new establishment clause test

[Steffen N. Johnson]

Sandy Levinson writes:

>>McConnell-Berg offer the following distillation of the caselaw:

>>"Aid is constitutional under the Establishment Clause if (1) it serves a
legitimate governmental purpose, such as education, (2) it is allocated on
the basis of neutral, secular criteria that neither favor nor disfavor
religion, (3) it is made available to a broad array of beneficiaries, both
religious and secular, on a non-discriminatory basis, and (4) any religious
element is the result of the genuinely independent and private choices of
individuals and families."

>>This test seems ideally crafted to justify tuition voucher or tax credit
programs and the like.  I'm not clear, though, about programs like the first
one struck down in Nyquist, which involved grants for building maintenance.
After all, the "religious element" in St. Mary's was not "the result of the
genuinely independent and private cuoices of individuals and families."  Nor
is this the case in the FEMA case, in which one wonders if federal aid
must-can-must not be given to rebuild the stained glass window of St. Mary's
that was shattered by the Oklahoma bombing.  I *would* allow-require federal
money to be given St. Mary's if every other building in the vicinity is
being rebuilt.  I assume that Michael would as well.  So what happens to
criterion (4)?

>>As for (3), how important is the empirical evidence as to the "broad array
of beneficiaries"?  After all, in Nyquist the majority pointed out that the
substantial majority of recipients under the program were church-related
schools.  Whereas Marshall in Witters adopted what Doug Laycock has
sometimes called the "little bit" theory and pointed out that only a small
portion of the aid program there at issue was used to pay for enrollment in
religious schools.  (This is certainly one of the major justifications of
rebuilding St. Mary's stained glass window.)  Is it enough that a program be
written in formally neutral language to pass muster, or do the numbers
matter?  (This is, incidentally, at issue in the Kiryas Joel replay, where
the New York legislature's rewriting of the statute to be "general" is under
attack on the ground that it *really* ends up applying either *only* to
Kiryat Joel or (is this, constitutionally speaking, the same thing?) to
Kiryat Joel and extremely few other areas.)

Sandy raises some valid points.

It seems to me the 4th element of the test could be improved by adding
"institutions," before "individuals and families."  That is, they too can be
legitimate recipients of government aid that is generally available under
secular criteria  (Kendrick).  Presumably this was left out because of the
Court's history with the direct/indirect distinction, but if the Court were
to adopt the McConnell-Berg test it seems that the distinction would no
longer be relevant.

As to the 3d element, i share Sandy's concern on empirical evidence.  How
broad is broad?  Brennan's statement in Texas Mo. that, so long as aid is
conferred on "a wide array of nonsectarian groups as well as religious
organizations in pursuit of some legitimate secular end," there's no primary
effect problem leaves the question open as well.  How wide is wide?  I would
propose some sort of equal protection approach, like Yick Wo.  Where aid
disproportionately assists religion over nonreligion, the burden shifts to
the government to show that it's not a gerrymander/intended to discriminate.

Steffen Johnson
Mayer, Brown & Platt
Chicago