Paul Yowell’s Constitutional Rights and Constitutional Design is a superb book, with a kind of lapidary intelligence. I think it advances the conversation in two ways, methological and substantive. Methodologically, it pursues the comparative-institutional method, conducted in an empirical rather than idealized register where empiricism is appropriate, as the subtitle suggests (“Moral and Empirical reasoning in Judicial Review”). There is no trace here of a standard professional deformation in conversations about judicial review — Harold Demsetz’ “nirvana fallacy,” which takes the form of comparing idealized courts with a realistically jaundiced view of legislatures, or for that matter the reverse (although the latter case is far less common among the knowledge class). Substantively, Yowell gives us an admirably clear account of central tradeoffs in constitutional design. Let me discuss some of these in turn.
(1) Information and bias
Yowell is excellent on the basic tradeoff between information and bias in the setting of legislatures and courts. Judicial procedures are designed to ensure equality of inputs, a form of evenhandedness. This is basically a leveling strategy of institutional design: no party has an advantage in presenting information to the judges. Moreover, the lack of any electoral connection on the part of judges gives them a remoteness from current politics and thus a kind of impartiality.
The price of this evenhandedness and remoteness, however, is a relative dearth of facts and tacit knowledge. The very political insulation that pushes the judges’ competence up also tends to deprive them of information, which pushes it down. By contrast, legislative fact-finding uses a relatively unstructured hybrid process, part adversarial and part administrative or inquisitorial, which generates more information overall at the cost of greater bias.
(2) Constitutional evolution vs. constitutional innovation
If I have a complaint about Yowell’s comparison of legislatures and courts — one can’t just praise everything — it is that it is insufficiently dynamic. There are two aspects to this.
First, one needs to compare legislatures and courts as the relevant environment changes, either informationally or in terms of actual underlying conditions. For exactly the reasons Yowell gives — the (partial) empirical dependence of so many constitutional questions — one cannot assume a fully static environment. What then becomes relevant are a class of positive models of the relationship between legislation and adjudication as the environment changes. In these models, judicial decisions can adjust constitutional rules incrementally, but statutes can innovate more rapidly and completely when there is an abrupt change of circumstances or a large-scale problem top be addressed. Once the fixed cost of enacting a statute has been paid, there is no extra cost to the legislature of enacting a new rule that is dramatically different than the old rule. Judges, by contrast, face increasing marginal costs for more dramatic departures from precedent, which implies that in a system of precedent incremental change will be the norm.
Given this, as the economists Ponzetto and Fernandez observe in the Journal of Legal Studies, “the fundamental trade-off is between evolution towards efficiency when the social optimum does not change, and rapid legal innovation when it does…. [S]tatutory intervention enables the legislature to adapt the law not only to the novel requirements of society, but also to its own biases, or the preferences of the special-interest groups that influence it. This price is only worth paying when social change in sufficiently intense.”
Of course models are only useful when they are useful. However I think this is a decent lens for understanding episodes of large-scale constitutional innovation in both the USA and (perhaps) the UK. Very tentatively, and subject to the correction of UK lawyers, I think of the Act of Settlement, the Parliament Act 1911, and the Human Rights Act as innovative constitutional statutes that take place at a kind of scale, and with leaps through the space of the possible, that it is hard to imagine judicial decisions replicating.
(3) Sentinel effects vs. epistemic moral hazard
The final tradeoff I will mention involves sentinel effects and epistemic free-riding or moral hazard. This is also an aspect of the difference between dynamic and static comparisons, involving the endogenous interaction between legislatures and courts. I didn’t see much about this in the book.
One possibility here is a form of epistemic moral hazard. Legislators who anticipate constitutional judicial review may rationally invest less in gathering and processing information. Precisely because they know that judges will be trying to catch their mistakes, they may commit more mistakes. The free-riding might even be mutual, in a case of “after you, Alphonse.” Legislators and judges might, on this picture, both reach fewer right answers than would either institution acting alone. This is of course just one possibility, however. Another one is a sentinel effect: the anticipation of review may induce higher-quality work than would otherwise occur. Yowell’s book would have been even richer with more discussion of these dynamic possibilities.
But this is to complain that no book is perfect — a standard none of us could meet. Yowell’s book makes a powerful contribution to the debate over judicial review, which now can no longer be idealized and stylized in the old manner.
Adrian Vermeule is Ralph S. Tyler Professor of Constitutional Law at Harvard Law School