Agency problem

Berle and Means gave the problem its name in The Modern Corporation and Private Property (1932): ownership had dispersed so widely across American corporations that control had passed to professional managers who owned almost none of the equity. Jensen and Meckling made it a cost in 1976, and their decomposition is still the cleanest one available. Agency costs are the sum of monitoring expenditure by the principal, bonding expenditure by the agent, and the residual loss that survives both.

The residual-loss term is the one that matters and the one that gets forgotten. Monitoring and bonding are visible: an audit fee, a compliance department, a lock-up, a clawback clause. The residual loss is what is still going wrong after all of that spending, and by construction it never appears in any account. An analyst who says a company has good governance because it spends heavily on compliance has measured the monitoring term and ignored the one that costs shareholders money.

The residual loss has no line in any ledger.

Limited liability creates a second, quieter conflict at the same moment. Once the owner's downside is capped, the owner's payoff becomes convex in the value of the firm: all of the upside, a floor on the downside. The lender's payoff is the mirror image, concave, capped at repayment. Two claimants on the same cash flows now disagree about risk, and they disagree most violently exactly when the firm is distressed.

That disagreement was created by the form, not by anybody's bad faith.

Who owns it, and who controls it

Ownership takes four shapes and each one changes which conflict dominates. Dispersed ownership, the classic American public company, leaves no shareholder with either the incentive or the ability to monitor, so the manager is the residual decision-maker and the shareholder-versus-manager conflict dominates. Concentrated ownership by an institution or a family gives somebody both the incentive and the votes to monitor, which suppresses the manager conflict and raises a different one. Promoter control, the Indian default, is concentrated ownership with a formal legal category attached: a promoter is a disclosed person or group under Indian securities law, and roughly three-quarters of Indian listed companies have one (check against current shareholding-pattern data before quoting the share).

Then there is state ownership, which adds an owner whose objective function includes goals that are not the share price.

How do you measure the wedge between votes and ownership?

Where a controller exists, the conflict that matters is controller against minority, and the sharpest measure of it is the wedge between voting power and economic exposure. Take the register from the opening. Northbridge Media (synthetic) has 46.0 million Class A shares at one vote each and 9.0 million Class B shares at ten votes each, so 55.0 million shares carry 136.0 million votes. The founder group holds all the Class B.

The ratio of voting power to economic ownership is 4.0441 times. A rupee of the founders' money carries four times the say that a rupee of yours does, which is the same as saying that the founders bear a quarter of the cost of any decision they make badly.

Now push it. Suppose each Class B share converts one-for-one into Class A when sold, which is the usual sunset design. Selling s million shares leaves the founders 10(9 - s) votes out of (46 + s) + 10(9 - s) = 136 - 9s total votes. Set that ratio to one half and solve: 90 - 10s = 68 - 4.5s, so s = 4.0 million. The founders can sell 44.44% of their holding, take their economic stake down to 9.0909%, and still hold exactly half the vote. Everything below that stake is a company where nine cents in the dollar decides.

Indian promoter control usually reaches the same place without a second share class, because a 45% to 55% single-block holding is enough when the rest of the register is fragmented and institutional turnout is imperfect. The wedge is then latent rather than structural, and the number to compute is the promoter stake against the actual voted share at the last few resolutions, not against the register.

Four conflicts, and how to tell which one you are looking at

Every governance story reduces to one of four disagreements. Naming the right one is most of the analysis, because the mechanism that fixes each is different.

The first is the shareholder against the manager. A manager wants scale, tenure and a quiet life while the shareholder wants return on capital, and the classic expression is the empire-building acquisition that lifts reported earnings and destroys value. The classic tell is a bonus scheme keyed to revenue or to earnings per share rather than to return on invested capital.

The second is the shareholder against the creditor. Once the debt is priced and issued, the shareholder gains from raising asset risk, because the upside accrues to equity and the downside beyond the equity cushion falls on the lender. Covenants exist to police exactly that, and the mechanism has its full treatment on the credit side of the curriculum.

The third, and the one that dominates in India and in any controlled company anywhere, is the controlling shareholder against the minority. A controller can extract value through channels that never touch the dividend: a related-party supply contract at an off-market price, a royalty to a promoter-owned brand entity, a loan to a group company repayable on demand, an issue of shares to a friendly entity at a price the board sets. The minority holds the same claim on paper and a smaller one in fact.

The fourth is the firm against everyone else. Employees, suppliers, customers, communities and the state hold claims that are contractual or statutory rather than residual, and a firm can transfer value from them too, by underpaying for an environmental externality or by squeezing a supplier's working capital. The environmental and social material further on is largely the analytical treatment of that fourth conflict.

Worked example: related-party transactions at Vindhya Ferro Alloys

Vindhya Ferro Alloys (synthetic), the Indian mid-cap scored in Worked Example 1, shows the third conflict at ₹412 crore of related-party transactions against ₹3,050 crore of revenue, which is 13.5082% of the top line moving through parties the controller also owns. That number proves no extraction on its own. It sizes the channel through which extraction, if it is happening, would have to run, and that is what a first-pass analyst is entitled to conclude.