Shareholder Voting and Social Choice

How does one-share-one-vote circumvent Arrow's impossibility theorem?

Here’s an economics puzzle. We know that Arrow's impossibility theorem implies there's no "perfect" social choice mechanism. But most companies are basically run through one-share-one-vote. Why is this? Does the share voting mechanism "solve" Arrow's impossibility in some sense?

This is a complicated question, but here's one proposed answer. Shareholder voting is not actually designed to solve a "social choice" problem, in Arrow's sense, of aggregating multiple parties' heterogeneous preferences into a social preference!

Instead, shareholder voting settings essentially take as given that all preferences are aligned, and that the objective function of all shareholders is identical: to run the company in a way that maximizes profits. But different shareholders may have different views on what actions of the firm maximize profits! Shareholder voting fundamentally solves an information aggregation problem, rather than a full social choice problem. Taking as given everyone ultimately wants the same thing, shareholder voting weights the firm's decisions towards what the biggest shareholders think will maximize profits.

In normal times, shareholders may disagree about whether a certain action of the firm is in the firm's best interests. But this is not that bad an issue — most of the more egregious failures of voting aren't too problematic in the shareholder voting setting.

Here’s the canonical thought experiment which shows how majority voting fails as a social choice mechanism. Suppose you ask a room of 10 people to vote whether to give each of 9 people a dollar and shoot the last person. This vote will pass with 9/10 votes, though it is obviously a stupid outcome. Here’s a few examples to see how similar forces play out in real shareholder voting settings.*

• Suppose Elon Musk owns 100% of Tesla and 51% of Solar City. Elon can then force Solar City to sell parts to Tesla at stupid low prices. This is terrible for Solar City shareholders, but great for Elon, since he gains the increased profits from the low-priced sales to Tesla. Elon is basically using his control of Solar City to extract from other Colar Sity shareholders.

• Suppose Luna equityholders control 51% of Curve voting rights. Luna can force Curve to provide lots of liquidity for Terra pools. This is not profit maximizing for Curve, and screws over some other Curve holders, but it increases Terra profits. Similarly, Luna is using its control of Curve to “siphon” profits from other Curve equityholders to Luna equityholders.

• As a non-monetary example, suppose an environmental group (E) controls 51% of Ford. E can then, say, make Ford shut down gas car production and only run electrics. This screws over the 49% of Ford owners, but benefits E since their objective is to save the planet. Similarly to the first two examples, E is “siphoning” profits from other Form equityholders, though here in a non-monetary way.

All these are essentially examples where majority share controllers use their voting power to "funnel/siphon" resources away from the 49%, towards themselves. They're in spirit similar to the shoot-a-person-for-a-dollar case example.

The point of these examples is to illustrate that shareholder voting only really works, when everyone is aligned on running the firm to maximize the firm's own profits, and the only question to be solved is what actions of the firm maximize profits. In any setting where the firm's actions have externalities on shareholders — and, thus, large shareholders have some incentives to run the firm in a non-profit-maximizing manner — shareholder voting can lead to much weirder outcomes!

---

*All examples hypothetical and purely for sake of illustration!