I wanted to make this an real essay about the rise of dual class share structures in tech companies and have an original take on how we should respond to it and What It Means, but after reading a bunch of papers about it I didn’t really have one, so here is a stream of consciousness instead.
The unit of the corporation is a hallmark of American capitalism. A corporation is a legal entity where people can come and go as they fill in the standard slots: the employees, the shareholders, the CEO, the board of directors, etc. Any individual person, no matter how key, is in theory expendable, so if a specific person fails the corporation has failsafes to keep from going down with him. The CEO is performing poorly? Whatever, the board fires the CEO and gets a new one.
The corporation is legally a person, but it is de-personalized.
The Silicon Valley startup of lore is not.
Everyone knows the story of how Apple fired its own founder, sank, brought him back, and then experienced the most remarkable turnaround in business. This is THE Silicon Valley story. Being ‘the guy who fired Steve Jobs’ is enough for permanent infamy. The startup founder is the lynchpin of the entire infrastructure that keeps the tech industry innovating, the person who quits their job to build their own vision.
In the past decade the term ‘founder friendly’ has gained prominence as investors compete to signal to entrepreneurs that they won’t kick the founder out of their company so that the founder will choose to take an investment from them. Part of this has been the rapid rise of what are called ‘dual class share structures’, which are designed to let founders keep control of their companies even after they IPO. Facebook, Google, Snap, LinkedIn, Square, Yelp, Zillow, and several other ‘unicorns’ all have this arrangement.
Here’s how it works. In a normal corporation, people own shares that correspond to ownership and claims on cash flows from the company. If I own 1% of the shares of a company, I have a claim on 1% of the dividends from that company. If the company goes up in value, the value of my holdings goes up by 1% of the amount that the company’s value increased. I also get 1% of the votes on corporate governance issues.
Generally speaking, the more shares I own, the more influence I have over the direction of the company, and depending on governance bylaws, if I own more than 50% of the company I may be able to have mostly uncontested control of what the company does. Pretty straightforward.
Now let’s imagine a company X with a visionary founder/CEO, Sarah, and two classes of shares, class A and class B. Class A shares operate according to the traditional one share, one vote principle. Class B shares have 100 votes. X IPOs with 1 class B share and 99 class A shares, with Sarah holding the class B share. The class A shares are now tradable on public markets. Sarah automatically has majority control of the voting shares of X because even if another person were to buy every single class A share, Sarah would still control a majority of votes (at least 100 out of 199 votes), even though she only has a claim to 1/100 of the equity cash flows from the company.
The company raises money off of the sale of the class A shares and the investors get to cash out, but Sarah is effectively an emperor. Someone who holds some class A shares introduces a plan to get X to issue some more shares? Fire the CEO? Change the board? Too bad, the holder of the class B share can vote it all down.
This is interesting to me because it’s a clear case of a cultural tendency being codified into legal economic institutions. The current boom of dual class in tech started with Google’s IPO in 2004, but the structure has existed in some form or another for a long time before that. Different countries around the world have different approaches to dual class, some going as far as total bans. Here’s a brief history of it in the US, from Democratic SEC Commissioner Robert Jackson:
“For most of the modern history of American equity markets, the New York Stock Exchange did not list companies with dual-class voting. That’s because the Exchange’s commitment to corporate democracy and accountability dates back to before the Great Depression. But in the midst of the takeover battles of the 1980s, corporate insiders “who saw their firms as being vulnerable to takeovers began lobbying [the exchanges] to liberalize their rules on shareholder voting rights.” Facing pressure from corporate management and fellow exchanges, the NYSE reversed course, and today permits firms to go public with structures that were once prohibited.
As you all know well, more and more companies choose today to go public with dual-class. Public companies using dual-class are today worth more than $5 trillion, and more than 14% of the 133 companies that listed on U.S. exchanges in 2015 have dual-class voting. That compares with 12% of firms that listed on U.S. exchanges in 2014, and just 1% in 2005.”
So is this good? Bad? Unimportant?
From what I can see of the research, most academics who study this think that perpetual dual class is probably bad (here’s a paper that I found to be comprehensive and persuasive). To use the example of company X above, a perpetual dual class structure would mean that the class B voting share will always have 100 votes. Let’s say Sarah runs the company until she gets tired of it and quits, at which point she writes a legal document that says, “I hereby bequeath my class B share to my son, Dave.” Now, Dave, whoever that guy is, owns a controlling share of the company. Do you, an owner of a class A share, think Dave is also a visionary founder who should be calling the shots? Well no one asked you, and Dave gets to be emperor anyway, and dynasties do have a habit of ending up like this.
For a concrete example of a perpetual dual class structure turning into a governance nightmare, see Sumner Redstone and the battle for CBS.
But dual class structures don’t have to be perpetual. You can do what’s called a sunset- stipulate that, say, 10 years after the IPO, the voting rights of the class B share are reduced to 1 vote, and so it will become a normal share. This seems more reasonable, as the founder does gets an extended period of time in which she gets to implement her visionary future, but after a while the company does get to eliminate the risk that the founder will lose her visionary touch or pass on the share to someone less competent.
There is another reason to be skeptical of dual class structures, and it’s what economists call agency costs. I’m gonna recycle an argument from this paper, which I think did a good job of explaining how that concept works here.
Agency costs basically mean that the value-maximization of a firm can be impaired if there is a difference between the management of the firm and the ownership of the firm. Let’s use the example where Sarah owns only 1/100 of X but has a controlling voting share. Let’s call the value of the company C and Sarah’s benefit from owning the company other than the equity cash flow she gets G. Sarah’s benefit from this arrangement is .01*C + G.
Let’s say a business opportunity Z comes along that increases C (the value of the company, which would benefit all shareholders) but also decreases G. Because Sarah is in the position to make the call on whether the company will do Z, she will only do it if the value of .01*C goes up more than G goes down. The rest of the shareholders could lose out on the opportunity because of this wedge between management and ownership. And the greater the wedge, the worse this problem gets.
The running example I have going here (Sarah owns 1% of the company despite having barely more than 50% of votes) is particularly wild; under Facebook’s arrangement, for example, Mark Zuckerberg has the majority of votes despite minority ownership because he owns a bunch of shares that are allotted 10 votes per share. However, this agency problem occurs whenever a minority shareholder has majority control.
Academic studies have also found that dual-class companies have lower stock prices than comparable single-class companies, though intuitively at least some of this might be explained by the fact that investors prefer to have governance say and so might avoid dual-class companies, which reduces the investor demand for the shares of the company without necessarily saying anything about the quality of the company’s management.
Several studies have also found (see link from previous paragraph) that dual-class companies are on average more likely to make value-reducing acquisitions and pay executives more, which does seem like the structure is insulating sub-optimal management from market discipline. But one could argue that this could happen because of the aforementioned poorly regarded perpetual dual stock structure, and that if you sunset this effect will be ameliorated.
Many people from all sorts of vantage points are opposed to dual class structures and it seems like founders are the only people who are really pushing for them to exist. Institutional investors, both passive ones like Vanguard and active ones like private equity firms, unsurprisingly, hate dual class companies and want them all converted into single class. In 2013, Senator Elizabeth Warren, who is a left populist with a strong anti-Wall Street track record, released a letter opposing perpetual dual class structures on the grounds that it would leave management unaccountable to investors.
I’m persuaded that perpetual dual class structures are bad. But sunsetting seems more reasonable to me. There has been a long term decline in the number of public companies, and good reason to suspect that part of the reason why is because of public shareholder demands to maximize short term profit at the expense of the long run. A recent working paper has found that dual class firms innovate more and better, as measured by number of patents and patent citations, in the first first years post-IPO than single class firms.
If you take out the legitimate concern about perpetual dual class structures leading to management rot, what seems to me to be the core of the issue here is a fight between founders and investors to maximize their control. The vast majority of stock in the US is owned by the richest 10% of households, and deployed through wealthy institutional investors. From a society-wide perspective, it doesn’t seem to me like there’s a compelling reason to stack the deck in favor either the tech founders or the investors here, so just let them fight it out. If a founder can persuade investors to let him get away with additional control for a while, then I don’t see why the law or a stock exchange should stop them from having that arrangement.
One possibility here is a trade of sorts: allow dual class in order to induce more founders to IPO and lead their companies for longer, and potentially other inducements for companies to go public as well, in exchange for more pro-worker corporate governance policies. Two recent ideas that stand out as interesting to me include Elizabeth Warren’s proposal to have workers make up 40% of boards of corporations with more than $1 billion in annual revenue and Matt Bruenig’s proposal for a federal social wealth fund that would pay dividends to every American.
Corporate governance has traditionally been conceived as a field in which negotiations occur between managers and owners, two classes of people who almost by definition are elites (pension funds and retail investors are a drop in the bucket here). I’m not intrinsically opposed to the founder-emperor pushing the investor a bit, but I want to see the concept deliver for more than just a few people at the top.