Constitutional Review and the Selective Promotion of Case Results

Text-only Preview

Constitutional Review and the Selective
Promotion of Case Results
Jeffrey K. Staton Florida State University
A significant majority of the world’s constitutional courts publicize their decisions through direct contact with the national
media. This interest in public information is puzzling in so far as constitutional judges are not directly accountable to voters.
I argue that the promotion of case results is consistent with a theory of judicial behavior in which public support for courts
can undermine incentives for insincere decision making. In this article, I develop a simple game theory model that identifies
how case promotion is linked to judicial choice. Results of a simultaneous equations model estimating the Mexican Supreme
Court’s merits decisions and its choices to publicize those decisions by issuing press releases to national media outlets support
an account of constitutional review in which judges believe they can influence their authority through case promotion.

Scholarshavelongrecognizedthepoliticalvalueof scholarswithanintriguingpuzzle.Insofarasconsti-
good public relations. In presidential studies, the
tutional judges are not directly accountable to voters, it
ability to go public is considered among the chief
is not immediately clear why they would promote their
executive’s most powerful tools (Kernell 1992). Legisla-
resolutions among the mass media. Why a constitutional
tive research similarly finds an important role for public
court would promote some resolutions and not others is
strategies of issue promotion (Cook 1996). Even bureau-
even less clear.
cratic analysis suggests that agency concerns over public
I argue that the selective promotion of case reso-
information as shaped by the media may influence policy
lutions is endogenous to judicial decision making. In
implementation (Carpenter 2002). Yet when it comes to
particular, case promotion offers judges solutions to po-
judicial politics, scholars largely ignore the interest high
litical problems induced by the separation of powers sys-
court judges take in shaping public information.1
tem and the way the media covers constitutional review.
This omission is especially striking for two reasons.
I develop this argument by extending Vanberg’s (2001)
First and foremost, high court judges engage in pub-
model of inter-branch relations, which integrates sepa-
lic relations activities. While John Marshall’s defense of
ration of powers and legitimacy theories. I then test a
the United States Supreme Court’s McCulloch decision
set of empirical implications derived from the model us-
is well known (Gunther 1969), modern constitutional
ing an original data set on all constitutional decisions of
court members frequently speak and publish work on
the Mexican Supreme Court between January 1997 and
judging (Barak 2002). Nearly every constitutional court
December 2002 in review of state or federal laws. Given
in the world maintains a Website through which they dis-
the proposed link between judicial decision making and
tribute information on case results, and like Marshall, a
case promotion, I seek to simultaneously explain the
number of these courts selectively promote their deci-
Supreme Court’s merits decisions and its choices to publi-
sions among national media outlets.2 Second, the nature
cize the results of those decisions by issuing detailed press
of constitutional court appointment and tenure, which
releases to print and television media. In what follows,
separates these judges from the voting public, presents
I first review how selective case promotion fits within a
Jeffrey K. Staton is assistant professor of political science, Florida State University, Tallahassee, FL 32306-2230 ([email protected]).
I would like to thank Gordon Hanson, Chris Woodruff, Jeffrey Weldon, Christopher Reenok, Steve Wuhs, Kathleen Bruhn, John Carey,
Randall Calvert, Lee Epstein, and three anonymous reviewers for their helpful comments and suggestions.
1Davis (1994) is a notable exception.
2An exhaustive search of all European and American high, constitutional, or international court Websites suggests that 71% of these courts
issue press releases summarizing at least some of their resolutions.
American Journal of Political Science, Vol. 50, No. 1, January 2006, Pp. 98–112
C 2006, Midwest Political Science Association
ISSN 0092-5853

theory of judicial choice. I then develop the theoretical
when the public is able to monitor the reactions of their
model and identify its empirical implications. In the sub-
representatives (2001, 347). In the sense that the public
sequent section, I discuss my research design and results.
is better able to monitor when the media covers a con-
I conclude with a few remarks on the implications of this
flict, the model suggests that media inattention presents
judges with a significant political problem. Without me-
dia coverage, courts may be forced to strategically de-
fer to political interests. While the Vanberg model nicely
integrates separation of powers and legitimacy theories,
Separation of Powers, Legitimacy,
he treats the likelihood that the public will be informed
and Public Relations
about interbranch conflicts as exogenous. But what if
courts are able to influence their media coverage? It would
Comparative researchers have increasingly turned to sep-
seem likely that judges might wish to exercise that power.
aration of powers models in order to explain why some
In the model that follows, I identify conditions under
constitutional courts appear unwilling to hold political
which public relations can address the problem posed
officials accountable for their behavior, no matter how
by media inattention and ask how the ability to influ-
suspect (Helmke 2002; Iaryczower, Spiller, and Tommasi
ence media coverage might influence decision making
2002; Vanberg 2001). Although these models assume that
judicial decisions are driven by policy preferences, they
If Vanberg is correct, media inattention presents con-
also posit that interbranch politics may influence choice
stitutional judges with a political problem, but because
(Epstein and Knight 1998, 22–51). Legislators and exec-
reporters do not perfectly translate jurisprudential lan-
utives ordinarily enjoy control over judicial institutions
guage, so does media attention. Although the press fre-
including jurisdiction, budgets, and tenure. This control
quently communicates merits decisions accurately, they
may grant politicians undue influence over the judiciary,
often misinterpret jurisprudential rationale, an outcome
especially where judicial institutions are relatively easy to
that judges themselves cite as frustrating (Davis 1994, 31).
change (Ferejohn 1999; Ramseyer and Rasmusen 2001).
Indeed, the Mexican Supreme Court, the empirical ob-
In light of these mechanisms of control, separation of
ject of this study, has explicitly stated a concern with the
powers models predict that high courts will be increas-
way reporters cover the jurisprudential theories that sup-
ingly likely to strategically avoid conflict by upholding
port their decisions.3 In the model that follows, I con-
challenged public policies as the importance relevant of-
sider how the incentive to ensure accurate reporting might
ficials assign to those policies increases (Helmke 2002,
interact with the more political motivation described
Separation of powers models are theoretically precise,
To summarize, the selective promotion of case reso-
but they cannot explain why courts sometimes challenge
lutions might be able to address two problems consistent
powerful officials. Judicial legitimacy theorists propose
with a model that integrates separation of powers and le-
that because the coercive sources of judicial power are
gitimacy theories. Promoting cases both may be a means
minimal, courts heavily rely on societal beliefs in their le-
of generating media attention when it is particularly useful
gitimacy to gain compliance. Legitimacy here is conceptu-
and ensuring accurate communication of jurisprudence
alized as diffuse public support or deep commitment to the
when the media decides to cover a decision. The following
institutional integrity of the judiciary (Caldeira and Gib-
model considers the degree to which case promotion can
son 1992, 638). Although legitimacy theory is designed to
resolve these problems.
explain personal acceptance of unpopular decisions, it is
a short step from the individual to the institutional level.
A Separation of Powers Theory
Vanberg integrates key insights from these two ap-
proaches (also see Friedman 2000). In his model, courts
of Case Promotion
and politicians interact under the sometimes-watchful
eye of a public, which is capable of punishing recalci-
Figure 1 depicts an extensive form game of incomplete
trant officials if aware of the conflict. In this sense, diffuse
information played between a constitutional court and
public support provides courts with a sort of political
a national executive, though the second player may be
cover; however, this is only true if people are informed
about the decisions they are purportedly enforcing. A key
This policy is published in “Compromiso, ´organo informativo del
Poder Judicial de la Federaci ´on,” n ´umero 1, julio-agosto 1999, pp.
empirical implication of this model is that constitutional
21–22. [Commitment: A Newsletter of the Judicial Branch of the
courts should be more likely to strike down public policies
Federation, 1(July-August):21–22].

Extensive Form of Baseline and Promotion Models
overage (p)
No Coverage (1-
edia C
A -C
A -K
A -K -C-K

Probability of media coverage U
Court s value of alternative policy
Uphold and promote
Cost of alternative policy for Executive
Strike down
Cost of public backlash
Strike down and promote
Cost of media errors
Cost of promoting a case
A ∈ℜ; A , B,E,K>0; |A |>E,K
Note: Figure 1 specifies the extensive form of both the Baseline and Promotion versions of
the game. The Actions available to the players only in the Baseline version are bolded.
conceptualized as any government official.4 I begin by
tion of the status quo, noncompliance may also involve
discussing a baseline model in which the court cannot
taking retributive actions against the court, the severity of
promote its opinion. I then analyze the same model but
which I parameterize in the court’s payoff function.
allow the court to promote its decision in the national
Before either player moves, there is a random draw
media. The bolded lines in Figure 1 indicate the only ac-
from a set that determines the level of media attention,
tions available in the baseline model. The court is asked
and by implication, public awareness of the case. Either the
to review a policy associated with the executive. It may
media covers the case and accurately communicates the
uphold the status quo or strike it down, substituting an
court’s merits decision or the media ignores the decision
alternative policy. If the court strikes it down, the ex-
and the public goes uninformed about the interaction.
ecutive chooses whether to comply with the alternative.
Denote the probability that the media covers the case p.
I conceptualize noncompliance broadly. Although non-
The court does not observe the outcome of the draw, but
compliance always involves the continued implementa-
it knows the distribution from which it is drawn (i.e.,
the court knows p). Since the executive moves after the
4I thank Georg Vanberg for his suggestions concerning how to sim-
decision it observes whether the media has covered the
ply endogenize case promotion in the context of his model.
case. As a result, the executive is perfectly informed about

the history of the game. Thus, a strategy for the court is a
Baseline Model Equilibrium Predictions
function sC, that assigns an action, Uphold or Strike Down,
to its single information set, depicted by the dashed lines
Executive Preferences:
in Figure 1. A strategy for the executive is a function s
Value of Status Quo Relative
that assigns an action, Comply or Defy, to each of its two
to a Public Backlash
singleton information sets.
I assume that judicial choice is primarily a function of
(AE > B)
(AE B)
policy preferences. I fix the value of the status quo policy
Court prefers
Case 1
Case 2
at 0 for both players. If the executive faithfully implements
the status quo For all p
For all p
an alternative, the court receives AC ∈ . This allows the
(AC ≤ 0)
sC = (Uphold)
sC = (Uphold)
court’s evaluation of the status quo relative to the alter-
sE = (Defy, Defy)
sE = (Comply, Defy)
native to vary across cases. The court pays a cost C > 0 if
Court prefers
Case 3
Case 4
the executive fails to comply. Although I assume that the
the alternative For all p
4A. For p
media accurately covers the merits decision, consistent
C + AC
(AC > 0)
with the judicial belief that reporters often mischaracter-
C = (Uphold)
sC = (Uphold)
ize jurisprudential details, I assume that the court pays a
E = (Defy, Defy)
sE = (Comply, Defy)
4B. For p >
cost E ∈ (0, |A
C + AC
C|) if the media covers the resolution. This
sC = (Strike Down)
cost represents the value the court places on reporters ac-
sE = (Comply, Defy)
curately communicating their decisional rationale. If the
media ignores the case, there is no rationale to misrepre-
Note: Table shows SPE strategy profiles in the baseline model for
given values of the parameters. The court assigns probability p to the
sent and the court saves E.5
history Media Coverage and 1 − p to the history No Coverage.
The executive is assumed to prefer the status quo over
any alternative and will pay a cost AE > 0 if she com-
plies with an unfavorable decision. If the executive fails
allow for since the heart of the diffuse public support
to comply and the media is covering the case, she pays a
concept is that people are willing to accept unpopular
cost among her constituents for defying the court, B >
decisions as long as the decisional source is legitimate.
0. A natural interpretation of these parameters incorpo-
For any given values of the parameters, there is a
rates the concepts of specific public support for the policy
unique pure strategy subgame perfect equilibrium (SPE).
under review (the public’s issue-related preferences) and
Table 1 lists the SPE strategy profiles in the baseline
diffuse public support for the court. If we assume that the
model for given values of the parameters (see appendix
executive is responsive to public preferences, then A
for proofs).
E may
be interpreted as specific public support for the status quo
Table 1 partitions the combinations of judicial and ex-
policy and B as diffuse public support for the court. In-
ecutive preferences over the status quo into four general
terpreted in this way, the model allows executives to defy
cases. The court prefers the status quo to the alternative
highly legitimate courts over valuable policies, but it also
in Cases 1 and 2 (AC ≤ 0) and prefers the alternative in
allows executives to implement decisions that are unpop-
Cases 3 and 4 (AC > 0). The value the executive assigns to
ular among their constituents, as long as diffuse public
the status quo is greater than the cost of the public back-
support is sufficiently large. This is a sensible result to
lash in Cases 1 and 3 (AE > B). In Cases 2 and 4 the value of
the status quo is smaller than the cost of noncompliance
(AE ≤ B). For Cases 1, 2, and 3 the court’s equilibrium
5I take the judicial interest in ensuring accurate coverage as a pro-
strategy is to uphold the status quo. In the first two cases, it
fessional end in itself; however, it may be that judges perceive the
does this sincerely. As should be expected, upholding the
inaccurate communication of their rationale as threatening to dif-
status quo over an inferior alternative is always optimal
fuse or specific public support. If the written opinion is the primary
tool through which courts persuade dissatisfied people or propagate
(Cases 1 and 2). This captures the reasonable implica-
the myth of impartiality (Gibson, Caldeira, and Baird 1998, 345), a
tion that courts do not veto public policies they sincerely
key component of legitimacy, inaccurate reporting may affect their
like—strategic judicial behavior is essentially deferential.6
control over diffuse support. If this is true, then a reasonable alter-
In Case 3, where the court prefers the alternative policy yet
native modeling choice is to have the court pay E > E when it strikes
down the policy, since this is where the court faces the largest set
the executive will defy all decisions, the court strategically
of dissatisfied citizens for whom the written opinion may be per-
suasive. While reasonable, this change only increases the incentive
to publicize decisions striking down the status quo and does not
6See Helmke (2005) for a model of judicial behavior in which courts
affect the decision-making results.
veto policies that they sincerely prefer.

upholds the status quo in order to avoid being defied.
Promotion Model
The model only predicts that courts will be influenced by
media coverage in Case 4, a world in which the expected
The promotion model, also depicted in Figure 1, is identi-
public backlash is sufficient to induce compliance and the
cal to the baseline model in all but one regard. I allow the
Court prefers the alternative policy. As Vanberg predicts,
court to promote a resolution among the national media
for relatively low values of p, p
, the court will
C + AC
at a cost K > 0. K measures the administrative costs as-
strategically uphold the status quo because it is insuffi-
sociated with promoting a case and the opportunity cost
ciently likely that the public will be able to monitor the
of promoting one decision over another. Maintaining the
interaction. For sufficiently large p, the court will take
assumption that constitutional politics is ultimately about
advantage of likely media coverage and strike down the
policy, I assume |AC| > K.7 I further assume that if the
status quo.
court promotes the resolution, the media accurately trans-
The results of the baseline model capture Vanberg’s
lates information about the case to the public. Clearly, this
analysis, though he does not explicitly discuss the con-
is a strong assumption, but it reduces the set of parameters
ditional effect I now introduce. If courts are unable to
and relaxing it will not materially affect the implications
influence their media coverage the following proposition
of the model. What matters is that case promotion at least
is consistent with the equilibrium results.
increases the chances of quality media coverage.8 The re-
maining structure of the model is identical to the baseline.
Proposition 1.
The relationship between media coverage
A strategy for the court continues to assign a single action
and judicial activism is conditioned by the importance of
to its single information set; however, the set of actions
the status quo relative to diffuse public support (i.e., the size
now includes the following choices: Uphold; Uphold and
of A
Promote; Strike Down; and Strike Down and Promote. An
E relative to B). When policy importance is relatively
small, courts should be more likely to strike policies down
executive strategy assigns an action, Comply or Defy, to
as p increases; however, when importance is relatively large,
each of its four singleton information sets.
there should be no relationship between p and the court’s
Like the baseline, the promotion model produces im-
plications for judicial decision making; however, it also
suggests implications for the selective promotion of case
results. More generally, it suggests limits to case promo-
In other words, courts should be more aggressive when
tion as a strategy for resolving the problems of media
the public is likely to be able to monitor the interaction,
inattention and reporter inaccuracy. As before, there are
but only if the status quo is not overly important. When it
four general cases to consider. Table 2 summarizes the
is, no amount of public monitoring will enhance judicial
results, which suggest the following proposition if courts
can control their media coverage.
The baseline model highlights the two problems
judges may be able to resolve through public relations.
First, it is the uncertainty surrounding the public’s ability
Proposition 2.
Courts should be less likely to strike down
to monitor the interaction that produces strategic judicial
policies as importance increases relative to diffuse public
behavior in Case 4. If the court were sure that the media
support; however, there should be no relationship between
would cover the resolution, it would have no incentive to
media coverage and decision making.
uphold the status quo. Thus, if judges are able to induce
media attention for their resolutions, they may be able
It is still the case that courts sincerely uphold the status
to eliminate the necessity to strategically defer in situa-
quo in Cases 1 and 2 and strategically uphold the status
tions like Case 4. Second, in all cases, the Court expects to
quo in Case 3; however, it is no longer the case that the
lose pE in equilibrium, the cost of imprecise media cov-
court will uphold the status quo for sufficiently low p in
erage discounted by the probability that the media will
report on the case. This problem may be resolvable via
This assumption reduces the number of cases I consider. Allowing
K > |AC| does not affect the relative incentives to promote decisions
public relations as well. The choice publicly to promote a
when courts either uphold or strike down; however, it would allow
decision may give judges the ability to clarify their resolu-
for some equilibria in which courts strategically uphold the status
tions, improving on the reporters’ description of the case.
quo in order to avoid having to pay the cost of promotion, a result
that only obtains for small A
I now turn to a version of the model in which the court
C. Since policy is generally perceived
to be paramount, the model assumes away this possibility.
is able to produce accurate media coverage through case
8Davis (1994) provides empirical support for this claim in the
promotion, which allows me to evaluate the plausibility
United States, and my own investigation of the Mexican data, which
of solving these problems through case promotion.
I briefly discuss below, supports it as well.

TABLE 2 Promotion Model Equilibrium Predictions
Executive Preferences:
Value of Status Quo Relative to a Public Backlash
(AE > B)
(AE B)
Court prefers the status
Case 1
Case 2
quo (AC ≤ 0)
1A. For p K
1A. For p K
sC = (Uphold)
sC = (Uphold)
sE = (Defy , . . . , Defy)
sE = (Comply, Defy, Comply, Defy)
1B. For p > K
1B. For p > K
sC = (Uphold and Promote)
sC = (Uphold and Promote)
sE = (Defy , . . . , Defy)
sE = (Comply, Defy, Comply, Defy)
Court prefers the
Case 3
Case 4
alternative (AC > 0)
1A. For p K
4A. For p C + AC K
C + AC E
sC = (Uphold)
sC = (Strike Down and Promote)
sE = (Defy , . . . , Defy)
sE = (Comply, Defy, Comply, Defy)
1B. For p > K
4B. For p > C + AC K
C + AC E
sC = (Uphold and Promote)
sC = (Strike Down)
sE = (Defy , . . . , Defy)
sE = (Comply, Defy, Comply, Defy)
Note: Table shows SPE strategy profiles in the promotion model for given values of the parameters. The court
assigns probability p to the history Media Coverage and 1 − p to the history No Coverage.
Case 4. Instead, when p is sufficiently low, the court will
resolution striking down the status quo if and only if p <
strike down the status quo and create its own media cov-
pS. First, note that pU does not depend on whether the
erage via case promotion. For sufficiently high values of
court upholds the status quo sincerely or strategically; it
p, the court merely strikes down the status quo taking ad-
is identical in Cases 1, 2, and 3. This means that whatever
vantage of the media’s expected coverage. In this sense, the
the reason for upholding the status quo, whether sincere
model suggests that public relations can solve the problem
affinity for the policy or strategic deference, the choice to
of strategic judicial deference associated with uncertainty
publicize the result is governed by the same logic. Second,
about public information (Case 4 behavior), but only if
recognize that if there is a p that satisfies the pU condition
the value of the status quo is sufficiently low. If the status
(and the Court promotes a decision upholding the status
quo is highly valued, then as before, no amount of public
quo), then the same p satisfies the pS condition (and the
monitoring will change the dynamics of the interaction.
Court promotes a decision striking it down); however,
In short, public relations can help constitutional courts
the converse is not true. There are two possibilities. Either
but there are limits—limits imposed by the political im-
the cost of promotion is less than the cost of media inaccu-
portance of the status quo.
racy or it is not (i.e., K < E or K ≥ E). If K < E, pU ∈ (0, 1)
Finally, the promotion model suggests that courts will
and will be satisfied for sufficiently high p, yet pS is satis-
promote both decisions upholding the status quo and de-
fied for all parameter values, because C + A K > 1. Alter-
C + A E
cisions striking it down; however, the incentives to pro-
natively, if K ≥ E, the court will never promote a decision
mote these different types of resolutions are distinct.
upholding the status quo, because pU ≥ 1, and no p is this
large. For the same values of K and E, pS ∈ (0, 1), and it
Proposition 3.
Courts should be more likely to promote
will be satisfied for sufficiently low p.
decisions that strike down the status quo than those that
The intuition behind this result is as follows. Whether
uphold it. This relationship should be especially strong when
the court upholds the status quo or strikes it down, it faces
the media is unlikely to cover a case.
the problem associated with imprecise media coverage
(i.e., the parameter E); however, when the court strikes
Using the results in Table 2, let pU = K and pS =
down a policy it faces the additional political problem of
C + A K . The court will promote a resolution upholding
C + A E
noncompliance. If the cost of promotion is sufficiently
the status quo if and only if p pU ; it will promote a
small, there will be cases in which the court promotes

decisions upholding the status quo, but any time it is op-
public policies than those upholding policies. This
timal to promote a decision upholding the status quo, it
relationship should be especially strong when the
is optimal to promote a decision striking it down. This
media is unlikely to cover the resolution.
explains the direct relationship between the decision type
and case promotion; the following is the intuition behind
I now turn to the data I use to test these implications.
the conditional relationship. Continue to assume that K is
small enough so that pU can be satisfied. At high values of
p, the court will promote all kinds of decisions, resolving
Research Design
reporter inaccuracy, but at low values of p the incentive
to promote in order to address inaccuracy disappears be-
cause it is unlikely that there will be any inaccuracy to
Mexico at the turn of the twenty-first century presents
correct. At such values of p, only decisions striking down
an excellent environment for testing the argument. As
the status quo will be promoted in order to deal with
Helmke (2002, 291) notes, strategic models are more likely
potential noncompliance.
to explain judicial behavior in settings where institutions
In addition to suggesting a testable empirical hypoth-
are relatively unstable. Although a monumental judicial
esis, this result suggests that there are different limits to
reform in 1994 significantly expanded the Court’s juris-
public relations as a strategy for resolving media inatten-
diction, tenure was changed from life to a nonrenewable
tion and reporter inaccuracy. As before, the policy im-
15-year term, its membership was reduced from 26 to 11,
portance of the status quo limits the ability of the court
all ministers were forced to resign, and a new Court was
to resolve the potential noncompliance problem result-
appointed. This was not the first such change during the
ing from media inattention. In contrast, the limit on the
twentieth century. In fact, the size of the Supreme Court
ability of courts to resolve reporter inaccuracy is only im-
changed four times between the adoption of the 1917
posed by the cost of promoting cases. That is, for arbi-
Constitution and the 1994 reform. Rules governing judi-
trarily small K, courts can always address this problem via
cial tenure changed five times over that period. In all, the
case promotion. To review then, the theoretical models
federal constitutional articles governing the judiciary have
suggest three testable hypotheses. Both models speak to
been amended over 69 times since 1928 (Carranco Z ´u˜
judicial decision making, but only the promotion model
2000, 97). Also, Mexican historians and political scien-
generates a prediction for case promotion.
tists point to at least three clear instances of institutional
change during the twentieth century that were designed to
reign-in an activist Supreme Court (Domingo 2000, 713;
Decisional Hypotheses
Schwarz 1973, 313). Thus, it would appear at least facially
valid to assume that the ministers of the Supreme Court
H1 (Baseline Model): If a court is unable to con-
who took the bench in early 1995 might have perceived a
trol its media coverage, it should be more likely
risk of further institutional tinkering.
to strike down public policies when the media
Second, the Supreme Court’s membership did not
is likely to cover the outcome; however, this re-
change between February 1995 and December 2003, when
lationship is conditioned by the importance of
two members of the original 11 stepped down. This ab-
the status quo policy for relevant political actors.
solute stability in membership allows me to control for
When the value of status quo is sufficiently high,
aggregate judicial ideology. Unless preferences are chang-
media coverage should have no effect.
ing over the course of the period studied, the aggregate
ideology of the court, whatever that may be, is controlled
H2 (Promotion Model): If a constitutional court
by design.
is able to control its media coverage, it should
In addition, three key assumptions of the theoreti-
be less likely to strike down important policies,
cal model are met in the Mexican case. First, the Court
but media coverage should not affect its decision
has formally stated that it wishes to ensure accurate me-
dia coverage of its decisions.9 Second, there is reason to
believe that Supreme Court coverage was largely neutral
Case Promotion Hypothesis
during the period I study. The Supreme Court’s own in-
ternal analysis of its daily coverage in 25 newspapers and
H3 (Promotion Model): If a constitutional court
magazines during the first half of 1999 classified 87% of
is able to control its media coverage, it should be
more likely to promote decisions striking down
9See note 25.

the 3,222 stories as neutral and 93% as either neutral or
positive. Of the 230 articles judged to include a negative
opinion of the Court, 173 (75%) involved a reporter quot-
I use an original data set on the universe of Supreme Court
ing an individual that disagreed with the decision, rather
constitutional decisions in review of state or federal laws
than a negative editorial.10 Further, many of the remaining
made in plenary session between January 1, 1997 and
57 articles had nothing to do with constitutional cases, but
December 31, 2002.14 The data set includes three kinds
rather concerned judicial reform proposals. Third, there
of constitutional actions: amparo appeals, constitutional
is evidence that the Court has been able to generate cov-
controversies, and actions of unconstitutionality. Amparo
erage through case promotion. My own analysis suggests
is a form of individual constitutional complaint wherein
that the newspaper La Jornada was roughly 1,300% more
persons, organizations, and corporations may seek redress
likely to cover a case in which the Court issued a press
in federal court against alleged governmental violations of
release than a case for which it did not. This result holds
individual rights. The Supreme Court hears amparo suits
up to a variety of multivariate specifications in which I
on appeal from lower federal district and circuit courts. In
control for elements of newsworthiness.11
contrast, the Supreme Court hears constitutional contro-
The final advantage of the Mexican case concerns ju-
versies and actions of unconstitutionality in the first in-
dicial legitimacy. Diffuse support is controlled by research
stance. In constitutional controversies, the Court resolves
design, unless it is rapidly changing over the short time
jurisdictional conflicts between levels of government and
period of the study. This result is highly unlikely given
across branches of government within a level (i.e., federal-
the slow way diffuse support has been shown to develop
ism and separation of powers issues). In actions of uncon-
(Gibson, Caldeira, and Baird 1998, 353). Still the study
stitutionality, the Court exercises abstract review over the
does require that diffuse support be greater than zero. If
constitutionality of state laws, federal laws, and interna-
it is not, then the court will have no incentive to inform
tional treaties.15 With one limited exception, the Supreme
the public, since a public backlash against a recalcitrant
Court’s jurisdiction in all three cases is mandatory.16 Thus,
official will be costless. Unfortunately, there is no pre-
the data set includes all constitutional claims that met the
cise measure of diffuse public support for the Mexican
Court’s statutorily defined jurisdiction.17 This is the first
Supreme Court during the period studied here.12 That
said, the available public opinion data suggests that dif-
vestigator for the 2000 wave of the World Values Survey. In-
fuse support may not have been completely absent. In a
formation on the Reforma sampling design can be obtained at
series of national opinion surveys conducted by the news- or by contacting the author. I
have been unable to locate national public opinion results for the
paper Reforma between December 2000 and March 2002,
period between 1997 and 2000.
respondents were asked to evaluate the job of the Supreme
14All replication material is available at
Court. During this period, the percentage of respondents
∼jstaton. Copies of full Supreme Court opinions are available on
expressing favorable opinions varied between 40% and
the Web at
50%, statistics that climb as high as 87% when excluding
15On the differences between these actions see Carranco Z ´u˜niga
nonresponses and neutral ratings.13
(2000) or Coss´ıo (2002).
16Pursuant to a 1988 constitutional amendment, the Supreme Court
enjoys a limited power of discretionary review through what is
10See “Qui´enes hablaron bien y quienes hablaron mal de la Supreme
called atracci´on (Mexican Constitution, Article 107, Section 8). If
Corte de la Naci ´on en el primer semester de 1999? [“Who spoke
the Court deems an issue raised in an appeal outside of its appellate
well and badly of the Supreme Court of the Nation during the first
jurisdiction fundamentally important to the law, it may exercise
half of 1999?”]. This report is on file with the author. It may be used
atracci´on. In addition, the Federal Attorney General may request
with the permission of the Supreme Court.
that the Court take up a case of national interest, and circuit courts
11Results available upon request.
may certify appeals for Supreme Court review.
12Unfortunately, neither the World Values Survey nor the Lati-
Between 1995 and 2002 the Supreme Court resolved 3,084 cases.
nobar ´ometro include items appropriate for measuring Supreme
The data set restricts the number of analyzable cases to 1,536. Many
Court support. The World Values Survey asks about pub-
cases with distinct case numbers address identical legal issues and
lic trust in the “legal system” (Question V137, 1995 Wave,
resulted in identical resolutions resolved by an identical vote on while the Latinobar ´ometro
the same day. For example, a major constitutional reform on in-
asks about trust in the “judiciary” (Question V.2.32, http://www.
digenous rights reform passed by Congress in the summer of 2001 Given the high levels of public confusion in
generated 292 distinct constitutional controversies filed by munici-
Mexico about what institutions fall with the judiciary or the legal
palities from across the country. Each of the 292 cases challenged the
system, these are highly imprecise measures of institutional com-
procedures Congress adopted when considering the amendment.
mitment to the Supreme Court itself.
These cases make up nearly 10% of the Court’s total constitutional
caseload, yet they were all formally resolved on September 6, 2002,
13During this period, the director of public opinion for Re-
by the same vote and for the same reason. These cases are clearly
forma was political scientist Alejandro Moreno, the country in-
not independent of each other.

quantitative study of which I am aware that includes the
care more about federal policies than state policies, and
full range of constitutional actions, an important feature
they care more about the policies challenged through con-
for the measurement of policy importance.18
stitutional controversies and actions of unconstitutional-
ity than they do about those challenged through amparo.
Given these assumptions, I generate a scale of increasing
federal importance, coded from 0 to 3, ranging from a
Measurement and Estimation
state law challenged via an amparo suit to a federal law
challenged under the action of unconstitutionality or con-
Since I will test both decisional and case promotion hy-
stitutional controversy.20
potheses, I require two dependent variables. The first,
The first assumption would appear relatively uncon-
Strike, indicates whether the Court invalidated a policy
troversial. On average, the Federal Congress likely cares
challenged through one of the three constitutional ac-
more about the federal penal code than it does about the
tions. Strike is coded 1 if the Court supported at least one
penal code of Nuevo Le ´on. On the other hand, the first
complainant argument made on the merits, and 0 oth-
assumption deserves further justification. Resolutions to
erwise. The second dependent variable, Press, is coded 1
amparo suits settle only the immediate controversy be-
if the Court issued a press release announcing the result,
ing adjudicated.21 In contrast, decisions in both consti-
and 0 otherwise. Press releases on case outcomes typi-
tutional controversies and actions of unconstitutionality
cally summarize the major components of the resolution,
have the potential of setting general effects.22 With respect
including the parties to the case, the constitutional ques-
to the measure, I argue that federal officials ought to care
tion, and a brief statement of the rationale. Formally, press
more about resolutions that have the potential of setting
releases are issued by the Court’s public relations office
precedent than about those that do not.
(DCS). Since the DCS director remains in close contact
Constitutional controversies and actions of uncon-
with the Supreme Court president, there is little reason to
stitutionality also deal with what we might understand
believe that the kinds of cases the DCS publicizes fail to
as generically more significant political issues. Typical
reflect the Court’s own preferences.19
constitutional controversies involve state-municipal con-
If the promotion model is correct, these variables are
flicts over the autonomy of local governments, state-state
jointly determined. We can estimate their joint distribu-
boundary disputes, and federal interbranch conflicts over
tion and test the hypotheses specified above by the fol-
competing claims on power (Fix-Fierro 1998). Actions
lowing recursive simultaneous equations model.
of unconstitutionality frequently involve political party
Strike =
challenges to the constitutionality of electoral laws, rules
1(Importance) +
quite essential to political interests. These are cases that
+ 3(Importance Coverage) + 1X1 + u1
affect large numbers of people and large sums of money.
Press = 4(Strike) + 5(Coverage)
The results are robust to an alternative coding scheme where I
6(Strike Coverage) +
2 X2 + u2,
combine the state policies into one category and where I use a
where Importance is a measure of political importance
dummy variable distinguishing between state and federal policies.
assigned to the challenged policy by Mexican federal offi-
21The Court may establish formal jurisprudential theses in amparo.
cials, Coverage measures the court’s expectations about
A jurisprudential thesis is binding on all lower federal court judges;
however, it is not clear that theses significantly affect bureaucratic
subsequent media coverage, and the Xi are vectors of
decisions to enforce laws. Legal principles may become part of the
control variables for each equation. As recommended by
jurisprudence when the Court establishes that it has invoked a simi-
Greene (1998), this model can be estimated via bivariate
lar principle in five consecutive cases (Ley de Amparo, Article 192).
These theses may be abrogated by a coalition of eight Ministers
voting contrary to provision under analysis (Article 194).
I measure federal policy importance by appealing to
the kind of constitutional claim under review and the level
When the Federation moves a constitutional controversy against
the laws or acts of a state or municipality, decisions supporting the
of the public official against which the claim was raised.
federation’s claim, when made by a super-majority of eight minis-
Two assumptions underlie the measure. Federal officials
ters, set general effects. Similarly, if a state moves a constitutional
controversy against a municipality, the Court’s decision will govern
18Coss´ıo (2002) uses a qualitative research design and neither
that state if it rules with the state by a super-majority vote of eight.
Magaloni and S´anchez (2001) nor Schwarz (1973) study all three
In all other cases, the Court’s decisions only affect the immediate
parties to the case. Similarly, for the action of unconstitutionality,
the effects of the decision are general if the Court resolves the case
19Although many press releases are drafted by DCS staff, the lan-
by a super-majority of eight votes. Otherwise, the decision only
guage is approved by individual ministers, the DCS director or his
affects the immediate parties to the case (Mexican Constitution,
immediate assistant.
Article 105).

In contrast, state entities have extremely limited standing
and 2002, I include dummies for the period of truly di-
in amparo; political parties have none. Also, individuals
vided national government under Vicente Fox (Divided)
are prohibited from challenging electoral codes through
and the period of unified PRI government, when the PRI
amparo. Finally, the Court itself has argued that amparo
only enjoyed a plurality of support in the legislature un-
suits are generically less important than constitutional
der Ernesto Zedillo (Unified-Pl). Thus, the period of truly
controversies and actions of unconstitutionality.23
unified government under Zedillo is the base category.25
I measure the Court’s expectations regarding subse-
In order to address the possibility that Supreme Court
quent media coverage by appealing to prior national print
decisions were affected by partisan factors (Magaloni and
media coverage of the case or the conflict that generated
S´anchez 2001), I also include a dummy variable coded 1
it. Here I assume that the Court estimates a higher prob-
if authority responsible for the challenged policy is either
ability of ex post coverage when the case has already been
a pri´ısta executive or a majority pri´ısta legislature and 0
covered by the print media. Because each member of the
Supreme Court receives a daily news briefing with photo-
I also control for the identity of the plaintiff moving
copied articles of every print story that even tangentially
the constitutional action, capturing the notion that the
relates to the federal judiciary, print coverage is a natural
quality of legal representation should vary according to
way for them to generate beliefs about subsequent levels
plaintiff type and using the assumption that political sta-
of public information (also see Epstein and Segal 2000).
tus is a reasonable proxy for litigant resources (McGuire
Specifically, I use the newspaper La Jornada as a proxy
1995; Sheehan, Mishler, and Songer 1992). The issue of le-
for general media attention.24 Coverage is coded 1 if La
gal representation is an especially relevant issue in Mexico,
Jornada ran an article on the controversy prior to the de-
where access to good counsel is limited and the dispari-
cision and 0 otherwise. Unsurprisingly, cases that received
ties between the legal representation of various parties is
prior coverage represent only 3% of the total. On the other
often great (Rubio, Magaloni, and Jaime 1994, 119–34).
hand, prior coverage is an excellent predictor of whether
Controlling for plaintiff identity involves estimating a se-
the resolution will receive subsequent media attention. La
ries of dummy variables that account for a large number
Jornada covered 72% of the resolutions to cases on which
of plaintiff categories.27 The base category is the modal
it had previously run a story; it only covered 3% of the
plaintiff, a private individual.
resolutions to cases that it had previously ignored.
In the Press equation, I control for a Supreme Court-
based theory of case promotion. In 1999, the Supreme
Court issued a formal policy statement defining the DCS
mission. In addition to its charge to ensure that media
coverage accurately reflects jurisprudence, the DCS was
In the Strike equation, I control for three potential influ-
asked to publicize the results of cases that resolve intrinsi-
ences on judicial decision making. Iaryczower, Spiller, and
cally “important” issues of Mexican law.28 Unfortunately,
Tommasi (2002, 704) suggest that courts should be more
the policy does not define what the ministers mean by an
deferential under unified government, because it is easier
intrinsically important issue, and there does not appear
for presidents and legislators to coordinate on a response
to be a ready-made measure of this concept. Still, it would
to undue judicial activism. To capture the three phases
appear that Importance, as defined above, might capture
of unified-divided government in Mexico between 1997
exactly what makes a particular constitutional question
23In April 2002, the ministers agreed to begin returning amparo ap-
more intrinsically important than another. The questions
peals to the benches, an effort designed to leave the full Court free
to consider what the ministers themselves claim raise more impor-
tant questions of constitutional law—constitutional controversies
The first period ended on September 1, 1997 and the second on
and actions of unconstitutionality. See Acuerdo [Accord] 4/2002,
December 1, 2000.
Suprema Corte de Justicia de la Naci ´on.
26Partisanship data for the state legislatures and governors
24Preliminary analysis on the internet and inside Mexico at re-
may be found at the Website for the Mexican Senate
search university libraries offered little confidence in the reliability
( and in Lujambio (2000).
of electronic searches for articles concerning the Mexican judi-
27These categories include: individual or private business, munic-
ciary. InfoLatina, Lexis-Nexis’s Spanish search engine and all utili-
ipality, union, civil or religious association, small political party,
ties maintained by the main national newspapers produced highly
corporation, large political party (PRI, PAN, PRD), university, state
erratic results, especially for the mid 1990s. Accordingly, the best
legislature, state executive, federal commission, federal congress,
strategy appeared to be a daily, full text search. La Jornada main-
state judiciary, federal judiciary, and federal executive. The results
tains a publicly accessible Website that contains issues dating from
are robust to various ordinal measures of complainant political
March 29, 1996 to the present. In addition, La Jornada is one of a
small number of widely circulated newspapers known for its inde-
pendence (Lawson 2002, 69).
28See supra note 3.