John Locke, in the 2d and 3d paragraphs of the “Introduction” to his 2d Treatise on Government: “On Civil Government” makes the radical claim that political power is only justified when it is for the benefit of the governed. He begins by denying that traditional or military hierarchical superiority is a good model for political power, than goes on to rest political power on the foundation of benefitting the governed:
To this purpose, I think it may not be amiss, to set down what I take to be political power; that the power of a magistrate over a subject may be distinguished from that of a father over his children, a master over his servant, a husband over his wife, and a lord over his slave. All which distinct powers happening sometimes together in the same man, if he be considered under these different relations, it may help us to distinguish these powers one from another, and shew the difference betwixt a ruler of a commonwealth, a father of a family, and a captain of a galley.
Political power, then, I take to be a right of making laws with penalties of death, and consequently all less penalties, for the regulating and preserving of property, and of employing the force of the community, in the execution of such laws, and in the defence of the commonwealth from foreign injury; and all this only for the public good.
What does the principle that political power must be “for the public good” to be legitimate mean for the legitimacy of democracy? First, it may be possible for a nondemocratic form of government to be legitimate if somehow it had strong institutional constraints that made it operate only for the good of the people. For example, if truly random-sample surveys were the foundation of authority rather than elections, and the survey organization were as incorruptible as the very best election authorities, the political power based on that survey data could be legitimate.
Second, democracy does not automatically yield legitimate political authority. Except in very rare cases, “the consent of the governed” is a myth because–almost always–a substantial minority voted against whatever candidate or policy was up for vote. The welfare of those who voted against a policy–or worse, were not allowed to vote at all, perhaps because of being underage or called “non-citizens”–needs to be taken into account, too, for the political power to be legitimate.
Let me try to be specific about a rule for limiting legitimate political power. It seems hard to me to justify the use of net-freedom-reducing government coercion for ends that lower net Utilitarian social welfare. Government action should either be freedom-enhancing, or raise Utilitarian social welfare, otherwise that action is illegitimate. The contrasting idea that will of the majority constitutes “consent of the governed” and can justify whatever the majority wants within a wide, wide range of actions seems quite questionable.
Of course, where there are disputes about whether something is freedom-enhancing and whether it raises Utilitarian social welfare, there has to be some decision-making mechanism, and democratic choices deserve some deference. But, for example, where it is clear to everyone who looks closely, that by a coercive action of government, members of the middle class are being benefitted a little each by hurting the poor a lot, I don’t see the legitimacy.