Dual-class shares have long been debated in corporate law, but it seems that commentators have not yet reached a conclusion. It is sometimes argued that one-share-one-vote system is in most cases socially more desirable than the dual class system. The controlling shareholders with multiple-vote shares are likely to pursue private benefits inefficiently, especially when the corporate law is not good enough to deter them. Such observation, however, does not necessarily lead that mandatory one-share-one-vote rule is socially preferable. Dual-class shares are likely to have several value-enhancing effects by coordinating various interests in many contexts such as IPO, strategic alliance, and joint venture investment. Mandatory enforcement of one-share-one-vote rule just eliminates suchp ositive effects. Even if the dark sides of dual-class shares should be fully taken into account, the prohibition would not be an answer in most situations. This paper joins the dual-class share debates with theories and empirical studies on corporate ownership structure and family firms, and suggests more modest approach, in which law should regulates the details of dual class shares. In fact, dual class shares, as an option given to founders, might be socially desirable even in emerging markets where corporate law is not that good enough to effectively protect investors. The point is that the corporate law which could hardly prevent controlling shareholders from pursuing private benefit is also likely to fail as well to restrict agency costs associated with professional management. Legislators could not know which agency cost is larger. In those situations, therefore, dual-class shares might be an optimal solution. At the same time, however, this paper emphasizes that the details of this multiple voting rights should be regulated by corporate law. While current conventional theory employs simple dual-class v. one-share-one-vote dichotomy, this paper argues that more attention should be paid to the detailed provisions of dual-class shares. One of the most important features to be considered is that the multiple voting rights should be granted only to the founders of company (or equivalent members), and should not be transferred to others nor inherited by the next generations. In fact, dual class shares in the United States, Canada, and Japan often contain those provisions. Market-oriented scholars are likely to argue that such features can be achieved contractually with investors and thereby priced by stock market, but a handful of empirical studies demonstrate that the market could not be so efficient as to evaluate such details. This paper therefore proposes rather a safer approach where corporate law or listing rules should play a role to alert investors to pay attention to the specific provisions of multiple-vote shares.