by Adrian Vermeule. New York: Oxford University Press, 2011. 232 pages. Hardcover $35. ISBN: 9780199838455.

Reviewed by Brian D. Feinstein, Judicial Clerk, U.S. Court of Appeals for the Seventh Circuit


Adrian Vermeule’s The System of the Constitution imports concepts associated with systems theory into the study of political institutions and constitutional design. The central insight of systems theory can be gleaned from the cliché “the whole is greater [or less] than the [dynamic] sum of its parts.” Add the words in brackets to this timeworn phrase, apply it to the study of political systems such as courts and legislatures, then apply it again to the study of “systems of systems” – here, the U.S. constitutional order – and you have Vermeule’s thesis: Individual members aggregate into institutions, and institutions aggregate into systems of government, in ways that are complex, interactive, and sometimes counterintuitive; accordingly, analysts must take these dynamics into account in their assessments of political institutions.

Examples of system effects – where “what is true of the members of the aggregate is not true of the aggregate, or vice-versa” – are familiar to social scientists. For instance, Condorcet’s paradox of voting shows that aggregating individuals’ transitive preferences may lead to intransitive preferences of the group; Condorcet’s jury theorem and the miracle of aggregation both show the circumstances under which a group may be more accurate than its individual members in expectation; the prisoner’s dilemma shows how individuals’ utility maximization can make the group worse off; and the invisible hand of the market provides an example of the converse.

According to Vermeule, these problems are pervasive in American politics at two levels, since governance based on separation-of-powers and federalism principles involves the interaction of multiple institutions, and these institutions, in turn, are comprised of multiple interacting members. Failure to understanding these systemic interactions can lead observers to erroneously believe certain fallacies of composition – e.g., that if individual judges are politically biased, appellate courts will issue biased rulings – and fallacies of divisions – e.g., that all components of the political system must be democratic in order for the system to be considered democratic.

Moreover, Vermeule continues, since institutions gradually change as a result of changes in their memberships, and since the overall system changes gradually as a result of changes in its component institutions, analysts must also be attune to dynamic effects. For instance, the First Amendment promotes tolerance of undemocratic ideals. If these ideas gain popularity, their adherents eventually may elect officials who act undermine democracy. By this sequence, the First Amendment could prove to be self-undermining. Through a similar sequence, the expansion of voting rights [*402] could prove to be self-reinforcing, as newly enfranchised groups will elect representatives with a self-interest in protecting the these groups’ gains. Although these insights are not novel – the voting rights example has echoes of Schattschneider, and the potentially self-defeating nature of free speech has been noted in ten thousand late-night dorm room discussions; see also Post (1991) – Vermeule’s classification of these and other examples under the system-effects rubric is analytically useful.

Most significantly, this way of thinking promotes a framework for institutional design that Vermeule terms “second-best constitutionalism.” Applying this framework, institutional designers need not assume that optimizing the extent to which a system has certain desirable features necessarily means maximizing the degree to which each component institution in that system possesses those features. Instead, the various components of a constitutional order may have offsetting failures. That is certainly plausible in theory, but I am not persuaded that it accurately describes the U.S. system in fact. Consider the evidence that Vermeule offers:

The Senate favors small states; the Electoral College favors groups with influence in battleground states, which may or may not be small states; the administrative state favors groups who can organize to influence agencies and congressional committees; the prestige and power of the Supreme Court benefit the legal elites. . . . [U]ndemocratic power is, in a sense, democratically distributed (p.51).

But the diverse-masters outlook looks less rosy if one questions whether, say, “groups with influence in battleground states” and “groups who can organize to influence” public officials really are all that different. More generally, wouldn’t those interests that are able to move outcomes towards their ideal points in one institution also possess advantages to do the same in other institutions as well? Why would we expect the undemocratic nature of various institutions to cancel out? Instead of showing how the presence of multiple biased institutions can lead to these institutions’ biases cancelling out – rather than reinforcing each other – at the systems level, Vermeule seems to assume it.

This assumption is a shortcoming, but it is also an opportunity for future analysts and advocates. As a general matter, Vermeule’s insights should discourage piecemeal, institution-by-institution reform efforts and encourage an emphasis on how various institutions interact, with a particular focus on macro-level comparative statics following a reform to a single institution. (Interestingly, there is some evidence that the Constitutional Convention of 1787 adopted this big-picture approach. Rather than create separate committees to study each potential institution in isolation, the founders established a Committee of Detail to consider the optimal system of institutions and issue a single report on the subject to the floor; see Farrand (1937).) More specifically, the concept of second-best [*403] constitutionalism should motivate institutional designers to think like hedge fund managers: identifying the biases inherent in potential components of a democratic system and seeking to assemble a portfolio of institutions, in which the components are not positively correlated. It is easy to see how Vermeule’s theoretical insights could motivate such a project; in scope and style, The System of the Constitution evokes Robert Dahl’s A Preface to Democratic Theory or Mancur Olson’s The Logic of Collective Action; slim volumes that became staples of graduate seminar reading lists by shedding light on intuitive, but overlooked, aspects of political institutions and serving as catalysts for the development of large formal-theoretic and empirical literatures to flesh out the details of the authors’ broad strokes. I can envision The System of the Constitution taking a similar place in the canon.


Dahl, Robert A. 1956. A Preface to Democratic Theory. Chicago: University of Chicago Press.

Farrand, Max, ed. 1937 (rev. ed. 1966). The Records of the Federal Convention of 1787, vol. 2. New Haven, Conn.: Yale University Press.

Olson, Mancur. 1965. The Logic of Collective Action. Cambridge, Mass.: Harvard University Press.

Post, Robert C. 1991. “Racist Speech, Democracy, and the First Amendment” William & Mary Law Review 32: 267-327.

Schattschneider, E.E. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt, Rinehart and Winston.

Copyright 2013 by the Author, Brian Feinstein