Explaining Policy Ties in Presidential Congresses: A Network Analysis of Bill Initiation Data
| DOI | 10.1111/j.1467-9248.2012.00964.x |
| Published date | 01 June 2013 |
| Date | 01 June 2013 |
| Author | Ernesto Calvo,Eduardo Alemán |
| Subject Matter | Original Article |
Explaining Policy Ties in Presidential Congresses:
A Network Analysis of Bill Initiation Datapost_964 356..377
Eduardo Alemán Ernesto Calvo
University of Houston University of Maryland
Policy networks formed by co-authoring and co-sponsoring bills reflect one of the most important types of
connection legislators develop while in office.Weexpect that in presidential countries, the probability of a tie between
two legislators should be influenced by partisan membership, territorial linkages and the policy areas in which they
develop expertise. Given the complex nature of relational data and the particular characteristics of bill initiation
networks, we propose a new approach – bootstrapping an exponential graph model using augmented data reflective
of the frequency of ties – to address the challenges of thinning dense networks.
Keywords: legislative politics; political parties; social networks; committees
Social science literature has typically argued that dense social networks with cross-cutting
affiliations provide a favorable environment for democracy and consensual politics (Lipset
and Rokkan, 1967; Mutz, 2002; Putnam, 1993). Policy networks derived from co-authoring
and co-sponsoring bills reflect one of the most important types of connection legislators
develop while in office (Crisp, Escobar-Lemmon et al., 2004). Prior works focused on bill
initiation data have examined how an actor’s relative position in the network affects
legislative success (Fowler, 2006;Tam Cho and Fowler,2010), how variations in ties reflect
changes in political polarization (Alemán, 2009; Zhang et al., 2008), how networks can be
utilized to map actors’ policy preferences (Alemán et al., 2009; Crisp, Kanthak et al., 2004)
or to identify programmatic leaders (Panning, 1983).While the growing literature on social
networks has illuminated some of the partisan and career traits that determine legislators’
behavior, there is still relatively little comparative research on the determinants of policy
networks. This article attempts to fill this gap by examining the main determinants of
policy collaboration networks in two presidential democracies. Given the complex nature
of relational data and the particular characteristics of bill initiation networks, the statistical
analysis of policy networks raises some interesting methodological questions that our article
also attempts to answer.
Policy networks are formed by purposive legislators, who connect with others with
whom they share some commonality of policy preferences as well as an interest in similar
policy areas or jurisdictions. These connections, we hypothesize, are well explained by the
institutions structuring legislative behavior in representative democracies. A vast literature
has hypothesized that electoral incentives explain both policy attention and policy effort.
This motivation,character ized by David Mayhew (1974) as an‘electoral connection’, is the
cornerstone of legislative accountability in both candidate-centric and party-centr ic
systems. Competing principals in the district and the nation impact legislators’ collaborative
efforts (Carey, 2007). The relative influence of each is supposed to stem from electoral
incentives embedded in the institutional rules that regulate candidate nomination and party
bs_bs_banner
doi: 10.1111/j.1467-9248.2012.00964.x
POLITICAL STUDIES: 2013 VOL 61, 356–377
© 2012The Authors. Political Studies © 2012 Political StudiesAssociation
competition, and the decentralization of political power. Inside congress, the organization
of legislative business, particularly the committee system, further structures legislators’
policy connections.
The empirical analysis presented in this article measures the determinants of policy
collaboration among legislators in Argentina and Chile. The analysis of policy networks
allows us to evaluate how partisan, territorial and jurisdictional deter minants of collabo-
ration operate in different institutional contexts – Argentina has a federal constitution,with
a volatile party environment, and a weakly specialized committee system, while Chile has
a unitary constitution with strong parties, stable coalitions and a serious committee system.
Through the analysis of policy networks in these two countries we are able to illustrate
within-party cohesion and cross-partisan similarities without single-mindedly focusing on
floor voting behavior. The results of our analysis provide evidence of partisan, territor ial
and jurisdictional effects on policy design that have not been measured in prior comparative
research.
The rest of this article is organized in the following manner. The first section introduces
the notion of a policy network and briefly discusses the motivations behind co-sponsorship.
The second presents our main argument about the formation of policy networks. The third
discusses some implications for legislative behavior in the congresses of Argentina and
Chile, and advances six testable hypotheses. The fourth addresses some methodological
challenges facing network analyses of bill initiation data. We propose a bootstrapped
exponential random graph model (B-ERGM) using augmented data reflective of the
frequency of ties to deal with the problem of thinning networks.Such an approach is used
to test for party,terr itorial and committee effects. The fifth section discusses the results from
the empirical analysis, followed by our conclusion.
Legislative Behavior: What Explains the Likelihood of a Policy Tie?
A social network is composed of actors and relations between these actors.We define policy
networks as the social networks formed by legislators during the process of initiating bills,
and the co-authoring or co-sponsoring of a bill by two legislators as a policy tie. The
relations that make up this network are joint public stances on policy which are explained
by partisan and electoral linkages between representatives and their local and/or national-
level constituencies. Bill initiation data, like voting or plenary floor speeches, carry infor-
mation about the policy stances of legislators.1
What motivates legislators to co-sponsor law initiatives?The legislative literature tends to
agree that co-authoring and co-sponsorship activities are carr ied out by legislators who are
attentive, and seek to claim credit, for the policy consequences of their actions. Yet
disagreement persists over the main audience for this behavior.While some scholar s argue
that legislators co-sponsor bills to signal their policy preferences to voters, others contend
that legislators’ co-sponsorship decisions are primar ily a signal to fellow party members.
The former perspective sees co-sponsoring as a position-taking device ( Balla and Nema-
check, 2000; Campbell, 1982; Koger, 2003; Mayhew, 1974).2Co-sponsorship conveys
information to voters regardless of floor consideration or legislative approval of initiatives.3
In co-sponsoring, as ‘in a large class of legislative undertakings, the electoral payoff is for
positions not effects’ (Mayhew, 1974, 132).
POLICY TIES IN PRESIDENTIAL CONGRESSES 357
© 2012The Authors. Political Studies © 2012 Political StudiesAssociation
POLITICAL STUDIES: 2013, 61(2)
To continue reading
Request your trial