systematically how much informational value legislatures produce and provide to
their members (Cox 2006 ).
Gilligan and Krehbiel ( 1990 ) develop the best-known formal model of informa-
tion specialization among legislators. By this account, individual legislators are
motivated to collect information on policies that improve outcomes for all in
exchange for policy concessions on the margin that can be translated into personal
electoral support. Committees serve as the seed beds both of policy expertise and,
via their control over the legislative agenda, of opportunities for their members to
secure advantageous policies on the margin. The informational model provides a
compelling account of the committee system in the US Congress (Krehbiel 1992 ).
The question for comparative legislative studies is to what extent committees play
this role elsewhere.
- 2 Tenure
The quantity and quality of information are diYcult to measure, but time is a
necessary condition for the development of policy expertise in legislatures. Studies
of the eVects of legislative term limits, for example, suggest that reelection is
necessary for legislators to develop expertise, and that short tenure weakens
legislatures relative to executives in shaping policy outcomes (Carey 1996 ; Carey,
Niemi, and Powell 2000 ; Kousser 2005 ). Comparative politics scholars have begun
to take an interest in reelection rates across legislatures. Reelection rates tend to be
substantially lower than in the United States, although longer legislative terms
(than those of the US House, at any rate) in most other assemblies mean that
diVerences in overall rates of tenure are somewhat less dramatic (Morgenstern and
Nacif 2002 ).
Samuels ( 2000 ) and de Luca, Jones, and Tula ( 2002 ) document low reelection
rates in Brazil and Argentina, respectively. Both are federal systems in which parties
and political careers are primarily organized at the sub-national level. Samuels
( 2000 ) shows evidence that state-level oYces attract many of the strongest politi-
cians away from the national congress. De Luca, Jones, and Tula ( 2002 ),
meanwhile, argue that state-level party bosses who control candidate nominations
in most Argentine states systematically rotate up-and-coming legislators oVparty
lists, and so out of Congress, before they can harness the institutional resources
there to challenge the primacy of state-level machines. In both these cases, explan-
ations for high legislative turnover hang partly on characteristics of federalism—
the lure of state-level oYce in Brazil, and turf guarding by party bosses in
Argentina. The US example demonstrates that long tenure is not impossible
under federalism, of course, but the comparative evidence suggests it may be
442 john m. carey