The A to B of dual share structures

Dual-class share structures are pretty much frowned upon by the JSE and are not found in abundance on the local bourse.

There are just a few dual share structures left — eMedia Holdings, Brimstone Investment Corp and Rex Trueform (plus its holding company, African & Overseas Enterprises).

But go back a few decades, to a time when business owners and founders were a tad more fixated on control, and there were plenty such structures on the JSE. Some involved not only dual shares but pyramid holding structures that would fortify any underlying company from a hostile takeover.

Pikwik, the holding company for Pick n Pay, was well known. The old Rembrandt Group involved three other listings. Liberty Life was also protected by layers of control. Seardel, the clothing and textile manufacturing company, was shielded by a dual share structure as well as a holding company structure — which may have contributed to the near-obliteration of value in that company.

The concept is fairly simple. A dual-class share structure is where you have A shares that carry one vote per share, or what most investors would classify as the standard voting regime for a public company, and B shares that carry 100 votes per share. In South Africa it was more common to have ordinary shares and N shares, the latter carrying no voting or limited voting power.

Dual-class share structures can be a double-edged sword. On the one hand you get founders such as Warren Buffett or Mark Zuckerberg who retain effective control to execute strategy over long periods, free from the machinations of proxy wars or short-term-oriented shareholder activism.

However, the amount of shareholder activism observed across equity markets does support the notion that, in certain instances, shareholder activism is necessary to enhance shareholder outcomes for all.

Not all shareholder activism is short term in nature, and there are many examples where activism has brought about long-term value in improving strategies to the benefit of all shareholders. The existence of dual-class structures can frustrate such shareholder activism or render it moot altogether.

One of the many problems with dual-class share structures is that they can concentrate power in the hands of a few shareholders or management, and allow for companies to be run in the interest of those rather than for all shareholders — maybe not explicitly, but implicitly, as the controlling shareholders know there is little to no check on their voting power given the dual-class structure.

Dual-class share structures are a feature of the technology sector, and the venture capital and private equity worlds. However, once a company transitions to the public markets, there is a case to be made that such structures are not the most effective capital architecture or in the best interest of all public shareholders.

When companies come to the public markets, they are knowingly coming to a marketplace that gives them enhanced liquidity, capital-raising ability and a profile, among other benefits. A potential sacrifice for these benefits could be to collapse or have time limits on dual-class share structures.

Taking a modern example, Airbnb came to market with such a structure in which the A share has one vote and the B share 20, creating what is known as the difference wedge.

The difference wedge is calculated as the difference between the percentage of total voting rights and the percentage of total equity ownership. The difference wedge quantifies the level of divergence between voting rights and economic rights created by the dual-class structure. At its most extreme — a difference wedge of 100%, for example — shareholders of the super-voting share class would control all the voting power but hold no equity or economic interest in the company.

Dual-class share structures are a feature of the technology sector, and the venture capital and private equity worlds

In Airbnb’s most recent proxy statement, there are 433,034,896 A shares and 185,611,689 B shares in issue. Airbnb’s three co-founders, Brian Chesky, Nathan Blecharczyk and Joe Gebbia, collectively own 7,231,018 A shares, or have an economic interest of 1.67% in the company. However, the trio collectively own 163,885,381 B shares, or 88.3% of all B shares.

Now, given that the B share carries 20 votes for every share, the three co-founders effectively speak for 79.2% of the votes when combining the votes on their A and B shares regarding any AGM item. So good luck to any activist trying to bring about change there.

A sunset clause will have Airbnb’s dual-class share structure collapse all B shares into A shares 20 years after the IPO date (December 2020). Sunset clauses on dual-class share structures have become common in IPOs coming to market in the US in recent years; however, some still come to market without a sunset clause.

Most sunset clauses are time bound, with the Council of Institutional Investors recommending a sunset period not exceeding seven years from the IPO date. However, as evidenced by Airbnb, this is not a mandated listing rule, and tech giant Meta has no sunset clause in its dual-class share structure.

Sunset clauses can also be triggered by the achievement of specific market capitalisation, revenue or profitability metrics, or the exit of the owner of the super-voting share class from the company, or that person’s death.

Overall, sunset clauses, in some form or another, are beneficial, as they provide a degree of certainty for the end of the structure for the holders of both sets of shares.

You only need to look at the furore in 2019 and the pushback from ordinary shareholders when Shoprite floated the idea of buying back deferred shares held by Christo Wiese that carry large voting rights but no economic rights.

The plan would have had Shoprite pay R3.3bn for the deferred shares and allowed for the collapsing of the structure. Ordinary shareholders rejected the plan before it was ever made a special resolution at an AGM. The deferred shares thus remain in place today.

A few sunset clauses are attached to the Shoprite deferred shares, but none of these is time based, and the other sunset clauses have not been triggered since 2019. So the deferred shares and the dual-class share structure just sit there in the background as an overhang on the capital structure.

The latest local example was Institutional Shareholder Services (ISS) recommending that shareholders vote against Johann Rupert and for re-election to the Richemont board at the AGM on September 10. The Rupert family owns 6,418,850 Richemont A shares and 537,582,089 B shares, representing 10% of the equity of the company and controlling 51% of the company’s voting rights. The family owns all the B shares, which carry 50% of the total voting rights, and the ISS’s recommendation did not lead to anything.

On balance, IM is not a fan of dual-class share structures for public equities. However, if they’re being implemented for public equities, clear guardrails are necessary, including stock exchange-mandated listing rules that require, at the very least, a time-based sunset clause and a sensible voting ratio.