In mid-2005 few investors in Porsche would have guessed that only 18 months later their firm would be the dominant shareholder in Volkswagen. Then again, it is none of their business, because they have little say over the sports-car maker’s managers or strategy. The Porsche and Piëch families have total voting control over the German firm, despite owning only half of its equity. The publicly traded shares, providing the other half of Porsche’s capital, have no right to vote.
After a recent European ruling, the “VW law”, which prevents any shareholder from exercising more than 20 percent of the carmaker’s votes, is likely to be repealed. If so, the Porsche and Piëch clans will have control of VW, because Porsche holds 27.4 percent of the voting shares. In shareholder democracies, as in political ones, poor turnouts mean that about one-third of the votes can grant control of a company.
That is not the full extent of the families’ influence. VW has a 20 percent stake in Scania, a Swedish truck and engineering group, which has two classes of voting shares, one with ten times as many votes as the other. Thanks to these super-voting shares, VW has 35.3 percent of the votes in Scania. In other words, the Porsche and Piëch families will also control Scania, thanks to an economic stake in the firm of a mere 2 percent.
Almost since the first company listed on the first public market, capitalists have sought to raise money without sacrificing control. Porsche demonstrates two techniques that families and founder-entrepreneurs commonly use. First is a cascade of companies — a ploy particularly favoured by Italians — to gain control of a larger group. Each stage of the cascade amplifies the capital governed by the tiny company at the top. The second, on flotation, is to issue shares to the public with vastly inferior — or no — voting rights. Both are legal.
Porsche is an extreme example of how a family can wield great power by harnessing large amounts of other people’s money. But it is not unusual. An American study concluded that 6 percent by number and 8 percent by capitalisation of quoted companies in America had dual-class shares in 2002. Only two-thirds of large European companies rigorously apply the principle that one share should command one vote, according to a Belgian study in 2005.
The European Commission, eager to create an open market for corporate control, is studying whether it should act against them. Israel stands out for having made the even allocation of votes a requirement of listing on the stock exchange. It is a tricky balance to strike. Liberalism argues that investors and owners should be free to enter into contracts as they wish. On the other hand, sheltering families that govern a company badly impedes restructuring. Isn’t that bad economics?
Sweden is the European country with the most dual-class shares: some two-thirds of its listed companies have them. The Wallenbergs, one of Sweden’s most famous business families, which controlled Scania before selling most of their stake to VW, could give a master-class in their use. Their foundations own super-voting shares in Investor, a quoted investment company that they control. Then there is another layer of special shares. Investor has helped develop Swedish companies such as Electrolux, a white-goods maker, and Ericsson, a telecoms-equipment company. Most of these firms have shares with disproportionate voting rights, a slug of which Investor owns.
Some of the best-known North American companies use dual-class shares. The Ford family may no longer own the company — it has a slender 3.75 percent of the shares — but nobody can doubt who is in control. When the firm went public in 1956 the family’s shares were converted into a special class that is guaranteed 40 percent of the voting power, no matter how many ordinary shares are in issue. The late Alex Trotman, who held the top job at Ford in 1993-98, used to say that to be successful, the boss of Ford needs to manage the family as well as the company.
Dual-class shares are no hangover from the past. When Google went public in 2004, its founders, Larry Page and Sergey Brin, insisted on a dual-class share structure. Google sold the public new A shares with one vote per share; its existing shares, mostly held by Mr Brin and Mr Page, became B shares with ten times as many votes per share. With around one-fifth of Google’s total equity, the founders still command about three-fifths of votes, and thus retain control.
Dual shares are designed to assure investors that a known group of managers will control the company indefinitely. They shield the long-term vision and strategy of founders from stockmarket forces. At Google, they are in effect founded on the idea that Mr Brin’s and Mr Page’s judgment is better than that of investors — and will remain so until the firm’s founding geniuses (or their heirs) decide differently. The duo clearly spelt out why they wanted such a structure, in their “Owner’s Manual” for prospective shareholders:
[It] will make it harder for outside parties to take over or influence Google…and easier for our management team to follow [a] long-term, innovative approach…[and] leave our team, especially Sergey and me, with increasingly significant control over the company’s decisions and fate.
Such share structures are often set up just before a company’s flotation, so you can hardly accuse owners of foisting them on unsuspecting investors. Investors pay less for shares in a company with dual-class shares than for shares in an otherwise identical company that endows each share with one vote. Their rights are weaker in a firm with super-voting shares, because they can not kick out the family. Whatever their drawbacks, therefore, dual-class shares are “fair”.
That, at least, is the theory. Sometimes, though, it seems hard to believe. Italian entrepreneurs use convoluted cascades of companies (also known as scatole cinesi, or Chinese boxes) to achieve the same end as dual-class shares. The company at the top of the cascade, the one that calls the shots, is usually private. This firm’s controlling shareholder has enough votes to secure an absolute (or de facto) majority at shareholders’ meetings at each level in the cascade. In effect, Chinese boxes remove control of the quoted company from its public shareholders, because the majority is fragmented into a series of minorities spread over the cascade.
This is the mechanism that the Agnelli family uses to control Fiat Group, a quoted company and Italy’s largest industrial group (see chart 1). The family’s cassaforte (literally, its strongbox), Giovanni Agnelli & C, a limited partnership, controls a 30 percent stake in Fiat through a cascade of two quoted companies, IFI and IFIL. This is enough to give it control of Fiat’s annual shareholder’s meeting. (For instance, at the meeting in 2006 less than half of those entitled to vote did so, thereby in effect doubling IFIL’s votes.)
A vivid illustration of the power of Chinese boxes came in 2003 when Fiat had needed a capital increase to prop up its ailing balance sheet. It was a lesson in how to turn €250m — which the Agnellis’ limited partnership injected into IFI to pump-prime the cascade of capital increases — into €1.84 billion, which was the amount Fiat raised. In other words, the minorities in the three quoted companies in the cascade provided the majority of the money.
There is a further twist. John Elkann, a grandson and main heir of Gianni Agnelli who ruled the Agnelli’s empire until his death in 2003, pulls the strings at Giovanni Agnelli & C. This is because he controls a company called Dicembre, which has 31 percent of the votes in the limited partnership. Tens of other descendants of Fiat’s founder, none of whom speaks for more than 5 percent, own nearly all the rest of shares. This makes Mr Elkann’s position unassailable without an unlikely alliance against him in the rest of the family. The extended Agnelli family, Mr Elkann apart, provides the capital but has no real say in what happens to it. Meanwhile, the man who controls Fiat owns a mere 3.5 percent of its equity.
Highly Preferred Stock
The economic costs of such schemes are complex. One attraction of family-run companies lies in their managers’ relationship with shareholders. In most companies, managers have an interest in creaming off profits through, for instance, lavish expenses and vanity projects, rather than paying them out to the owners. By contrast family managers are the owners. The more of a family’s wealth that is tied up in the business the closer the alignment of managers’ and shareholders’ interests. Deterring family-run companies from listing in the public markets or encouraging them to take on debt rather than raising equity might therefore be inefficient.
Super-voting shares, however, undermine this argument. Dual-class shares and cascades tempt the powerful to benefit themselves, rather than to pay money out to everyone equally. An extreme example, American prosecutors allege, is Hollinger International (now called Sun-Times Media Group), once one of the world’s largest newspaper groups, with titles including the Daily Telegraph, a British daily, and the Jerusalem Post. Lord Black, its former chief executive, is standing trial this week on charges, which he denies, of siphoning tens of millions of dollars out of the firm.
The group’s capital structure was a factor in the alleged frauds. Through a chain of companies, Lord Black indirectly controlled around 20 percent of Hollinger’s shares. But he had around 70 percent of the votes; each of his B shares had ten times the voting rights of A shares held by other investors. Nobody could stop Mr Black from packing Hollinger’s board with prominent and well-paid contacts, who seemed not to have asked too many questions. Mr Black had no counterweight, because nobody could fire him.
Despite this increased risk, calls for the reform of super-voting shares inside companies are sporadic. Perhaps this is because a company’s investors know they have no chance of success without some negotiating leverage.
But when their position is strong, investors do force change. Part of the defence of Arcelor, then the world’s number two steel group, against a hostile bid in 2006 from Mittal Steel, the world’s largest steel group, was to criticise the bidder’s governance. Arcelor’s main target was the disproportionate voting rights of Mittal Steel’s B shares, owned by Lakshmi Mittal , the controlling shareholder. To help his bid, Mr Mittal agreed that all shares in the new firm, Arcelor Mittal, would have equal economic and voting rights.
Another governance risk is in companies that look to the second generation for management. The worry is that Junior has got to the top thanks to Pops. This leaves no room for talented people with the wrong surname. American research suggests that second-generation companies tend to fare worse than others do.
The Lagardère family in France has found the ultimate way to pass on the quoted family firm, a magazines-to-missiles group, of which it controls just 7 percent. This is a French limited partnership (société en commandite par actions). It has two types of partners: general partners and limited partners — the shareholders. The general partners have exclusive power to appoint the managing partners (gérants), who are responsible for running the business. (The gérants are appointed for a six-year term, which may be renewed.)
Lagardère’s general partners are Arnaud Lagardère, son of the late Jean-Luc Lagardère, who set up this structure, and Arjil Commanditée-ARCO (ARCO), a company that the Lagardères control. Unsurprisingly, the general partners have appointed themselves as Lagardère’s gérants. The firm’s supervisory board, which the shareholders appoint, cannot dismiss Mr Lagardère as general partner unless he first decides to sack himself as gérant. And ARCO’s position as general partner is untouchable so long as the Lagardère family controls it.
Doing lip service to the demands of corporate governance, the company’s articles provide that shareholders approve the appointment (or reappointment) of a gérant. Ultimately, however, shareholders can refuse a candidate only by a majority of at least two thirds. If they do, Mr Lagardère and ARCO assume responsibility for running the business in their capacity as general partners. In other words, the Lagardère family can run the business for as long as it likes. It would lose control only if someone acquired more than two-thirds of the firm’s shares and then applied successfully to the courts to change the statutes which give it control. Another of France’s main companies, Michelin, the world’s largest tyre-maker, is also a limited partnership of this sort.
Other French entrepreneurs or families have a less extreme way of entrenching their grip. PSA Peugeot Citroën, a quoted car company, grants double-voting rights to shares held in the same name for at least four years. The main beneficiaries of this rule are the Peugeot family: their 30 percent stake commands 45 percent of the votes.
This system was supposed to encourage loyalty in shareholders and to punish speculation. But it has mainly benefited large family-shareholders, such as the Arnaults, owners of just under half of LVMH, a luxury-goods group; their rivals, the Pinaults, who control PPR; and the Bouygues, at Bouygues SA, a construction-to-telecoms conglomerate. Martin Bouygues, one of the sons of Bouygues’s founder, runs this company, just as François-Henri Pinault, one of Mr Pinault’s sons, runs PPR.
The share-price discount of dual-class stock reflects the risk that the lack of control may lead to abuses. The evidence on how often that happens is mixed. For instance, a study of firms that switched from giving each share one vote to a dual-class structure concluded that on average the change was good for shareholders. Another of a sample of dual-class companies going public found that the share price of these companies outperformed comparable firms with single-class shares.
But the most comprehensive piece of work on the subject, a recent study by Wharton School, Stanford University and Harvard Business School of all dual-class firms from 1995 to 2002, concludes that shares with voting rights in excess of economic rights are bad for other shareholders, and that the bigger the gap the worse the damage. “What you would really like to do is give managers a lot of economic ownership in a company, but no votes, which is the opposite of what you see in most dual-class companies,” says Andrew Metrick, one of the study’s three authors. Using data from the study, The Economist has compiled a list of large dual-class companies, ranked by market capitalisation (see chart 2). Big media companies dominate the list and, as at the New York Times Company, occasionally attract hostility. (Full disclosure: The Economist, albeit a private company, has four classes of shares.)
Dominant families are one thing. Managers who wield control even though they have no particular long-term interest in the company are more alarming still. The depository-receipt system in the Netherlands splits economic rights from voting rights and thereby, in effect, enables managers to have voting control without owning any share capital. When a company with this system issues shares, it does so to a voting trust, which retains the voting rights. The trust then issues depository receipts, which have economic rights mirroring those of the underlying shares.
The depository-receipt holder can tell the trust how to vote. But few holders give instructions, so the trust dominates the voting. “The voting trusts are a vehicle to get a ‘yes’ vote on every item on the agenda,” says Peter Paul de Vries, head of the Dutch Investors’ Association (VEB). In 30 years of observing shareholders’ meetings, Mr de Vries has not seen a trust vote against managers.
Fewer Dutch firms use the system these days. But the 10 percent that still do include ABN Amro (under siege from hedge funds) and ING, two banks, and Unilever, a consumer-goods giant. ING claims, as do others, that depository receipts prevent a minority from taking control. But the VEB is fiercely critical of the system, mainly because it deprives shareholders of influence. “A dictator will rarely do away with his own dictatorship,” says Mr de Vries.
The Dutch system is perverse, but other systems of control are more finely balanced. Families understandably want to keep control. On the other hand, there is a whiff of inconsistency about a family seeking public-market funds, but having such little faith in investors that it requires protection. Moreover, what looked like a good idea at the time of the company’s listing might not seem so clever five decades on. At Ford, Henry’s memory has been eclipsed by his descendants’ endless struggle with decline.