In deciding whether to invalidate the Religious Freedom Restoration Act (RFRA) in City of Boerne v. Flores, 117 S. Ct. 2157 (1997), the Supreme Court faced an unhappy dilemma. On the one hand, it had before it an immensely popular law. Anytime an individual could show that a federal, State or local law "substantially burdened" a religious practice, RFRA came into play -- compelling the government to justify the burden by running the gauntlet of strict scrutiny. Senators Hatch and Kennedy signalled the bipartisan support for the law by cosponsoring it, and ultimately just three members of Congress opposed the law's passage. The spark that prompted RFRA, moreover, was the Court itself. As the language of the statute proclaimed, RFRA was an attempt to "restore" free exercise law to its alleged prior status by "overruling" the Court's 1990 decision in Employment Decision v. Smith, 494 U.S. 872 (1990). Add to these pressures the normal deference that the Court properly gives the legislature in exercising its policymaking prerogative, and it was fair to doubt the Court's inclination to strike the law.

At the same time, the Court had before it a unique piece of legislation. In enacting RFRA, Congress relied on section 5 of the Fourteenth Amendment, which only gives it "power to enforce, by appropriate legislation, the provisions of this article." Unable to establish the normal predicate for a section five law -- that the States had been, or currently were, violating their citizens' constitutional rights -- Congress claimed that section 5 authorized legislation that protected the "values" of the free exercise clause or that established a "statutory buffer zone" around the clause. But this vision of section five power had no stopping point, particularly when applied to the utilitymen of the Constitution -- the equal protection and due process clauses. RFRA's breadth of application also made it an envelope-pushing exercise of section five power. It restricted all laws, whether federal, State or local, and applied to any law that had been passed or that ever would be passed in the future. RFRA, in short, had all the appearances of a constitutional amendment in section five clothing.

In choosing between upholding an exceedingly popular law and maintaining a limit on congressional power, the Court properly chose the latter. In doing so, Justice Kennedy's carefully-reasoned majority opinion made two essential points about the limits of section five authority. First, the power "to enforce" the Fourteenth Amendment is only a preventive or remedial power. Short of a constitutional wrong, there is nothing Congress may constitutionally remedy. Second, section five laws must be proportional in nature, requiring a "congruence and proportionality" between the injury that must be remedied and the means for doing so.

The Court concluded that RFRA in the end met neither requirement. It did not rest on any finding of State action abridging free exercise rights; it merely tried statutorily to change the meaning of the free exercise clause itself. And, in view of the wide applicability of RFRA, it was exceedingly doubtful whether even the most egregious of the free exercise wrongs could have justified such a far-reaching exercise of federal power. Conspicuously missing from the dissent was any disagreement with Justice Kennedy's interpretation of section five; indeed, one of the dissenters, Justice O'Connor, wholeheartedly agreed with it. The three dissenters instead opposed the result on the basis of their previously stated disagreement with or ambivalence about Smith itself.

The Boerne decision is entirely correct. Section five combines a broad power (to pass "appropriate legislation") with a broad limitation on that power (to do so only when "enforc[ing] the . . . provisions of this article"). The verb "enforce" is an important one and a confining one. While it permits Congress to compel obedience to something -- here, to require the States to adhere to the prohibitions contained in the equal protection and due process clauses -- it does not give Congress the power to "define" constitutional rights or to "expand" them. As Justice Kennedy properly recognized, section five is a remedial power, not a definitional one: the Court remains the interpreter of the Constitution; Congress simply retains remedial power to ensure that constitutional violations cease or are cured.

The decision also strikes a welcome blow for States' rights. Together with Printz v. United States, 117 S.Ct. 2365 (1997) (which invalidated a portion of the Brady Bill), the case serves as a reminder that federalism remains a vital, not a quaint, component of the Court's jurisprudence. It now ought to be clear that preserving the Constitution's structural delegations of power matters to this Court, whether due to the importance of securing a separation of powers among the three branches of the federal government or due to the importance of preserving the dual sovereignty of the federal government and the States.

The failure to invalidate RFRA would have been devastating to these views and to the underlying end of federalism itself: the protection of individual liberties by dividing government and limiting its aggregation in any one body. Left intact, the RFRA version of section five power would effectively have transformed the federal government from one of limited powers to one of unlimited powers. For if Congress were allowed to enforce constitutional values, as opposed to constitutional violations, it is hard to fathom any limiting principle to its power. Congress could federalize all law enforcement in the name of Fourth and Fifth Amendment values, all punishment in the name of Eighth Amendment values, all speech-related regulations in the name of First Amendment values, to say nothing of an assortment of family and social issues in the name of substantive due process values.

The only limits to this boundless theory of section five power would have been those that Congress was willing to place on itself. Yet only a most feeble version of federalism would depend on congressional self-restraint. It was precisely because the Framers doubted the ability of all government (whether local or national) to restrain itself that they carved up authority in so many different ways. In this respect, federalism is no less a countermajoritarian right than the first eight amendments to the U.S. Constitution. Both serve the goal of safeguarding individual liberty; they just do so in different ways.

Finally, as Boerne also reveals, federalism decisions with muscle can comfortably co-exist with a belief in the "passive virtues" of judicial restraint. Whenever the Court strikes down a federal law on structural grounds, the decision invariably prompts a charge of inconsistency. How can those who believe in judicial restraint, the argument runs, also believe in striking democratically-enacted laws in the name of federalism? The charge is a serious one, but it submits to a responsible answer.

A dispute between the States and Congress over the meaning of a structural guarantee creates a zero-sum situation, in which either a State or a federal lawmaking prerogative must fall. The facts of Boerne illustrates the point. At issue was not just the validity of RFRA, but also the validity of many local laws -- for example, the zoning regulations of the citizens of Boerne, Texas at issue in the case and the generally-applicable prison regulations of virtually all State correction facilities. From Congress's perspective, it had authority to pass RFRA and to nullify these local exercises of lawmaking authority. From the States' perspective, they had authority to regulate these areas free from congressional overreaching. The national government's gain in these types of cases invariably becomes the State's loss, and vice versa. The Court had no choice but to invalidate a democratically enacted law -- either, under RFRA, the City of Boerne's zoning laws or, under the Constitution, RFRA itself.

Nor would it be appropriate to give greater deference to one sovereign's reading of the Constitution over the other in this situation. Because the Constitution does not treat State and local officials as second-string readers of the Constitution, they (no less than their federal counterparts) are entitled to respect in interpreting their oath to obey the Constitution. Any presumption that Congress has correctly interpreted the Constitution necessarily slights the States' ability to interpret the same document. One or the other interpretation must prevail, and a Court that purported to be exercising a judicial restraint in upholding a federal law in this setting could only claim to be speaking half the truth.

Far from upholidng an "ersatz federalism," as Professor McConnell charges, the Court's decision in City of Boerne ensures that States will be free to strike their own balance in accommodating religious objections to neutral, generally applicable laws. Congress's attempt to override the States in this area, and to impose a uniform national rule that exceeded the requirements of the Free Exercise Clause, was properly invalidated as wholly in excess of Congress's limited powers.