Search

We study the introduction of a new control-enhancing mechanism in Italy, a country still characterized by family-controlled firms but with an increasing importance of institutional investors. Since 2014, Italian firms have been able to adopt loyalty shares, which allow a double voting right if shares are continuously held for at least two years. We find that about 20% of listed firms have introduced loyalty shares, and family-controlled firms are the most likely adopters. Loyalty shares neither anticipate acquisitions, nor equity issues by the adopting firm. Instead, they allow controlling shareholders to reduce their equity stake without losing control. We report no evidence of an adverse wealth effect both at the adoption and in the years following it. As expected, institutional investors vote against the introduction of loyalty shares. Yet, they do not reduce their holdings afterwards, as incremental governance costs are outweighed by the superior performance of adopting firms. Overall, our evidence suggests that bolstering family control is the main effect of the introduction of loyalty shares.

The high separation of ownership from control achieved through the concurrent use of non-voting shares and stock pyramiding could favor acquisitions made to increase private benefits of the controlling shareholders rather than all shareholders’ wealth. A standard event study methodology is carried out on three different samples of Italian acquisitions during the 1989-1996 period in order to test this hypothesis. We find evidence that a worst market reaction characterizes acquiring firms with a higher separation of ownership from control, while more value-enhancing transactions are undertaken by those smaller in size and with higher prior-performance. An entrenchment effect seems to determine a significant U-shaped relationship between the market reaction and the ultimate shareholder ownership. When the sample is restricted to acquiring firms with a dual class equity structure we find that non-voting shares report significantly negative excess returns in contrast to significantly higher positive returns for voting shares. Such evidence seems to indicate that the average acquisition has been overpaid, as suggested by the negative market reaction of the non-voting shares, while it was expected to lead to higher private benefits to the majority shareholders, as suggested by the revaluation of the voting shares. Finally, the market reaction to acquisitions made within pyramidal groups seems to indicate that the price is set so as to transfer wealth towards the companies located at the upper levels, where majority shareholders own greater fractions of the companies’ cash flows.

Called to vote for a reduction in their dividend privileges, the non-voting shareholders of one of the largest Italian firms appeared to expropriate themselves and favor the voting class of shares. However, what at first seemed to be self-expropriation turned out to be self-interest, as soon as the media coverage, the voting decision, and the dual-class ownership of 36,361 shareholders were investigated. Firstly, our new anecdotal evidence shows that the media can mislead investors and help the approval of harmful proposals when they are not independent of the companies they report on. Secondly, most of institutional investors who cast the “for” vote were in conflict of interest either because of ownership ties with the controlling shareholders or because they held both classes of shares. Finally, we find that both retail and institutional dual-class shareholders are more likely to vote for self-expropriating one class of shares if they benefit from the other class in their portfolios. When conflict of interest in dual-class voting is not regulated, dual-class ownership protects from dual-class wealth transfers but favors the approval of the operations it is meant to hedge from.