Mr. China is a memoir of doing private equity investing in China in the late 80s and early 90s, when the economy was free market enough that there was such a thing as private equity, but still had a long way to go before it worked like the US. The book is basically a series of vignettes that go one of two ways:

1. We bought a company that was profitable, and as soon as we bought it their profits started going down. After careful investigation, we caught the manager embezzling money, or

2. We bought a company that was profitable, and it remained profitable. One day, the CEO transferred all of the company's cash to an offshore bank account and disappeared.

These scenarios sound like the kinds of things an over-literal child would dream up the first time they learned about the concept of investing in a business. But they really happened, and similar things happen all the time. Control of a company exists partly as a set of legal norms, partly through a web of social relations and traditions, and partly in the form of who has access to which passwords.

The broad evolution of managerial control starts with the convergence of two different models: sole proprietorships and family businesses scaling to the point that they need outside capital, and the quasi-state trading companies they gradually converged with. The original meaning of a corporation was basically a partial delegation of state power: the government might turn over control of some asset, like a bridge or a monopoly on some product, to a private actor who would promise to operate it in exchange for collecting profits. This scaled up rapidly when it became a default way for countries to have overseas trading relationships.¹ Founder control is a very primitive kind of governance: whoever defines a thing as independent is implicitly laying some kind of claim to it. And a board of directors elected by stockholders appointing a professional manager on their behalf is a comparatively modern idea. Corporate governance has long been in tension between these two ideas, of personal and individual rule or impersonal and collective institutions.

There were times when banks, as providers of marginal capital, were the most powerful participants in the debate: they could dictate whether or not a company could refinance its bonds or roll over its loans, so they had an effective veto over shareholders. Postwar companies tended to have powerful CEOs and boards, which had the positive effect of ensuring that companies were run by people who understood how they worked, and the negative effect that sometimes it's good for a spreadsheet jockey to chime in and note that even if a given business is run exceptionally well, it'll produce a worse return than t-bills, and that the main determinant of its value is how soon management recognizes this and liquidates it.

The rise of corporate raiders in the 1980s was a kind of paradoxical fight between two models of personal rule, one of which was that a CEO was inviolably tied to the company he ran, and one of which was connected to whoever could buy enough shares to be able to threaten a takeover, or at least to threaten adding some ornery new board members.

Early shareholder activists were very good at spinning their campaigns as a populist insurgency against complacent managers. Oddly enough, this rhetoric is less common today even though changes in the tax code have made the most shareholder-unfriendly raider maneuever—taking "Greenmail," or an above-market tender offer available only to them—impractical. But the threat of shareholder activism has led more companies to create elaborate multi-class structures for their voting rights, such that outsiders can get economic exposure much more easily than they can get control.

There are limits to such a legalistic approach: when Sam Altman was fired by OpenAI's board in late 2023, they were acting within their legal rights and performing their legally-sanctioned role. But Altman's biggest asset wasn't formal power, but the informal power of being able to convince people to threaten to leave with him (or convince them to convince other people of this). If the CEO has the personal loyalty of key employees, it doesn't really matter what contracts say, especially in a state with minimal enforcement of noncompetes.

Control is always both a legal and a social fact. In practice, companies are constrained by public opinion and potential legal action even if they theoretically have wide latitude to pursue their aims. On the margin, technicalities about vote count can make a big difference, especially for a company whose assets skew towards physical ones or brand names rather than pure human capital. But corporate control is ultimately the very literal question of whose instructions will be followed in the event of a conflict. Whoever wins that tie vote is in charge of the company, and if legal structures ratify that fact, it's in the form of recognizing existing reality.

The Diff has covered corporate governance in a few different places:

• The market in talent and access is far from efficient ($).

• Activist investors sometimes target bad companies, but sometimes end up going after companies with a longer-term outlook than their peers ($).

• There used to be worst matching between managers and potential value creation.

• One of the things that can change effective control is access to capital, as the junk bond boom demonstrates ($).

Join the discussion!

1  This is one of many examples of a regulatory arbitrage that turned into a good business. One of the reasons these companies existed was that for a foreign company, doing business with the Mughals or Ming entailed actually going to the throne room and acknowledging them as sovereign. It would raise awkward questions if a direct representative of the English government referred to England as a tributary of the Mughal Empire, but an independent contractor who was merely state-sponsored could. So, they kowtowed in Agra and Beijing and realized high-90s gross margins in London and Lisbon.