did not represent shareholders against management; it represented an argument to a small set of intermediaries who hold other people's residual control rights under a policy. That is what corporate control looks like when ownership has the composition Chapter 2 measures. Control rights follow the holdings table, so Part IV's ecology of holders is also the governance map — and a change in that ecology is a change in governance, whether or not anyone intended it as one. The passive revolution was a fee story, a distribution story, and an asset-pricing story before it was a governance story. It became a governance story anyway, because the claims it accumulated came with votes attached.

In Chapter 1 §1.2's terms, the price of a vote moves as a constraint binding rather than as news: the shares did not change and the companies did not change, the mandate did.

***

> **Box 24.1 — Case: securities litigation as a governance backstop**
>
> A shareholder who buys at a price inflated by a false statement has, in the United States, a private right of action under Rule 10b-5. What makes it a systemic instrument rather than a private remedy is the class action, and what makes the class action possible is *Basic v. Levinson* (1988), which held that in an efficient market a purchaser may be presumed to have relied on the integrity of the market price rather than on the statement itself. That presumption — **fraud on the market** — is an application of Chapter 7's semi-strong efficiency doctrine imported into evidence law, and it converts individually unprovable reliance into a certifiable common question. The Private Securities Litigation Reform Act of 1995 tightened pleading standards and installed a lead-plaintiff mechanism intended to put large institutions in charge of the litigation; *Halliburton II* (2014) preserved *Basic* while letting defendants rebut the presumption with evidence of no price impact at the certification stage — which has had the effect of moving event-study evidence, of exactly the kind Chapter 7 §7.3 describes, to the front of the case. Filings run to a few hundred a year in the federal courts and aggregate settlements to a few billion dollars.
>
> The critique is that the remedy is close to circular. The defendant is usually the corporation, which pays out of the assets of its *current* shareholders, to a class of *former* purchasers, most of whom are diversified and therefore sit on both sides of the transaction across their portfolios — so the expected transfer nets toward zero while the attorneys' fees and directors-and-officers premiums do not. Individual officers and directors almost never pay from their own pockets. If the case for the mechanism is deterrence rather than compensation, then it must run through channels other than the damages: reputational and career costs to the individuals involved, the disclosure practices firms adopt in anticipation, and the screening the plaintiff bar performs. Whether the deterrence is worth the cost, how the pleading and certification standards should be set, and what a non-circular remedy would look like are questions of legal design rather than of asset pricing, and they are developed in the prospective *Law and Economics* volume in this series. What belongs here is the placement: litigation is the state-contingent control right that vests in a *court*, and it is the only instrument in this chapter that a shareholder can invoke without owning enough to matter.

> **Box 24.2 ★ — Rothschild-Stiglitz: screening and the separating equilibrium**
>
> Every contract in this chapter is written by a party who knows less than the other. The canonical analysis of what that does to a competitive market is Rothschild and Stiglitz (1976), stated here in the insurance setting Chapter 16 §16.2 uses, because the insurance case is where the geometry is clearest.
>
> A consumer has wealth $$W = 100$$ and faces a loss $$L = 50$$. Two types exist: a high-risk type with accident probability $$\pi\_H = 0.4$$ and a low-risk type with $$\pi\_L = 0.2$$. The consumer knows her type; the insurer does not. Preferences are von Neumann-Morgenstern with $$u(c) = \sqrt{c}$$ over consumption $$c\_{\mathrm{N}}$$ in the no-accident state and $$c\_{\mathrm{A}}$$ in the accident state. Without insurance every consumer sits at $$(100, 50)$$.
>
> A contract is a pair $$(c\_{\mathrm{N}}, c\_{\mathrm{A}})$$, and a contract breaks even against a type with accident probability $$\pi$$ when the premium collected in the good state equals the expected net indemnity paid in the bad one:
>
> $$
> (1-\pi)(W - c\_{\mathrm{N}}) = \pi\bigl(c\_{\mathrm{A}} - (W - L)\bigr)
> $$
>
> a line through the no-insurance point with slope $$-(1-\pi)/\pi$$. Two facts do all the work. Risk-averse consumers fully insure when offered their own fair odds, so the type-$$\pi$$ optimum is $$c\_{\mathrm{N}} = c\_{\mathrm{A}} = W - \pi L$$. And the high-risk type's indifference curves are *flatter* at every point than the low-risk type's, because she weights the accident state more heavily — the single-crossing property that makes screening possible.
>
> No **pooling** contract can survive competition. If one contract is offered to both types at the population-average odds, an entrant can offer a contract with slightly less coverage at slightly better odds, which the low risks accept and the high risks decline, leaving the incumbent with the high risks and a loss. Any candidate equilibrium must therefore **separate**. The separating pair is pinned down by two constraints: high risks get full insurance at their own fair odds, $$(80, 80)$$, since nothing they would accept can be denied them; and low risks get the *most* coverage at their own fair odds that leaves the high-risk type just unwilling to switch. Solving $$0.6\sqrt{c\_{\mathrm{N}}} + 0.4\sqrt{c\_{\mathrm{A}}} = \sqrt{80}$$ along the low-risk fair-odds line gives $$(98.55, 55.80)$$: a premium of $1.45 buying a gross indemnity of $7.25, which is about fourteen and a half percent of the loss.
>
> Read the result. The low-risk type is rationed, and severely. She is offered a sliver of coverage at actuarially fair terms and would happily buy far more; no insurer will sell it to her, because any contract she would prefer the high risks would take too. Her certainty equivalent rises from 88.63 uninsured to 89.03, against the 90 that full insurance at her own odds would deliver — under thirty percent of the available gain. The distortion falls entirely on the party with nothing to hide, and it takes the form of a *quantity* restriction rather than a price. That is the general lesson, and it recurs whenever a contract is designed against a privately informed counterparty: an insurer's deductible, a lender's collateral requirement, a venture capitalist's milestone-contingent tranche, and the retention a securitization sponsor keeps in its own deal (Chapter 10 §10.8) are all the same instrument.
>
> Two warnings complete the box. First, the separating pair may not be an equilibrium at all: if high risks are a small enough share of the population, pooled odds sit close to low-risk odds, and the pooling contract low risks would prefer defeats the separating pair — while pooling is itself defeated by cream-skimming, so no equilibrium in pure strategies exists. In the numbers above that happens when high risks are less than about ten percent of the population. Second, the model's welfare reading is not that information is bad but that competitive screening cannot reach the full-information allocation; whether a mandate improves on it depends on who is being cross-subsidized, which is the argument every insurance market has.

***

## Close: Two Halves of One Contract

Chapter 23 asks how a firm's cash flows are divided into claims and paid out. This chapter has asked who decides — and the answer has been, at every step, that the same contract does both jobs. A covenant is a term in a loan agreement that prices like a credit spread and functions like a transfer of authority. An option grant is compensation expense and a change in the manager's risk preference. A share is a residual cash-flow claim and a ballot. Chapter 21 §21.4 said that a security is an allocation of cash-flow rights and control rights across states of the world; Chapters 23 and 24 are that sentence read twice, once down each column.

The reason to keep both columns in view is that they can move apart. Cov-lite migration changed the control content of leveraged loans while leaving their cash-flow content nominally intact, and the pricing question is whether the spread noticed. The passive revolution changed who holds equity's control right without changing the instrument at all, and it did so as a by-product of a decision about fees. In each case a claim was repriced because its control dimension moved, and in each case the holders of Part IV are the reason it moved. That is the same proposition this book has been making from the pricing side since Chapter 3's closing question about whose marginal utility sets prices: the identity and the constraints of the holder are not a detail of implementation. Here they determine not what a claim is worth but what its owner may do.

What remains is the prior question. This chapter and the two before it have assumed a firm with a market value, a board, and a class of public shareholders — an entity that has already chosen among the claims on Chapter 2's map. Most firms have not, and cannot. The overwhelming majority of firms in any economy are financed by a bank, a private lender, a family, or retained earnings, and they will never issue a security that trades. Which firms get access to which claims, on what terms, and at what point in their lives, is Chapter 25's subject, and it is where the design problem of Part V meets the distribution of firms that actually exists.

***

## Elsewhere in the Series

* **Securities litigation, the design of the class action, and the deterrence-versus-compensation question** — the prospective *Law and Economics* volume in this series. Box 24.1 states the mechanism and its critique and stops there; the legal-design argument is developed there, not here.
* **This book.** The proposition this chapter applies — a security is an allocation of cash-flow *and* control rights across states — is Chapter 21 §21.4, with the agency-cost apparatus named and deferred at §21.5. The takeover market as an external governance mechanism, the poison pill, the staggered board, and the governance-index evidence: Chapter 22 §22.6, which this chapter cites and does not restate. Default, recovery, and the pricing of the state in which control transfers: Chapter 10, especially §§10.1-10.2; the retention a sponsor keeps in its own securitization, as a screening device: §10.8. The loan-holder ecology behind cov-lite: Chapter 19 §19.4, for the securitization vehicle and its funding constraint; private credit as its competitor: Chapter 18 §18.5, with the borrower's view at Chapter 25 §25.3. How informed trading gets into prices, which is what makes governance by exit work: Chapter 7 §7.2. The master holdings table from which the votes are read: Chapter 2, Table 2.5. Index-fund stewardship capacity, common ownership, and proxy-adviser concentration: Chapter 17 §17.7. Mandates as holder constraints, and the insurance material Box 24.2 serves: Chapter 16 §§16.2 and 16.6. Delegation and the agency problems of benchmarking, which sit underneath the stewardship problem: Chapter 16 §16.3. Diversification and why idiosyncratic risk is not priced, against which §24.2's managerial risk preferences are measured: Chapter 4. Capital structure and payout as the cash-flow half of the same contract: Chapter 23. Which firms get access to which claims at all: Chapter 25.

***

## Summary

1. **A covenant is a state-contingent transfer of control, not a promise about cash.** Maintenance covenants are tested every quarter and move authority to the lender on deterioration alone; incurrence covenants are tested only when the borrower acts, and transfer nothing when the business simply worsens (Table 24.1).
2. **What a lender gets on a breach is a bargaining position, not repayment.** Acceleration is rarely used; the waiver is sold, and its price includes restrictions on investment, payout, and sometimes management.
3. **The regression-discontinuity evidence around covenant thresholds is the cleanest identification in corporate finance.** Capital expenditure and leverage fall and chief-executive turnover rises just after a violation, with the effect concentrated where lender monitoring incentives are strongest — creditor discipline applied by a party with no interest in the upside.
4. **Cov-lite went from a small share of the US institutional loan market before 2007 to the market standard**, driven by holders who cannot run workouts and by competition from private credit. Removing the covenant does not remove the distress; it postpones the meeting and moves the fight to the definitions.
5. **Bankruptcy is the last transfer.** Absolute priority is the formal rule and the plan is negotiated in its shadow; deviations toward junior classes are common, which is why Chapter 10's recovery rate is a distribution.
6. **Jensen-Meckling's agency cost is a deadweight loss capitalized at the moment of sale.** In §24.2's example, diluting to a forty percent stake raises perquisite spending from twenty to fifty and destroys nine of value — all of it borne by the owner-manager through the price outsiders will pay. That is why she buys monitoring and bonding voluntarily.
7. **Free cash flow is where the problem bites.** Cash in excess of positive-net-present-value projects gets spent rather than returned; payout and debt are the two claim-design commitments that stop it, which is the governance content of Chapter 23's payout policy.
8. **The manager's risk preferences differ from a diversified holder's in both directions.** Undiversified human capital makes her too conservative; convex option payoffs make her too aggressive. No single instrument fixes both.
9. **Board independence has little average effect and real effects on discrete adversarial decisions** — dismissal, bid evaluation, accounting failure. Composition is endogenous, so cross-sectional performance regressions are the wrong test. Staggered boards, by contrast, are associated with lower value (Chapter 22 §22.6).
10. **Pay-performance sensitivity rose by roughly an order of magnitude after Jensen and Murphy's three dollars and twenty-five cents per thousand**, driven by options — whose accounting treatment, convexity, and grant-date discretion produced the era's three documented distortions, and whose lack of indexation pays managers for luck.
11. **Blockholders govern by voice and by exit.** Voice is undersupplied because its cost is private and its benefit shared; exit works through the price and therefore requires informative prices and liquidity — and is unavailable to a holder who cannot sell.
12. **Hedge-fund activism shows positive, non-reversing announcement returns and subsequent operating improvement**, and the five-year evidence does not support the short-termism critique, though identification and scope objections remain live (Table 24.3).
13. **The Big Three's share of votes cast exceeds their share of shares**, because index funds vote nearly everything and retail holders mostly do not. Governing at that scale means governing by published policy — which makes control addressable, and makes a fifty-million-dollar stake sufficient to win three board seats at a quarter-trillion-dollar company.
14. **Control rights follow the holdings table.** Pass-through voting, the common-ownership question, and the rise and retreat of environmental and social voting support are all consequences of who holds the claim rather than of what the firm did. Part IV's ecology of holders is also Part V's governance map.

***

## Key Terms

* **Covenant**: A term in a credit agreement or indenture that constrains the borrower and, on breach, transfers specified decision rights to the lender; the fine structure of Chapter 21 §21.4's state-contingent control allocation
* **Maintenance covenant**: A financial test applied every period regardless of the borrower's actions, so that deterioration in the business alone transfers control to the lender
* **Incurrence covenant**: A test applied only when the borrower takes a specified action — new debt, an acquisition, a dividend — and evaluated pro forma for it; silent when the business simply worsens
* **Cov-lite**: A loan carrying incurrence covenants only, now the standard structure in the US institutional leveraged loan market; a repricing of the lender's control right, not a documentation detail
* **Agency cost**: In Jensen and Meckling's accounting, the sum of the principal's monitoring expenditures, the agent's bonding expenditures, and the residual loss that survives both; capitalized into the price at which an owner-manager can sell equity
* **Free cash flow**: Cash flow in excess of what is required to fund all projects with positive net present value at the relevant cost of capital; Jensen's diagnosis of where the agency conflict is most severe, and the reason payout and leverage are governance instruments
* **Pay-performance sensitivity**: The change in a manager's wealth per one thousand dollars of change in shareholder wealth; about three dollars and twenty-five cents for the median large-firm chief executive in Jensen and Murphy's 1990 estimate, and roughly an order of magnitude higher after the options era
* **Blockholder**: A shareholder whose stake is large enough to internalize a meaningful share of the gains from monitoring, and therefore to overcome part of the free-rider problem that afflicts diffuse ownership
* **Voice and exit**: The two channels through which a blockholder disciplines management — direct engagement and the vote, versus the credible threat to sell and depress the price on which the manager's wealth depends
* **Proxy contest**: A campaign to elect directors nominated by someone other than the incumbent board, decided by a vote of shareholders; the instrument a staggered board is designed to blunt
* **Universal owner**: A holder diversified across essentially the whole market, whose stake in any one firm is small but whose aggregate control rights are large, and who therefore governs by general policy rather than by company-specific judgment
* **Fraud on the market**: The presumption, from *Basic v. Levinson*, that a purchaser in an efficient market relied on the integrity of the price rather than on the misstatement itself — the doctrine that makes the securities class action possible (Box 24.1)
* **Separating equilibrium**: In a screening market, a menu of contracts under which each privately informed type selects a different one; achieved by rationing the low-risk type's coverage until the high-risk type prefers her own contract (Box 24.2)

***

## Readings

### Required

* Jensen, M. C. and W. H. Meckling (1976). "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." *Journal of Financial Economics* 3(4): 305-360. *The founding text of the subject and the source of §24.2's model. Read the first thirty pages, where the owner-manager's perquisite problem is set up and the point that matters is made: the outside investors price the agency cost, so the entrepreneur bears it and has an incentive to bond against it. The later sections on the agency cost of debt are what Chapter 23 uses.*
* Shleifer, A. and R. W. Vishny (1997). "A Survey of Corporate Governance." *Journal of Finance* 52(2): 737-783. *The organizing survey for the whole chapter, and the one to read if only one is read. Note that this is their governance survey in the June issue, not the limits-of-arbitrage paper in the same volume's first issue. Their framing question — how do suppliers of finance get their money back? — is the control-rights question of Chapter 21 §21.4 asked from the investor's side, and their comparative treatment of legal protection and concentrated ownership is the bridge to Chapter 25.*

### Recommended

* Jensen, M. C. (1986). "Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers." *American Economic Review* 76(2): 323-329. *Six pages, and the most cited six pages in corporate finance. The argument for debt and payout as bonding devices, and the theoretical engine behind both Chapter 22 §22.6's account of the 1980s and this chapter's §24.2.*
* Jensen, M. C. (2005). "Agency Costs of Overvalued Equity." *Financial Management* 34(1): 5-19. *The agency cost that runs the other way. When the stock is worth more than the business can justify, managers are obliged to manufacture the growth the price implies, and acquisitions, accounting aggression, and value destruction follow — which extends §24.2 and connects it to the hubris material in Chapter 22 §22.5.*
* Jensen, M. C. and K. J. Murphy (1990). "Performance Pay and Top-Management Incentives." *Journal of Political Economy* 98(2): 225-264. *The source of the three-dollars-and-twenty-five-cents estimate. Read it for the measurement — how the sensitivity is constructed from salary, bonus, option holdings and share holdings — because everything since has argued about the construction as much as the number.*
* Bebchuk, L. A., Y. Grinstein and U. Peyer (2010). "Lucky CEOs and Lucky Directors." *Journal of Finance* 65(6): 2363-2401. *Option grants dated at local price minima far more often than chance allows. A measurement warning for §24.3: a pay-performance sensitivity is only as good as the grant dates it is computed from, and for a period those dates were partly chosen after the fact.*
* Gompers, P., J. Ishii and A. Metrick (2003). "Corporate Governance and Equity Prices." *Quarterly Journal of Economics* 118(1): 107-155. *The governance index, and the paper that made governance an empirical asset-pricing variable. Read alongside Chapter 22 §22.6's note that the return result did not persist out of sample while the value and operating results held up better.*
* Bebchuk, L., A. Cohen and A. Ferrell (2009). "What Matters in Corporate Governance?" *Review of Financial Studies* 22(2): 783-827. *The successor to the governance index of the preceding entry. Six of the twenty-four provisions do the work — staggered board, limits to bylaw amendment, supermajority requirements for mergers and for charter amendments, poison pills, golden parachutes — and the other eighteen are not associated with anything. Read it for the method as much as the entrenchment index: this is what narrowing an index down to what it measures looks like.*
* Brav, A., W. Jiang, F. Partnoy and R. Thomas (2008). "Hedge Fund Activism, Corporate Governance, and Firm Performance." *Journal of Finance* 63(4): 1729-1775. *The paper that established the modern activism facts. The hand-collected sample and the classification of campaign objectives are as valuable as the returns, because they show how rarely activists demand the thing they are accused of demanding.*
* Bebchuk, L. A. and S. Hirst (2019). "Index Funds and the Future of Corporate Governance: Theory, Evidence, and Policy." *Columbia Law Review* 119(8): 2029-2146. *Long, and the second half can be skimmed. The incentive argument in the first half is the foundation of §24.4 and of Chapter 17 §17.7, and the stewardship-resource evidence is the empirical core.*
* Bolton, P. and F. Samama (2013). "Loyalty-Shares: Rewarding Long-Term Investors." *Journal of Applied Corporate Finance* 25(3): 86-97. *A designed instrument for the horizon problem §24.4 treats in the abstract: rights that vest with holding period, so that the control content of a share depends on how long it has been held. Read it as security design (Chapter 21 §21.4) applied to governance rather than to financing.*
* Becht, M., P. Bolton and A. Röell (2003). "Corporate Governance and Control." In *Handbook of the Economics of Finance*, Volume 1A. Elsevier. *The comparative complement to Shleifer-Vishny, written with the European blockholder case in view rather than the dispersed American one. The right survey for §24.3's blockholder subsection and for anyone who found the separation-of-ownership-and-control premise too readily granted.*
* Leuz, C. and P. D. Wysocki (2016). "The Economics of Disclosure and Financial Reporting Regulation: Evidence and Suggestions for Future Research." *Journal of Accounting Research* 54(2): 525-622. *Disclosure regulation as the third governance instrument beside boards and pay, and the one §24.3 does not have. Their treatment of what is and is not identified in the mandatory-disclosure literature is the honest counterpart to this chapter's own audit.*
* Fisch, J. E. (2010). "The Overstated Promise of Corporate Governance." *University of Chicago Law Review* 77: 923-958. *The skeptical review, and the counterweight §24.3's toolkit framing needs. The argument is that the governance instruments are weaker, more contingent, and worse identified than either the indices or the reform advocacy admit.*
* Gilson, R. J. and R. H. Kraakman (1984). "The Mechanisms of Market Efficiency." *Virginia Law Review* 70(4): 549-644. *The doctrinal source of the fraud-on-the-market premise Box 24.1 rests on, and the piece that took efficiency apart into the specific mechanisms — informed trading, professional arbitrage, derivatively informed trade — by which information reaches a price. The natural bridge from Chapter 7 §7.2 into securities litigation.*

***

## Discussion Questions

1. **Can index funds govern?** Section 24.4 argues that the capacity constraint which makes index managers weak firm-specific monitors is the same constraint that makes them a concentrated and reachable audience. Take both sides. First, build the case that policy-based voting by three managers is *better* governance than the dispersed alternative — consider consistency, the elimination of a free-rider problem across thousands of small holders, and the fact that a published policy can be criticized while a private judgment cannot. Then build the case that it is worse, using the incentive argument of Bebchuk and Hirst, the proxy advisers' role, and the observation in §24.4 that support for one class of proposals rose and fell within two seasons without any change in the firms. Finally, state what evidence would move you, and say whether pass-through voting is a test of the question or a change of subject.
2. **Is cov-lite a repricing of a control right, or a giveaway?** A loan whose maintenance covenants have been removed is a different security from one that retains them. If credit markets are efficient in Chapter 7's semi-strong sense, the spread should compensate lenders for the lost control right, and the migration in Table 24.2 is a market clearing at a new point. Set out what the "efficient repricing" view predicts about (a) the spread differential between otherwise identical cov-lite and covenanted loans, (b) recovery rates in default, and (c) the timing of restructurings. Then set out the competing view — that the migration reflects a shift in *who holds the claim* rather than a change in its price — and identify which of the three predictions distinguishes them. Which of Chapter 18's holder facts does the second view rest on?
3. **Whose discipline was it?** Section 24.1 reports that capital expenditure falls and chief executives are replaced more often after a covenant violation, and §24.3 reports that the same things happen after a hedge fund files a 13D. Both are described as governance working. Are they the same phenomenon? Consider who bears the cost of the intervention, who captures the gain, what each party's objective function is, and in which states of the world each instrument is available. Then answer a harder question: if creditor discipline and shareholder discipline both cut investment, how would you tell an efficient correction from an inefficient one?
4. **The convexity problem.** Section 24.2 argues that undiversified managers are too conservative about idiosyncratic risk and that convex option payoffs make them too aggressive, and that no single instrument fixes both. Design a compensation contract that does better than either — you may use restricted stock, options at various strikes, indexed instruments, deferred cash, clawbacks, and vesting schedules. State what your contract does to the manager's incentive at high and low stock prices, what it costs the firm, what it is vulnerable to, and why boards might not adopt it. Does the answer change if the firm is a bank?
5. **Litigation without owners.** Box 24.1's circularity critique observes that a securities class action moves money from current shareholders to former purchasers, most of whom are the same diversified institutions. Suppose the entire equity market were held by three index funds. What would the class action then be doing, and what would remain of the deterrence argument? Use your answer to say something general about which governance instruments survive universal ownership and which do not.

***

## Problems

**Problem 1 — Perquisites and the price of dilution.** Use §24.2's model. A firm's value with no perquisite spending is $$V\_0 = 500$$ (millions of dollars). The manager's private benefit from perquisite spending $$F$$ is $$B(F) = bF - \tfrac{1}{2}cF^2$$ with $$b = 1.4$$ and $$c = 0.02$$. The manager retains a fraction $$\alpha$$ of the equity and sells the rest to rational outside investors.

(a) Derive $$F^{\ast}(\alpha)$$ and compute it for $$\alpha = 1$$, $$0.6$$, $$0.4$$, and $$0.2$$. Compute the firm's market value $$V\_0 - F^{\ast}$$ in each case.

(b) For $$\alpha = 0.4$$, compute what outside investors pay, the value of the manager's retained shares, her private benefit, and her total wealth. Verify that her total wealth equals $$V\_0$$ less the residual loss.

(c) Compute the residual loss for each $$\alpha$$ in (a), defined against the sole owner's choice. Plot or tabulate it against $$\alpha$$ and describe its shape.

(d) The manager can hire an auditor who costs 3 per year and mechanically caps perquisite spending at 35. At $$\alpha = 0.4$$, does she hire the auditor? At $$\alpha = 0.2$$? State the general principle your two answers illustrate, and identify which of Jensen and Meckling's three components the auditor's fee is.

(e) Now suppose outside investors *cannot* observe $$\alpha$$ and assume $$\alpha = 0.4$$ regardless of the truth. Show that the manager's privately optimal $$\alpha$$ is not 1, and explain in two sentences what this does to the interpretation of part (b).

**Problem 2 — A covenant threshold and what it buys.** A borrower has trailing EBITDA of $120 million, total debt of $480 million, and annual cash interest of $34 million. Its credit agreement contains two maintenance covenants: net debt to EBITDA no greater than 4.5 times, and EBITDA to interest no less than 3.0 times, both tested quarterly.

(a) Compute the current level of each ratio and the headroom in each, expressed as the percentage decline in EBITDA that would trip it. Which covenant binds first?

(b) EBITDA falls 15 percent. State which covenants are breached. The lender does not accelerate. List three things it can extract in a waiver negotiation, and explain in one sentence for each why the borrower would rather concede it than face acceleration.

(c) The waiver is priced at a 50 basis point fee on the outstanding debt plus a 100 basis point increase in the margin for the remaining three years of the facility. Compute the total cost to the borrower, undiscounted, and express it as a percentage of the year's EBITDA after the decline.

(d) Now suppose the loan is cov-lite: incurrence covenants only, tested pro forma on new debt, acquisitions and restricted payments. Repeat (a) and (b). State precisely what event, if any, transfers control to the lender under this documentation, and what the lender's practical remedy is in the meantime.

(e) The borrower in (d), with liquidity running low, moves a valuable subsidiary outside the restricted group and raises new debt secured against it. Explain what this does to the original lenders' recovery, why the cov-lite structure made it available, and how a maintenance covenant would have altered the sequence of events. Relate your answer to Chapter 10 §10.1's recovery-rate distribution.

**Problem 3 — Pay-performance sensitivity.** A chief executive holds 2,000,000 shares of her firm outright and options on a further 3,000,000 shares with an average delta of 0.6. The firm has 400 million shares outstanding, trading at $50.

(a) Compute the firm's market capitalization and the manager's *effective* ownership in delta-equivalent shares, as a fraction of shares outstanding.

(b) Compute pay-performance sensitivity in dollars of change in the manager's wealth per one thousand dollars of change in shareholder wealth. Compare with Jensen and Murphy's estimate for the median large-firm chief executive.

(c) The stock falls 40 percent and the options move deep out of the money, cutting their average delta to 0.2. Recompute (a) and (b). The board proposes to reprice the options back to the money, restoring a delta of 0.6. Compute the new sensitivity and give one argument for the repricing and one against.

(d) A director proposes replacing the option grant with 1,800,000 shares of restricted stock. Compute the sensitivity under the new package. State what the change does to the manager's incentive to take idiosyncratic risk, and connect your answer to §24.2's discussion of managerial risk preferences.

(e) The options are struck at $50. The firm's stock rose 30 percent over the year and its industry index rose 28 percent; value the options at intrinsic value at the end of the year. Compute the manager's gain under the original package and under a package identical except that the options' strike is ratcheted by the industry index. Use the difference to state precisely what "pay for luck" costs shareholders, and give one reason indexed options are nonetheless rare.

**Problem 4 — Voting arithmetic.** A company has 4.2 billion shares outstanding, trading at $60. Ownership and expected turnout are as follows: the three largest index managers hold 20 percent of shares and vote 99 percent of them; other institutions hold 55 percent and vote 90 percent; retail holders hold the remaining 25 percent and vote 30 percent.

(a) An activist acquires a stake of 0.02 percent of the shares. Compute the number of shares and the dollar value of the position.

(b) Compute the total votes cast, as a percentage of shares outstanding, and the three index managers' share of *votes cast*. Explain in one sentence why it exceeds their share of shares outstanding.

(c) A dissident slate needs a majority of votes cast. Compute the number of votes required. Then compute what fraction of the non-index vote the dissident must win (i) with the three index managers' support and (ii) without it.

(d) Suppose pass-through voting is introduced and 30 percent of the index managers' shares are voted directly by underlying investors, who vote at a 20 percent rate and split evenly between the two slates. Recompute (b) and (c)(i). Does the reform help or hurt the dissident here, and what does your answer depend on?

(e) The activist's campaign costs $30 million. Compute this as a percentage of its own stake and as a percentage of the company's market capitalization. Then compute the increase in the company's value that would be needed for the campaign to break even *for the activist alone*. Use the three numbers to explain why campaigns of this kind are rational only under the ownership structure described in §24.4.

**Problem 5 ★ — Screening contracts.** Use Box 24.2's setting: $$W = 100$$, $$L = 50$$, $$u(c) = \sqrt{c}$$, high-risk probability $$\pi\_H = 0.4$$, low-risk probability $$\pi\_L = 0.2$$.

(a) Compute each type's certainty equivalent with no insurance, and the full-insurance contract each would buy at its own fair odds.

(b) Suppose high risks are 40 percent of the population. Compute the pooled fair-odds full-insurance contract. Show that the low-risk type prefers the separating contract $$(98.55, 55.80)$$ to it, and that the high-risk type prefers the pooled contract to $$(80, 80)$$.

(c) Verify that the separating pair satisfies both incentive constraints and that each contract breaks even against the type that takes it.

(d) Find the share of high risks below which the low-risk type would prefer the pooled contract to her separating contract. Explain why the existence of that region is a problem for the concept of equilibrium rather than for the consumers.

(e) A regulator mandates that everyone buy the pooled full-insurance contract at the population-average odds, with 40 percent high risks. Compute each type's certainty equivalent and compare with the separating outcome. Who gains, who loses, and by how much? State what the calculation does and does not settle about whether the mandate is a good idea.

***

## Selected Solutions

*Solutions to Problems 1 and 4 follow. Solutions to the remainder are in the instructor materials.*

**Problem 1.**

(a) The manager maximizes $$\alpha(V\_0 - F) + B(F)$$, giving the first-order condition $$-\alpha + b - cF = 0$$ and $$F^{\ast}(\alpha) = (b-\alpha)/c = (1.4-\alpha)/0.02$$.

**Table 24.4: Perquisite spending and firm value in Problem 1**

| $$\alpha$$ | $$F^{\ast}$$ | Market value $$V\_0 - F^{\ast}$$ |
| ---------- | ------------ | -------------------------------- |
| 1.0        | 20           | 480                              |
| 0.6        | 40           | 460                              |
| 0.4        | 50           | 450                              |
| 0.2        | 60           | 440                              |

*Source: Author's calculation, on §24.2's parameters — a private-benefit slope of 1.4, a curvature of 0.02, and a no-perquisite firm value of 500.*

(b) At $$\alpha = 0.4$$: market value is 450, so outsiders pay $$0.6 \times 450 = 270$$ and the manager's retained shares are worth $$0.4 \times 450 = 180$$. Her private benefit is $$B(50) = 1.4(50) - 0.01(2500) = 70 - 25 = 45$$. Total wealth $$= 270 + 180 + 45 = 495 = 500 - 5$$. The sole owner's total is $$500 + \[-20 + B(20)] = 500 + 4 = 504$$, so the manager is 9 worse off — exactly the residual loss computed in (c).

(c) Perquisite surplus is $$S(F) = -F + B(F) = 0.4F - 0.01F^2$$, maximized at $$F = 20$$ with $$S = 4$$. Then $$S(40) = 0$$, $$S(50) = -5$$, $$S(60) = -12$$, and the residual losses against the sole owner are 0, 4, 9 and 16 for $$\alpha = 1$$, $$0.6$$, $$0.4$$, $$0.2$$. The loss is quadratic in the dilution: $$S(20) - S(F^{\ast}(\alpha)) = 0.01(F^{\ast}(\alpha)-20)^2 = 25(1-\alpha)^2$$. Check: $$25(0.6)^2 = 9$$ at $$\alpha = 0.4$$ and $$25(0.8)^2 = 16$$ at $$\alpha = 0.2$$. ✓

(d) At $$\alpha = 0.4$$ the cap binds ($$F^{\ast} = 50 > 35$$). Surplus rises from $$S(50) = -5$$ to $$S(35) = 0.4(35) - 0.01(1225) = 14 - 12.25 = 1.75$$, a gain of 6.75 against a cost of 3, so she hires the auditor: net gain 3.75. At $$\alpha = 0.2$$, $$F^{\ast} = 60$$, and the cap raises surplus from $$-12$$ to $$1.75$$, a gain of 13.75 against the same cost of 3. She hires it, and more eagerly. The principle: the value of monitoring rises with the severity of the agency problem, so bonding and dilution are complements — the entrepreneur who sells more of the firm buys more monitoring, and does so *voluntarily*, because the price outside investors pay reflects it. The fee is a **bonding expenditure** in Jensen and Meckling's taxonomy, since it is incurred by the agent, not the principal.

(e) If investors pay $$450$$ per unit of value sold regardless of $$\alpha$$, the manager's wealth is $$(1-\alpha)450 + \alpha\bigl(V\_0 - F^{\ast}(\alpha)\bigr) + B(F^{\ast}(\alpha))$$. Evaluating: $$504$$ at $$\alpha = 1$$, $$496$$ at $$\alpha = 0.6$$, $$495$$ at $$\alpha = 0.4$$, $$496$$ at $$\alpha = 0.2$$, $$498.06$$ at $$\alpha = 0.05$$, and $$499$$ at $$\alpha = 0$$. The function has a *minimum* exactly at the believed value $$\alpha = 0.4$$, and rises in both directions: above it she is selling a firm worth more than the price, below it she is selling one worth less. So the honest $$\alpha$$ is the one choice she would never make, and the global optimum here is to keep the whole firm. Part (b)'s clean result — that the manager bears the entire agency cost — holds only because investors' beliefs about $$\alpha$$ are correct *in equilibrium*. That is a rational-expectations assumption, not an accounting identity, and where it fails the agency cost is partly a transfer after all.

**Problem 4.**

(a) $$0.0002 \times 4{,}200 \text{ million} = 840{,}000$$ shares, worth $$840{,}000 \times 60 = 50.4$$ million dollars.

(b) Per 100 shares outstanding, votes cast are $$20 \times 0.99 = 19.8$$ from the index managers, $$55 \times 0.90 = 49.5$$ from other institutions, and $$25 \times 0.30 = 7.5$$ from retail: **76.8 percent of shares outstanding are voted**. The index managers' share of votes cast is $$19.8/76.8 = 25.8$$ percent. It exceeds their 20 percent ownership because they vote nearly everything while the least-voting block, retail, is diluted almost out of the count.

(c) A majority of votes cast is $$76.8/2 = 38.4$$. The non-index votes cast total $$76.8 - 19.8 = 57.0$$. **With** the index managers, the dissident needs $$38.4 - 19.8 = 18.6$$ of those 57.0, or **32.6 percent** of the non-index vote. **Without** them, it needs all 38.4 from the same 57.0, or **67.4 percent**. Winning over three stewardship teams moves the required support among everyone else from more than two thirds to less than a third.

(d) Of the index managers' 20 shares, 6 are now voted by underlying investors at a 20 percent rate, contributing $$6 \times 0.20 = 1.2$$ votes cast, split 0.6 to each slate; the managers still vote $$14 \times 0.99 = 13.86$$. Total votes cast fall to $$13.86 + 1.2 + 49.5 + 7.5 = 72.06$$, and a majority is 36.03. The managers' share of votes cast falls to $$13.86/72.06 = 19.2$$ percent. With their support plus the 0.6 pass-through votes, the dissident needs $$36.03 - 13.86 - 0.6 = 21.57$$ of the remaining $$57.0$$, or **37.8 percent** — worse than the 32.6 percent of (c). Pass-through *hurts* the dissident here, because the votes it removes from the managers' bloc are largely not cast at all, and the ones that are cast split evenly. The answer depends entirely on the participation rate of the pass-through investors and on how they split: raise their turnout and tilt them toward the dissident and the sign reverses.

(e) Market capitalization is $$4{,}200 \times 60 = 252$$ billion dollars. The campaign costs 59.5 percent of the activist's own stake of 50.4 million dollars and about 0.012 percent of the company. To break even on its stake alone, the activist needs the company's value to rise by $$30\text{m}/0.0002 = 150$$ billion dollars, or 59.5 percent — which it plainly does not expect. The campaign is therefore not an investment in the stake; it is an investment in the fund's franchise, and it is feasible only because the persuasion target is three stewardship teams applying published criteria rather than millions of dispersed holders. Under dispersed ownership the same thirty million dollars buys a mailing.

***

## Data Exercise: Reading Control Rights Out of Public Filings

Parts A through C run entirely on free data from the SEC's EDGAR system. Part D is the licensed extension.

**Part A — The ownership table behind the vote (free, EDGAR).** Choose a large US listed company and retrieve its most recent definitive proxy statement (form type DEF 14A) from EDGAR's full-text search.

1. Locate the beneficial ownership table. Record every holder listed at five percent or more, with its stake, and record the aggregate holdings of directors and officers as a group.
2. Compare the five-percent holders against Chapter 2 Table 2.5's holder columns: which sector does each belong to, and what does the sum of the fund-sector holders come to? State what share of the company's votes the three largest asset managers control if all shares are voted.
3. Find the compensation discussion and analysis. Record the chief executive's holdings of shares and of vested and unvested options, and compute the pay-performance sensitivity of Problem 3(b) using the share count and price from the same filing. Note what you had to assume about option delta and how much the answer moves if you assume 0.4 instead of 0.6.
4. Record the board's structure: is it staggered or annually elected, what fraction of directors is independent, and what were the votes for and against each director at the last meeting?

**Part B — How the Big Three actually voted (free, EDGAR).** Form N-PX is the annual voting record every registered fund must file, and it is free.

5. Pick a contested item — a dissident director slate, a merger, or an environmental or social shareholder proposal — at a company of your choice, and retrieve the N-PX filings of one large index fund from each of the three largest managers for the year in question.
6. Record how each fund voted on each item, and whether the manager published a vote bulletin explaining it. Compare the three.
7. Repeat for the same three funds on a *comparable* proposal two years later. Section 24.4 claims that support for environmental and social proposals rose through 2021 and fell back afterward. Does your pair of observations agree? State honestly what two observations can and cannot establish, and design the sample you would need to test the claim properly.

**Part C — Covenants in the document (free, EDGAR).** Credit agreements are filed as exhibits to the 8-K or 10-K of the borrower.

8. Find the credit agreement of a leveraged borrower. Locate the financial covenants article. Classify each covenant as maintenance or incurrence using Table 24.1's test, and record the levels and any step-downs.
9. Locate the definition of EBITDA in the agreement and list every permitted add-back. Estimate, from the definition alone, how much reported EBITDA could exceed the accounting figure. Explain what this does to the effective tightness of the covenant you recorded in (8).
10. Find a company that disclosed a covenant waiver or amendment. From the amendment document, record what the lenders received: fees, margin changes, new covenants, collateral, restrictions on investment or payout. Write one paragraph relating what you find to §24.1's claim that a breach transfers a bargaining position rather than a repayment right.

**Part D ★ — Governance panels (WRDS).** With ISS (formerly RiskMetrics) governance and incentive-lab data, Compustat, and CRSP:

11. Build a panel of board characteristics and antitakeover provisions for the S\&P 1500 over as long a window as the data allow. Construct the entrenchment index and regress firm value (market-to-book) on it with industry and year fixed effects. Report the coefficient, and then repeat the exercise on returns rather than valuations, splitting the sample at 2000. Relate what you find to Chapter 22 §22.6's out-of-sample note.
12. Construct pay-performance sensitivity for each chief executive-year from holdings and option deltas. Plot the cross-sectional median from the earliest year available to the present, and mark the mid-2000s accounting change. Then regress sensitivity on firm size and report the elasticity; explain why a size elasticity below one is what Jensen and Murphy's dollars-per-thousand metric is sensitive to.
13. Merge a Schedule 13D filing sample onto the panel and reproduce the announcement return in Brav, Jiang, Partnoy and Thomas with Appendix A's event-study machinery. Then follow targets for five years on operating profitability, capital expenditure, and research and development, against a matched control group. Report which of the short-termism critique's predictions your data support and which they do not, and state the identification problem you could not solve.