Content uploaded by Frédéric Varone
Author content
All content in this area was uploaded by Frédéric Varone on Dec 18, 2018
Content may be subject to copyright.
1
Do MPs' Ties to Interest Groups Matter for Legislative Co-sponsorship?
Manuel Fischer (corresponding author) Eawag (Swiss Institute for Water Science and
Technology) and University of Bern, Überlandstrasse 133, 8600 Dübendorf, Switzerland. +41
58 765 56 76, manuel.fischer@eawag.ch
Frédéric Varone, University of Geneva, Boulevard du Pont-d’Arve 40, 1205 Genève,
Switzerland. +41 22 379 83 82. frederic.varone@unige.ch.
Roy Gava, University of Geneva, Boulevard du Pont-d’Arve 40, 1205 Genève, Switzerland.
+41 22 379 83 82. roy.gava@unige.ch.
Pascal Sciarini, University of Geneva, Boulevard du Pont-d’Arve 40, 1205 Genève,
Switzerland. +41 22 379 83 78. pascal.sciarini@unige.ch.
Abstract: This article investigates whether linkages between members of parliament (MPs)
and interest groups matter for MPs' activities of co-sponsoring legislative proposals. Based on
statistical models for network data, the study builds on classical explanations of co-
sponsorships highlighting the role of similar ties between MPs, such as party membership,
legislative committee assignments, electoral district or gender. It shows that, on top of these
traditional forms of homophily, MPs' ties to interest groups also make a difference. MPs with
ties to a similar type of interest groups are more likely to co-sponsor their respective
proposals. The same holds for MPs with ties to groups active in the same policy domain.
These findings have implications for the study of groups' lobbying, legislative behaviour and
representative democracy.
2
Keywords: Co-sponsorship; parliament; parliamentary intervention; interest groups;
networks; homophily
3
1. Introduction
This article studies whether members of Parliament (MPs)' co-sponsorship behaviour varies
according to their ties to interest groups. It assesses whether MPs with links to the same types
of interest groups work together in formulating and introducing legislative proposals on specific
policy issues. Answers to this question are relevant and innovative for the literature on
legislative behaviour and elite networks, as well as for the literature on the policy influence of
interest groups (Varone et al. 2017).
With respect to legislative behaviour, MPs need to recruit co-sponsors for their policy
proposals, which allows them to increase the likelihood of attracting partisan, media and citizen
attention and, eventually, support for their policy agenda. Accordingly, several studies have
focused on MPs co-sponsorship behaviour and have demonstrated that Representatives of the
US Congress who share the same political ideology, who are members of the same legislative
committee, who are elected in neighbouring districts and who have similarities in (a minority
status of) gender or race are more likely to co-sponsors each other's legislative proposals (Koger
2003; Bratton and Rouse 2011: 446). Partisan cohesion, policy expertise and electoral
responsiveness to local voters are also key determinants of bills co-sponsorship in Argentina
and Chile (Alemán and Calvo 2013: 371-2). At the European level, MPs representing the same
Member-State are likely to engage in co-sponsorship (Baller 2017). However, previous
scholarship did not pay attention to another potentially important source of homophily, namely
MPs' ties to interest groups. We argue, and we show, that MPs with ties to a similar type of
interest groups are more likely to co-sponsor their respective proposals. The same holds for
MPs with ties to groups active in the same policy domain.
Interest groups try to influence policy-making in order to realise the policy preferences
of their respective constituencies. MP-group linkages can be conceptualised as exchange
relationships. Groups provide technical expertise about policy issues, as well as strategic
4
information about the policy position of their constituency. As a counterpart, MPs grant groups
a privileged access to the parliamentary venue, or even commit themselves to actively support
policy proposals promoted by groups (Berkhout 2013). Empirical evidence from various
countries corroborates the (self-reported) importance of ties between interest groups and elected
representatives (Rasmussen and Landeboom 2013; Wonka 2017; Eichenberger and Mach
2017). However, previous studies did not scrutinise whether the linkages between MPs and
interest groups matter for legislative co-sponsorship. This study fills a gap by focusing on a
specific legislative behaviour that can be observed directly and, thus, avoiding the
methodological bias of MPs misreporting (Bundi et al. 2016).
Besides its relevance to studies on legislative behaviour and interest group influence,
the potential influence of interest groups on MPs' co-sponsorship also matters from a normative
perspective. This relates to the broader question of which interests MPs do represent in
parliament. If MPs promote the policy agenda of sectional groups instead of representing the
preferences of their broad electorate, then the discrepancy between electoral constituency and
group preferences could be a concern for representative democracy (Giger and Klüver 2016).
The question addressed in this study is thus also relevant for assessing the quality of political
representation and the way MPs fulfil their delegates' function.
The article is structured as follows. The theoretical section discusses the existing work
on legislative co-sponsorship and introduces two research hypotheses about the role of MPs'
ties to similar types of interest groups and ties to interest groups active in the same policy
domain for co-sponsorship activities. The methodological section describes the Swiss
parliament as an empirical setting and explains how our data capture co-sponsorship relations,
MP-group ties and further variables. The empirical section tests the two hypotheses, while the
concluding section put the results in a broader perspective and identifies the next research steps.
5
2. Theoretical framework
Some scholars consider legislative co-sponsorship as a form of ‘cheap talk’ since most bills do
not pass and co-sponsoring a policy proposal is a low-cost action for MPs (Kessler and Krehbiel
1996). Others, however, argue that legislative co-sponsorship serves as a signal to agenda
setters and a form of position taking for constituents (Koger 2003), and that it accurately reflects
the network of peer support between elected representatives (Fowler 2006). Bratton and Rouse
(2011: 423) even claim that focusing on the co-sponsorship stage of the law-making process is
highly relevant to capture how MPs ‘introduce and endorse policy proposals that help them
garner electoral support from relevant constituencies, serve to advance their political career,
and allow them to translate their own policy interests into outcomes’. Thus, co-sponsorship
activities allow MPs to increase their chances for office-seeking, for example to secure their re-
election, as well as for policy-seeking, by promoting their policy preferences (Fenno 1973;
Kirkland 2011; Bratton and Rouse 2011; Ringe et al. 2013). This study sheds new light on the
explanatory factors underlying MPs' co-sponsoring activities.
2.1. Homophily between MPs
The literature that aims at explaining legislative co-sponsorship put forward different types of
homophily. Homophily refers to the phenomenon observed in many different types of networks,
according to which actors who are similar with respect to some characteristic tend to form ties
(McPherson et al. 2001). Theoretical models based on homophily relations assume that MPs
who are similar to each other are more likely to work together to influence the parliamentary
policy agenda and, in particular, to co-sponsor legislative proposals (Bratton and Rouse 2011;
Alemán and Calvo 2013; Baller 2017). Briatte (2016) summarises the findings as follows:
‘Legislators are more prone to co-sponsor the work of other legislators when they share some
characteristics, such as ethnicity (Bratton and Rouse, 2011), gender (Clark and Caro 2013),
6
constituency (Alemán and Calvo 2013; Baller 2017) or committee membership (Kirkland and
Gross 2014; Baller 2017).’ Various types of homophily were assessed in existing research on
legislative co-sponsorship networks.
First, MPs from the same party family are more likely to co-sponsor a policy proposal
than MPs from different party families. MPs from the same party or party family have similar
preferences regarding which policy priorities should be addressed by the parliament and,
furthermore, about which legislative proposals are desirable solutions to these policy problems.
In addition, co-sponsoring the same bills allow MPs to showcase party cohesion about the
policy agenda and to increase their ‘issue ownership’ (Petrocik 1996). A co-sponsorship
strategy eventually contributes to achieve the policy-seeking and office-seeking goals of
political parties (Strøm 1990).
Second, MPs who are members of the same legislative committee are more likely to co-
sponsor a policy proposal than MPs not seating in the same committee (Kirkland and Gross
2014). Serving in the same legislative committee offers many opportunities to interact
personally, to share policy information, to put arguments forwards and to negotiate acceptable
policy solutions. Membership in permanent legislative committees fosters the specialization of
MPs (Gillian and Krehbiel 1987; Searing 1987; Strom 1998), who acquire policy expertise that
grant them power and prestige among party peers and media. Repeated meetings constitute a
favourable context for legislative co-sponsorship between committee members.
Third, it has been shown that the likelihood of co-sponsorship is higher between MPs
who are competing in the same electoral district or region than between MPs living in different
electoral jurisdictions or regions. The rationale for this hypothesis is also quite straightforward.
MPs competing in the same district have a strong incentive to defend the policy interests of
their local constituencies, in order to be re-elected (Alemán and Calvo 2013). For instance, if a
distributive policy aims at delivering benefits to their region of origin, through financial
7
subsidies, the building of new infrastructures or service delivery, then MPs from this district
will probably co-sponsor this legislative proposal. Such an agenda-setting behaviour is highly
attractive for MPs who want to claim the effective representation of their local electoral
constituency.
Fourth, MPs are more likely to mutually co-sponsor their respective policy proposals
than to co-sponsor a policy proposal formulated by an MP who does not return the favour.
Reciprocity is one of the most basic phenomena in social networks of any type and an important
component of social capital and the development of trust (Burt, 2005). MPs’ tendency to
reciprocate is well assessed in the US congress, in both Chambers, and over the years (Fowler
2006: 463-464; Brandenberger, 2018).
2.2. Ties between MPs and interest groups
Previous studies focusing on MPs agenda-setting behaviour have suggested that ties between
MPs and interest groups increase the likelihood of legislative co-sponsorship, but have not
tested that idea empirically. For example, Louwerse and Otjes (2015: 479) claim that external
agents such as lobbyists have an impact on the co-sponsoring attitudes of MPs working on the
same policy area. Other authors also speculate that policy subnetworks at the issue level,
including interest groups, might influence MP agenda-setting activities. Bratton and Rouse
(2011) suggest that a ‘small world’ dynamic underlies informal networks collaboration, and
Fowler (2006) insists on the issue connectedness of MPs. Our arguments below are based on
this idea, which was developed originally by David Truman in The Governmental Process
(1951: 345) and according to which informal legislative groups of MPs and interest
associations' representatives with similar types of information and similar preferences facilitate
the circulation of trustworthy information, the mitigation of gridlocks and the negotiation of
legislative proposals (Fiellin 1962; Ringe et al. 2017).
8
Interest groups try to influence policy-making but, by definition, do not compete for
political office. Consequently, they cannot make binding political decisions and – among other
strategies – must cooperate with MPs in order to influence legislative agenda-setting (Heaney
2014). One lobbying strategy consists in delivering policy information and political intelligence
to MPs. On the other side, elected MPs hold a formal decision power, but they are prone to
interact with interest groups in order to increase their information resources (Pfeffer and
Salancik 2003; Berkhout 2013). By building ties with interest groups, MPs are for instance able
to gather insiders' information about the very nature of a policy problem to be solved by policy-
makers, the expected effectiveness of a potential policy solution and its acceptability by groups'
constituency. Previous scholarship on groups' lobbying has demonstrated that, in the US
Congress, the information transmitted by groups to MPs predominantly concern (problems
with) the feasibility and implementation of public policies (Baumgartner et al. 2009). Yet
groups do not only deliver technical policy expertise to MPs, who can strategically use this
policy information for their own agenda-setting activities. When groups lobby MPs, they also
provide political intelligence about the policy-making process itself, such as procedural advice
to monitor the strategic move of other MPs or interest groups (Hall and Deardorff 2006). This
political intelligence is a key asset for resource-constrained MPs and is thus likely to influence
their co-sponsorship behaviour. The repeated interactions between groups and MPs lead to the
development of informal relationships or formal ties such as groups' membership or even a seat
in a group's board.
We argue that MPs' ties to interest groups generate an additional form of homophily,
besides the ones presented above, e.g., the ones relating to party family or electoral district.
More specifically, we identify two dimensions of interest groups homophily that are likely to
matter for co-sponsorship activities. Interest groups can be similar with respect to their
members' constituencies and related interests, and with respect to the policy domain in which
9
they are active. We assume that MPs tied to either similar types of interest groups or interest
groups from the same policy domain are more likely to collaborate, and to co-sponsor the
legislative proposals.
First, MPs with links to similar types of interest groups (e.g. business groups, as opposed
to citizen groups) are exposed to the same types of information and have developed loyalties to
similar constituencies, i.e. individual or corporate group members. As a consequence, MPs
affiliated to similar groups might promote each other's policy agendas and positions. Our first
hypothesis focuses on the first dimension of homophily, namely similar types of interest groups.
It posits that two MPs with ties to similar types of interest groups are more likely to mutually
co-sponsor their proposals than two MPs with ties to dissimilar types of interest groups. This
means that MP A is more likely to co-sponsor a legislative proposal formulated by MP B, if
both MPs are affiliated to business groups, than if MP A is tied to a business group and MP B
has linkages with a citizen group.
Second, interest groups active in the same policy domain share commonalities with
respect to political agendas and policy priorities that might go beyond commonalities based on
party family or district (Scott 2013). As highlighted by Scott (2015), a policy domain consists
of lobbying organizations that ‘are likely to have close-knit relationships among each other’
(2015: 6-7) and, furthermore, ‘close-knit relationships develop over, and are characterised by,
shared interests over time, and these shared interests over issues should then translate into joint
activity and coalitions’. MPs with strong ties to groups active in the same policy domain benefit
from privileged technical information about the groups' policy portfolio, and tend to develop a
specialization and expertise in that domain. MPs are thus likely to collaborate with peers with
whom they share a domain of policy specialization. For instance, they may work together to
elaborate a legislative proposal that is a political compromise between divergent policy
positions. Hypothesis 2 postulates that two MPs with ties to interest groups active in the same
10
policy domain are more likely to co-sponsor policy proposals than two MPs without ties in that
policy domain. For example, MP A who is a board member of an environmental group may co-
sign a legislative proposal on the promotion of renewable energies introduced by MP B, who is
occupying a leading position in an electric utilities association.
3. Case selection, data and method
Previous scholarship on legislative co-sponsorship has focused either on presidential systems
like the United States (Kessler and Krebhiel 1996), Argentina and Chile (Alemán and Calvo
2013), or on parliamentary systems like the Netherlands (Louwerse and Otjes 2015) and Finland
(Pajala et al. 2016) to test the effects of classical homophily hypotheses regarding party, policy
specialization, or electoral constituency. The present study renews this perspective by analysing
MPs' co-sponsorship activities in Switzerland.
The Swiss parliament is a particularly interesting case to study, since it is a sort of
intermediary case between strong and weak parliaments (Vatter 2014). On the one hand, the
Swiss parliament is institutionally strong. In Switzerland, the government system tends towards
a separation of powers system (Shugart and Carey 1992; Schwarz et al. 2010), in which MPs
are less constrained than in traditional parliamentary systems, i.e. they are not necessarily
expected to align to the party line. Moreover, MPs are granted with powerful agenda-setting
and law-making instruments. Therefore, the Swiss parliament is capable of actually creating
legislation, ‘a classical parliamentary function almost forgotten by some national parliaments’
(Corbett et al. 2007: 9). In particular, MPs have powerful institutional instruments to initiate
legislation at their disposal. Empirically, they give the impetus to about a fourth of legislative
processes, whereas government initiates about half of the processes (Sciarini et al. 2002).
Legislative co-sponsorship is thus more than just a ‘cheap talk’ and can be highly consequential
11
for law-making. In addition, in a comparative perspective, Switzerland displays a very high
share of parliamentary initiatives that are co-sponsored by other legislators (Briatte 2016).
On the other hand, the Swiss parliament is structurally weak. According to the ‘militia
system’, MPs’ involvement is part-time and incidental to a principal professional activity. That
is, MPs are ‘amateurs who combine their professional activities with their parliamentary duties’
(Kriesi 2001: 60). As a result, the Swiss parliament usually meets only four times a year, for a
three-week session (plus possible extraordinary sessions). Moreover, the information and
control resources that Swiss MPs have at their disposal are especially low in comparative
perspective (Vatter 2014). This has favoured the development of MPs' ties to interest groups
(Bailer 2011). However, permanent legislative committees have been institutionalised since
1992 and MPs have become increasingly competent in those policy fields covered by the
specialised committees to which they belong. This has reduced MPs' dependence on interest
groups (Linder 2007), and has reinforced parliament's power in decision-making processes
(Sciarini 2014; Sciarini et al. 2015). Therefore, the Swiss parliament offers an ideal setting for
investigating the relationships between interest groups and elected MPs, as well as the potential
influence of groups' advocacy on legislative co-sponsorship.
3.1. Networks of co-sponsorship
The dependent variable of this analysis is the network of co-sponsorship among MPs in the
Swiss Lower Chamber, the National Council. In order to construct this network, we collected
all legislative proposals submitted by MPs of the Lower Chamber between October 2011 and
November 2015, covering thus one entire legislative period. Swiss MPs have different types of
instruments at their disposal to initiate legislation. First, parliamentary initiatives represent the
strongest instrument available to MPs, as these proposals allow them to launch a legislative
process and draft a bill themselves. Second, motions allow the parliament to demand the
12
government to take legislative action. Third, postulates require the government to deliver a
report on a given issue. MPs can submit any of these parliamentary proposals in their name,
and the parliamentary plenum decides on whether to accept or not a given parliamentary
proposal. In the legislative period 2011-2015, the 226
1
Swiss MPs handed in a total of 2108
legislative proposals (with a minimum of zero proposals by 17 MPs and a maximum of 49
proposals by one MP). On average, a MP handed in about 9.5 interventions during this period
(median value of 7 interventions). A ‘standard’ proposal has received about 18 co-sponsors, on
average.
We then coded the co-signatures of other MPs on a given proposal to construct our
network among MPs. The resulting network of co-sponsorship is a valued network, as MP A
can sign one or many proposals of MP B. Indeed, while most MPs do not co-sponsor many
other MPs, one MP signed a maximum of 34 interventions of another peer. Rather than the
absolute number of signatures, however, the relative number of signatures is more appropriate
to express co-sponsorship ties (Kirkland 2011). We thus calculate the percentage of MP B's
interventions that are co-signed by MP A. If MP A co-signs a high share of MP B's interventions
– regardless of whether MP B submitted one or 20 interventions – this can be interpreted as a
strong relationship. The relationship is weaker if only a small share of MP B's interventions are
signed by MP A. On average, MPs sign nine percent of each other's interventions. We thus
dichotomise the network at a value of 10 percent, but test for robustness of results with higher
thresholds. This network is composed of 221 nodes, 10401 edges, and a resulting density of
0.21. Figure 1 shows the resulting network, with node colours indicating party family, and node
size indicating the number of ties to interest groups dealing with economic issues, as an
example.
--- Figure 1 about here ---
13
3.2. Independent variables
Our main independent variables are coded as follows:
Ties to interest groups: Swiss MPs are requested to declare their affiliation (i.e.
membership, seat in a group’s board) with interest groups. The resulting official ‘register of
interests” is a rich source of observational data that allows matching individual MPs with
specific groups. Gava et al. (2017) show that the average number of interest ties per MP has
more than doubled in the last decade, from 3.5 in 2000 to 7.6 in 2011. Behind these values,
there is considerable heterogeneity across individual MPs. Only 14 MPs did not declare any
kind of ties to groups, whereas, at the other extreme, a single MP dominated the ranking of
affiliations with ties to 51 different groups.
Types of interest groups: We capture the interest groups' diversity by means of seven
subcategories, based on the categorization developed by the Interarena project, which is a well-
tested instrument to capture the diversity of interest groups' population (Binderkrantz et al.
2015):
2
(1) Business groups (e.g. Swiss Bankers' Association); (2) professional groups formed
by individuals who share the same occupation (e.g. Swiss Medical Association); (3) Unions;
(4) Identity groups that primarily seek benefits for their own members or restrained
constituencies (e.g. Swiss Automobile Club); (5) Public interest groups that focus on the
attainment and protection of common goods, as for example consumer groups or environmental
or humanitarian organizations; (6) religious groups including spiritual organizations, such as
churches or other associations focused on the support of religious communities (e.g. Swiss
Evangelical Alliance); and (7) institutional groups representing Swiss cantons or municipalities.
For our present purposes we distinguish between three types of groups that differ with respect
to constituencies and related policy preferences. That is, we aggregate the first two
subcategories into a ‘Business groups’ category, the subcategories 3 to 6 constitute together the
14
‘Citizen groups’ category, and the subcategory 7 (i.e., ‘Institutional groups”) constitutes its own
category. For each MP, the number of ties to groups of each category was added up. Interest
group homophily is then measured as the absolute difference between the numbers of respective
ties: The smaller the difference, the more similar two MPs' profiles.
Interest groups' main domain of activities: For each interest group, we coded the main
domain of activity based on the classification scheme of the Comparative Agendas Project
(CAP).
3
This classification of policy domains, which identifies 20 main issue topics, has been
widely used in comparative policy and politics, in order to assess the relative issue attention in
a wide range of agenda, from the political (parliamentary or government) arena to the media
arena (Green-Pedersen and Walgrave 2014). We again aggregated the 20 main issue topics in
a more limited – and more tractable – number of categories: economic issues (macro-economy,
taxes, labor, finances, agriculture), social issues (health, education, environment), law and order
(crime, individual rights, political institutions) and foreign affairs (international relations,
European integration, defense). Again, as with interest group types, the number of ties to groups
of each category was added up for each MP. From that we then calculated the absolute
difference between the numbers of respective ties to measure the degree of similarity of MPs'
profile towards interest groups in a given policy domain.
Additional variables in the basic models, which do not refer to our two hypotheses but
correspond to the set of more established hypotheses in the respective literature on homophily
relationships, are operationalised as follows. First, party family homophily is introduced by a
nodematch term assessing whether there is a match between MPs belonging to the same party
family. We distinguish between left parties (most importantly Social Democrats, Green Party),
moderate right parties (most importantly Free Democrats, Christian Democrats), and far right
parties (most importantly Swiss People’s Party) according to the three most important camps
in Swiss politics (Sciarini et al., 2015). Second, legislative committee homophily is assessed
15
with an edge covariate term representing the number of joint committee memberships of two
MPs. Third, regional homophily is also assessed through a nodematch term, testing if there is a
match between MPs from French (including Italian) and German speaking cantons. Fourth,
reciprocity is included in the models as an endogenous network parameter, assessing the degree
to which co-sponsorship ties from MP A to MP B do coexist with co-sponsorship ties from MP
B to MP A. Further controls are a) a nodematch term to capture gender homophily (Clark and
Caro 2013), b) an in-degree term for assessing whether legislative proposals by more senior
MPs get signed more than proposals by more junior MPs
4
, and c) a GWESP term that assesses
an endogenous network mechanism in that it accounts for shared partners of two MPs that are
themselves connected in the network of co-sponsorship.
Hypotheses are tested based on two types of parameters assessing interest group
homophily, one regarding the three types of interest groups (business, citizen, institutional), and
one regarding the four policy domains (economic, social, law & order, and foreign policy).
More specifically, the respective parameter assesses the opposite of homophily, that is, whether
MPs with dissimilar interest group profiles tend to co-sponsor each other or not. The respective
parameter (absdiff) assesses whether the probability that MPs mutually co-sponsor their
proposals relates to the difference between MPs in terms of the number of ties to interest groups
of a given type or from a given policy domain. A negative parameter – indicating that the larger
the difference between MPs, the less they co-sponsor each other – would thus lend support to
our hypotheses. Moreover, we need to control for the number of MP's ties to interest groups of
a given type or from a given policy domain. Specifically, the number of ties is likely to be
positively associated with the number of proposals that a MP co-sponsors. This is controlled
for by the out-degree parameter. Similarly, the in-degree parameter controls for the likelihood
that MPs with a high number of ties to interest groups of a given type or from a given policy
domain attract more co-sponsors for their proposals.
16
3.3. Exponential Random Graph Models
We test our hypotheses by estimating Exponential Random Graph Models (Robins et al. 2007;
Lusher et al. 2013; Cranmer et al. 2017). ERGMs allow for statistical inference on network
data, which by definition are non-independent. Non-independency among observations in
network data means that the probability of a co-sponsorship tie between two MPs might depend
upon the structural properties of the network in which the two MPs are embedded. Standard
regression models are unable to take this dependency into account and would erroneously
attribute explanatory power to exogenous variables (Cranmer and Desmarais 2011; Lusher et
al. 2013). Given the dependency among observations, error terms would be correlated across
observations, standard errors would be too small, and p-values for exogenous variables too
optimistic.
In order to avoid the assumption of relational independence, ERGMs model the
probability of observing a given configuration of the network, as compared to all other possible
network configurations with the same number of nodes and network density (Cranmer and
Desmarais 2011). The structure of the network is modelled based on actor-level variables (in-
and out-degree parameters in our models), dyadic variables (homophily parameters in our
models), and endogenous network structures (reciprocity parameter in our models). The relation
between the probability of a network m and the network statistics in can be expressed by the
following formula, where is the vector of k parameters that describe the dependence of P(Ym)
on the network statistics in (Cranmer and Desmarais 2011) :
ERGMs integrate an exponential family form log-likelihood function. Due to the very
high number of possible network configurations, computing the exact maximum likelihood is
17
however computationally too demanding (Cranmer and Desmarais 2011). Therefore, we
estimate ERGMs using Markov Chain Monte Carlo Maximum Likelihood (MCMC-MLE),
which approximates the exact likelihood by relying on a sample from the range of possible
networks to estimate the parameters. Goodness of fit and MCMC diagnostics for all models
appear in the Appendix.
4. Analysis
The first column of table 1 presents the results of the basic model, which includes the main
classical homophily variables, as well as further controls. The second column shows the results
of a model additionally including homophily parameters relating to each of the three types of
groups, as well as in- and out-degree parameters for each of the three types of interest groups.
Results for the model including domains of interest groups appear in table 2.
5
Results from the basic model confirm the traditional hypotheses of legislative co-
sponsorship. The strongest effect can be attributed to homophily between party families. Not
surprisingly, MPs from the same parliamentary group tend to co-sponsor their proposals. Also
committee membership is an important factor, as MPs sitting in the same legislative committee
co-sponsor their proposals at higher rates than MPs that do not sit in the same committee. Also
region homophily appears as an important predictor for co-sponsorship among Swiss MPs, as
MPs from the same region tend to co-sponsor their proposals. Finally, the significantly positive
reciprocity parameter shows that – independently of other factors – MPs tend to co-sponsor the
proposals of other MPs who also co-sponsor their proposals.
---Table 1 about here ---
Hypothesis 1 claims that MPs with ties to similar interest groups tend to mutually co-
sponsor their proposals as they are exposed to the same type of information and interests. The
18
results of table 1 provide partial support for the hypothesis. They show that MPs with similar
numbers of ties to business groups have a higher likelihood to co-sponsor their respective
proposals: The more similar the number of ties to business groups between two MPs, the more
they mutually co-sign their legislative proposals. The same is true for citizen groups. Shared
affiliations to institutional groups, by contrast, are not related to the co-sponsorship behaviour
of MPs. Institutional groups mainly defend federalist issues, (i.e. division of power across levels
of government) as they represent cantons and municipalities. These actors might have other,
more direct access options to parliamentary agenda-setting and policy-making at the national
level. Still, for two out of three types of interest groups under consideration, there is evidence
that the interest group profile of MPs matters for understanding their co-sponsorship behaviour.
The in-degree and out-degree parameters are included as controls, to clearly isolate potential
homophily effects.
6
Hypothesis 2 relates to another dimension of interest groups, namely to the MPs'
affiliations to interest groups active in a given policy domain. We expected to observe
homophily among MPs with relations to interest groups active in the same policy domain,
whereby we distinguish between four broad policy domains (i.e. economic issues, social issues,
law and order issues, and foreign policy issues). The hypothesis is partly confirmed. We see
from table 2 that interest group homophily plays a role in two out of four policy domains. The
more similar two MPs are with respect to their number of affiliations to groups dealing with
economic issues and law & order issues, the more they tend to co-sponsor each other's
proposals. By contrast, our results reveal no effect for groups dealing with social and foreign
policy issues. In the Swiss context, social policies and related redistributive issues lead to vivid
political controversies (e.g. private versus public health care system and funding). Similarly,
foreign affairs are also highly conflictual. One emblematic example is the recurrent conflict
between supporters and opponents to Switzerland's European integration policy. On such
19
sensitive issues that articulate entrenched political cleavages, policy compromises are
especially difficult to reach.
7
This presumably accounts for the non-significant relationship
between MPs' ties to interest groups active in the same policy domain and the likelihood of co-
sponsorship.
Additional controls appear at the end of the tables: First, a weak, but still significant
homophily effect can be observed for gender: MPs of the same gender tend to co-sponsor their
proposals, independently of other factors. Second, the significantly negative seniority parameter
indicates that the longer an MP sits in parliament, the less co-sponsors (s)he gets on her or his
proposals. Finally, the GWESP parameter
8
accounting for triadic closure is significant and
positive in all but the basic model.
Overall, we thus see a tendency for MPs with similar interest group profiles to co-
sponsor their proposals, but only for given types or policy domains a group is active in.
Regarding the size of the effect, the parameters for interest group homophily from table 1 or 2
suggest that their effect is less strong than the homophily effects related to party family,
legislative committees, or regions. The size of effects, that is, the probability of observing a tie
can be obtained by calculating the exponential function of effects.
9
Thus, for example, the
probability of co-sponsorship increases by 67% (e(0.69)/(1+ e(0.69)) = 0.67) if two MPs are from
the same region, as compared to two MPs that are not from the same region. Figure 2 shows
the marginal effect plots of the three interest group variables that have a significant effect in
our models (Czarna et al. 2016; Leifeld et al. 2018). It shows how the probability of a tie
decreases with an increase in the difference between MPs’ respective interest group profile.
10
For example, the probability of two MPs having a co-sponsorship tie decreases from 23% to
13.5% if the difference in their business interest group profile increases from 0 to 11 (left plot
in Figure 2).
20
--- Figure 2 about here ---
Besides the considerable size of these effects, the interest group variables contribute to
the explanatory power of the model, as indicated by the AIC and BIC coefficients that are lower
in both homophily models as compared to the basic model. Also, the fact that effects of interest
group homophily show up while controlling for the other, more traditional factors influencing
co-sponsorship – and especially party family and committee homophily – can only reinforce
our confidence in our results.
The results are also robust to a higher threshold for the percentage of interventions
signed (minimum of 50 percent, instead of 10 percent). The effects of homophily of interest
group profile are similar or slightly stronger, with the exception of the ‘law & order’ domain
homophily, which loses significance. Furthermore, dropping committee homophily – which,
like interest group homophily effects, is related to the issue specialization of MPs – does not
affect the results, but slightly reduces model fit. This suggests that the two interest group
homophily variables capture effects beyond shared committees and, therefore, complement the
latter variable (see also Varone et al. 2018 for similar findings).
--- Table 2 about here ---
5. Conclusion
This study confirms previous scholarship on legislative co-sponsorship by demonstrating that
traditional homophily relationships among MPs are strong predictors of co-sponsoring
behavior. MPs belonging to the same party family, seating in the same legislative committee,
or representing the same electoral district are more likely to co-sign legislative proposals. Thus,
our findings from the Swiss case, which is neither a presidential nor a parliamentary system,
and where the parliament is a sort of intermediary case between a strong and a weak parliament,
21
improves the external validity of homophily models that were already tested in the US
presidential system or in European parliamentary democracies.
On the more innovative side, we find strong support for the hypothesis that MPs with
ties to similar types of interest groups are more likely to co-sponsor legislative proposals, and
we find partial support for the hypothesis that MPs' ties to interest groups specialised in similar
policy domains increase the likelihood of co-sponsorship. These new findings improve the
explanation of legislative co-sponsorship. Moreover, the fact that they hold while controlling
for the more classical forms of homophily – and more especially party, committee or district
homophily – can only reinforce their relevance.
Compared to the ties existing between MPs based on fundamental similarities such as
party membership, linkages between MPs and interest groups should rather be interpreted as
weak ties (Granovetter 1973, 1983). Indeed, they bring together policy actors with diverse
professional background, personal objectives, organizational incentives, but also different roles
in the policy-making process and levels of access to institutional venues. However, such weak
ties have a high value, since they foster the exchange of non-redundant policy information
between MPs and groups' representatives, the confrontation of different points of view on a
policy domain and, ultimately, the development of creative policy solutions. These policy-
specific cooperation ties increase information dissemination and diversity, and facilitate the
negotiation of policy compromises. As highlighted by Kirkland (2011: 889), ‘(the) value of
weak ties is a result of their novelty of information or influence’. Assuming that both business
and citizen groups are key informants for MPs and provide technical information and political
intelligence, MP-group linkages stimulate informal cooperation between MPs and the co-
sponsoring of legislative proposals. Yet our results suggest that the correlation between MPs'
ties to interest groups active in a similar policy domain and legislative co-sponsorship is weaker
22
if that domain articulates a major political conflict (i.e., in the Swiss case conflicts on the
redistributive, left-right dimension or on the integration-demarcation dimension).
From a normative point of view, our findings deliver an encouraging and optimistic
message. Indeed, MPs co-sponsor legislative proposals introduced by MPs from the same party
family and electoral constituency. So doing, they endorse a delegate role and represent the
policy preferences of their voters, in line with the ideal-typical view of substantive
representative democracy (Miller and Stokes 1963). MPs' ties to interest groups do not push
elected representatives to defect from their electoral constituency when they co-sign a
legislative proposal, but they may foster the elaboration of legislative proposals that are
supported across party lines and electoral districts. In other words, interest groups may play a
positive role in transmitting the importance of policy issues for various societal stakeholders in
general, and for legislative committees in particular.
Further interesting research questions regarding the relationship between MPs' interest
group profile and their co-sponsorship activity abound. First, further to controlling for classical
homophily variables (i.e. party family, legislative committee, electoral district and gender), we
might also consider the interaction effects between the MPs' specialization in a policy domain,
on the one hand, and their respective party affiliation or electoral district, on the other hand. For
instance, MPs who are ideologically close from each other and, thus, who have similar policy
positions, will probably co-sponsor one another if they specialise in the same policy domain.
Specialization in a legislative committee has a decisive influence on the policy issues on which
the MPs are going to work together and, at the same time, party affiliation has an impact on
choosing with whom (i.e. which committees members) MPs will co-sponsor a policy proposal
(Louwerses and Otjes 2015:481).
The impact of MPs' policy specialization is also highly relevant for the link between
MPs-groups ties and legislative co-sponsorship. In the Swiss case, it has been shown that MPs
23
seating in a legislative committee hold relevant group ties. There is a topical match between
MPs affiliation to groups from a specific policy domain and the substantive area of competence
of the legislative committee they are a member of. For example, MPs seating in the
environment, infrastructure and energy committee have formal contacts with business
associations (e.g. associations representing private electricity producers, public utilities or
renewable technologies), but also with environmental and consumers associations (Gava et al
2017). Furthermore, this topical congruence is the consequence of both elected MPs carrying
relevant group ties into the legislative committees, and MPs being proactively recruited by
interest groups once they are assigned to a specific committee (Eichenberger and Mach 2017).
In sum, upcoming studies should look in depth at interaction effects between MPs' policy
specialization, traditional homophily variables and MPs' ties to interest groups.
Finally, we acknowledge that explaining why MPs co-sponsor each other's bills is not
the same as explaining why co-sponsored bills are eventually accepted by Parliament. It makes
sense to investigate how and why some MPs work together to introduce policy proposals on
which all MPs will then vote (Louwerese and Otjes 2015). But the natural step forward should
be to assess whether the composition of co-sponsorship network in general, and MPs'
affiliations to interest groups in particular, might also explain the success of co-sponsored
legislative proposals. Previous scholarship on whether the number and types of co-sponsors has
a positive impact on bill passage delivered rather mixed evidence (see Bratton and Rouse 2011;
Craig 2015). Integrating MPs-groups ties as additional variable in these explanatory models is
a promising research avenue.
24
References
Alemán, Eduardo and Ernesto Calvo (2013). ‘Explaining Policy Ties in Presidential
Congresses: A Network Analysis of Bill Initiation Data’, Political Studies, 61, 356-
77.
Bailer, Stefanie (2011). ‘People's voice or information pool? The role of, and reasons for,
parliamentary questions in the Swiss parliament’, The Journal of Legislative Studies,
17(3): 302-14.
Baller, Inger (2017). ‘Specialists, party members, or national representatives: Patterns in co-
sponsorship of amendments in the European Parliament’, European Union Politics,
online.
Baumgartner, Frank, and Bryan Jones (1993). Agendas and Instability in American Politics.
Chicago: Chicago University Press.
Baumgartner, Frank, Jeffrey Berry, Marie Hojnacki, Beth Leech, and David Kimball (2009).
Lobbying and policy change: Who wins, who loses, and why. Chicago: University of
Chicago Press.
Berkhout, Joost (2013). ‘Why interest organizations do what they do: Assessing the explanatory
potential of ‘exchange’ approaches’, Interest Groups & Advocacy, 2(2): 227-50.
Binderkrantz, Anne, Peter Christiansen, and Helene Pedersen (2015). ‘Interest Group Access
to the Bureaucracy, Parliament, and the Media’, Governance, 28(1): 95–112.
Bratton, Kathleen and Stella Rouse (2011). ‘Networks in the Legislative Arena: How group
Dynamics Affect Cosponsorship’, Legislative Studies Quarterly, 36 (3): 423-60.
Briatte, François (2016). ‘Network patterns of legislative collaboration in twenty parliaments’,
Network Science, 4 (2): 266-71.
Bundi, Pirmin, Frédéric Varone, Roy Gava, and Thomas Widmer (2016). ‘Self-Selection and
Misreporting in Legislative Surveys’, Political Science Research and Methods, online.
25
Brandenberger, Laurence (2018). ‘Trading favors—Examining the temporal dynamics of
reciprocity in congressional collaborations using relational event models’, Social
Networks 54: 238-53.
Burt, Ronald (2005). Brokerage and Closure: An Introduction to Social Capital. Oxford/New
York: Oxford University Press.
Clark, Jennifer, and Veronica Caro (2013). ‘Multimember Districts and the Substantive
Representation of Women: An Analysis of Legislative Cosponsorship Networks’,
Politics & Gender, 9:1-30.
Craig, Alison (2015). Lone Wolves and Team Players: Policy Collaboration Networks and
Legislative Effectiveness in the House of Representatives. Unpublished manuscript,
Ohio State University.
Corbett, Richard, Francis Jacobs, and Darren Neville (2007). The European Parliament.
London: John Harper Publishing.
Cranmer, Skyler, and Bruce Desmarais (2011). ‘Inferential Network Analysis with Exponential
Random Graph Models’, Political Analysis, 19: 66-86.
Cranmer, Skyler, Leifeld, Philip, Scott McClurg, and Meredith Rolfe (2017). ‘Navigating the
Range of Statistical Tools for Inferential Network Analysis’, American Journal of
Political Science, 61 (1): 237-51.
Czarna, Anna, Philip Leifeld, Magdalena Smieja, Michael Dufner, and Peter Salovay (2016).
‘Do Narcissism and Emotional Intelligence Win Us Friends? Modeling Dynamics of
Peer Popularity Using Inferential Network Analysis’, Personality and Social
Psychology Bulletin, 42 (11), 1588-99.
Eichenberger, Steven, and André Mach (2017). ‘Formal ties between interest groups and
members of parliament: Gaining allies in legislative committees’, Interest Groups &
Advocacy, 6(1): 1-21.
26
Fenno, Richard (1973). Congressmen in Committees. Boston: Little, Brown.
Fiellin, Alan (1962). ‘The Functions of Informal Groups in Legislative Institutions’, The
Journal of Politics, 24(1): 72-91.
Fowler, James (2006). ‘Connecting the Congress: A study of Cosponsorship Networks’,
Political Analysis, 14: 456-87
Gava, Roy, Frédéric Varone, Andé Mach, Stephen Eichenberger, Julien Christe, and Corinne
Chao-Blanco (2017). ‘Interests groups in Parliament: Exploring MPs’ interest
affiliations (2000-2011)’, Swiss Political Science Review, 23(1): 77-94.
Giger, Nathalie and Heike Klüver (2016). ‘Voting Against Your Constituents? How Lobbying
Affects Representation’, American Journal of Political Science 60(1): 190-205.
Gillian, Thomas, and Keith Krebhiel (1987). ‘Collective Decisionmaking and Standing
Committees: An Informational Rationale for Restrictive Amendment Procedures’,
Journal of Law, Economics & Organization, 3(2): 287-335.
Goodreau, Steven, Mark Handcock, David Hunter, Carter Butts, and Martina Morris (2008).
‘A statnet Tutorial’, Journal of Statistical Software, 24(9): 1-26.
Granovetter, Mark (1973). ‘The Strength of Weak Ties’, The American Journal of Sociology,
78 :1360-80.
Granovetter, Mark (1973). ‘The Strength of Weak Ties: A Network Theory Revisited’,
Sociological Theory, 1: 201-33.
Green-Pedersen, Christoffer and Stefaan Walgrave (eds). Agenda Setting, Policies and Political
Systems: A Comparative Approach. Chicago: University of Chicago Press.
Hall, Richard and Alana Deardorff (2006). ‘Lobbying as legislative subsidy’, American
Political Science Review, 100(1): 69-84.
Heaney, Michael (2014). ‘Multiplex Networks and Interest Group Influence Reputation: An
Exponential Random Graph Model’, Social Networks, 36: 66-81.
27
Kessler, Daniel and Keith Krebhiel (1996). ‘Dynamics of cosponsorship‘, American Political
Science Review, 90(3): 555-66.
Kirkland, Justin (2011). ‘The Relational Determinants of Legislative Outcomes: Strong and
Weak Ties Between Legislators’, The Journal of Politics, 73(3): 887-98.
Kirkland, Justin, and Justin Gross (2014). ‘Measurement and theory in legislative networks:
The evolving topology of Congressional collaboration’, Social Networks, 36: 97-109.
Koger, Gregory (2003). "Position Taking and Cosponsorship in the U.S. House." Legislative
Studies Quarterly, 28(2): 225-246.
Kriesi, Hanspeter (2001). ‘The federal parliament: The limits of institutional reform’, West
European Politics, 24(2): 59-76.
Leifeld, Philipp, Skyler Cranmer, and Bruce Desmarais (2018). ‘Temporal Exponential
Random Graph Models with btergm: Estimation and Bootstrap Confidence Intervals’,
Journal of Statistical Software, 83 (6): 1-36.
Linder, Wolf (2007). Schweizerische Demokratie. Bern: Paul Haupt Verlag.
Louwerse, Tom and Simon Otjes (2015). ‘The Impact of Parliamentary Specialization on
Cosponsorship’, The Journal of Legislative Studies, 21 (4): 476-94.
Lusher, Dean, Johan Koskinen, and Garry Robins (2013). Exponential Random Graph Models
for Social Networks. Cambridge: Cambridge University Press.
McPherson, Miller, Lynn Smith-Lovin, and James Cook (2001). ‘Birds of a Feather:
Homophily in Social Networks’, Annual Reviews of Sociology, 27: 415-44.
Miller, Warren and Donald Stokes (1963). ‘Constituency Influence in Congress’, American
Political Science Review, 57(1): 45-56.
Morris, Martina, Mark Handcock, David Hunter (2008). ‘Specification of Exponential-Family
Random Graph Models: Terms and Computational Aspects’, Journal of Statistical
Software 24: 1-24.
28
Pajala, Antti, Elena Puccio, Jyrki Piilo, and Michele Tumminello (2016). ‘Party Comrades and
Constituency Buddies: Determinants of Private Initiative Cosponsor Networks in a
Parliamentary Multiparty System’, https://arxiv.org/pdf/1612.06648v1.pdf.
Petrocik, John (1996). ‘Issue Ownership in Presidential Elections, with a 1980 Case Study’,
American Journal of Political Science, 40(3): 825-50.
Pfeffer, Jeffrey, and Gerald Salancik (2003). The external control of organizations: A resource
dependence perspective. Stanford: Stanford University Press.
Rasmussen, Anne, and Gert-Jan Lindeboom (2013). ‘Interest group-party linkage in the twenty-
first century: Evidence from Denmark, the Netherlands and the United Kingdom’,
European Journal of Political Research, 52: 264-89.
Ringe, Nils, Jennifer Victor, and Justin Gross (2013). ‘Keeping Your Friends Close and Your
Enemies Closer? Information Networks in Legislative Politics’, British Journal of
Political Science, 43: 601-28.
Ringe, Nils, Jennifer Victor, and Wendy Tam Cho (2017). ‘Legislative networks’, in Jennifer
Victor, Alexander Montgomery, and Mark Lubell (eds.), The Oxford Handbook of
Political Networks. Oxford: Oxford University Press, 471-98.
Robins, Gary, Pip Pattison, Yuval Kalish, and Dean Lusher (2007). ‘An introduction to
exponential random graph (p*) models for social networks’, Social Networks, 29: 173-
91.
Schwarz, Daniel, André Bächtiger, and Georg Lutz (2010). ‘Switzerland: Agenda-Setting
Power of Government in a Separation-of-Powers Framework’, in Bjorn Erik Rasch
and George Tsebelis (eds), The Role of Governments in Legislative Agenda Setting.
London: Routledge, 127-43.
Sciarini, Pascal (2014). ‘Eppure si muove: The changing nature of the Swiss consensus
democracy’, Journal of European Public Policy, 21 (1): 116-32.
29
Sciarini, Pascal, Sarah Nicolet, and Alex Fischer (2002). ‘L'impact de l'internationalisation sur
les processus de décision en Suisse: Une analyse quantitative des actes législatifs 1995-
1999’, Swiss Political Science Review, 8 (3/4): 1-34.
Sciarini, Pascal, Manuel Fischer, and Denise Traber (2015). Political Decision-Making in
Switzerland. The Consensus Model under Pressure. Basingstoke / New York: Palgrave
Macmillan.
Scott, John (2013). ‘Social Processes in Lobbyist Agenda Development. A Longitudinal Social
Network Analysis of Interest Groups and Legislation’, Policy Studies Journal, 44(4):
608-35.
Scott, John (2015). The social process of lobbying. Cooperation or collusion? New York /
London: Routledge.
Searing, Donald (1987). ‘New Roles for Postwar British Politics: Ideologues, Generalists.
Specialists, and the Progress of Professionalization in Parliament’, Comparative
Politics, 19(4): 431-52.
Shugart, Mathew, and John Carey (1992). Presidents and Assemblies. Cambridge: Cambridge
University Press.
Strøm, Kaare (1990). ‘A Behavioral Theory of Competitive Political Parties’, American
Journal of Political Science, 34(2): 565-98.
Strøm, Kaare (1998). ‘Parliamentary Committees in European Democracies’, The Journal of
Legislative Studies, 4(1): 21-59.
Truman, David (1951). The Governmental Process. New York: Knopf.
Varone, Frédéric, Karin Ingold, Charlotte Jourdain, and Volker Schneider (2017). ‘Studying
policy advocacy through social network analysis’, European Political Science, 16:
322-36.
30
Varone, Frédéric, Pirmin Bundi and Roy Gava (2018). ‘Policy evaluation in parliament: Interest
groups as catalysts’, International Review of Administrative Sciences,
https://doi.org/10.1177/0020852317750461
Vatter, Adrian. 2014. Das politische System der Schweiz. Stuttgart: Nomos UTB.
Wonka, A. (2017) "German MPs and interest groups in EU multilevel policy-making: The
politics of information exchange", West European Politics 40(5): 1004-24.
31
1
The Lower Chamber contains 200 MPs, and 26 MPs replaced others during the legislative period. 5 MPs were
excluded from our analyses because of missing data on their interest group relations.
2
See http://interarena.dk.
3
See http://www.comparativeagendas.net.
4
This variable is assessed as the number of months a MP has been a member of parliament until the end of the
legislature (fall session of October 2015).
5
The percentage of the variance due to MCMC is 0 for all parameters in the Interest Group Type and the Interest
Group Domain Models. 1% of the variance is due to MCMC for the exogenous variables in the Basic Model.
MCMC diagnostics and goodness-of-fit diagnostics for all three models appear in the Appendix. The models do
not well simulate the dyadwise and edgewise shared partner distributions, but they reconstruct the basic form of
the respective distributions.
6
Without controlling for in-degree and out-degree of the respective attributes, identified homophily effects could
simply be due to the specific activity or popularity of given types of nodes in the network.
7
The conflict over redistributive issues and the conflict on the desired level of integration versus demarcation into
the European Union are the two main conflict lines among the Swiss elite according to a study of the most
important decision-making processes of the early 2000s (Sciarini 2014, Sciarini et al. 2015).
8
Geometrically weighted dyad-wise and edge-wise partner statistics (GWDSP, GWESP) are the most appropriate
and commonly used parameters to assess triadic closure in networks, and their geometric form reduces
convergence problems (Morris et al. 2008). Whereas GWDSP assesses the basic probability of two nodes to have
a share partner, GWESP assesses whether two linked nodes tend to have shared partners. In our models, we rely
on a GWESP parameter only. Including a GWDSP parameter does improve the goodness of fit, but the respective
parameter displays poor convergence statistics. For the GWESP, we rely on a weight parameter (alpha value) of
0.5. We have tested lower and higher values, also in combination with the GWDSP parameter, but respective
models showed poorer fit or convergence statistics.
9
The size of effects can only be interpreted assuming that all other covariate values are the same, that is, that the
rest of the network is fixed (conditional log-odds). For more detailed explanations, see Goodreau et al. (2008).
10
The maximum dissimilarity value varies between the three types of interest groups and between the four policy
domains of interest groups. It is calculated as the maximal observable difference between a MP with zero ties to a
given type or policy domain of interest groups and the MP(s) with the maximal number of ties to the corresponding
32
type or policy domain of interest groups. Maximal dissimilarity values range from 5 points for interest groups in
the foreign policy domain and 29 points for interest groups in the social policy domain.
33
Table 1. Basic model and IG Type Model
Basic Model
IG Type Model
Party family homophily
1.51 (0.03)
1.51 (0.03)
Party family (Left)
0.29 (0.02)
0.27 (0.02)
Party family (Far right)
0.30 (0.02)
0.35 (0.02)
Joint committees
0.28 (0.03)
0.28 (0.03)
Region homophily
0.69 (0.03)
0.69 (0.03)
Region (German speaking)
-0.14 (0.02)
-0.17 (0.02)
Reciprocity
1.60 (0.04)
1.58 (0.04)
H1 : Difference Business IG
-0.05 (0.01)
Business IG out
0.01 (0.01)
Business IG in
0.04 (0.01)
H1 : Difference Citizen IG
-0.01 (0.00)
Citizen IG out
0.04 (0.00)
Citizen IG in
0.01 (0.00)
H1 : Difference Institutional IG
-0.01 (0.01)
Institutional IG out
-0.03 (0.01)
Institutional IG in
0.04 (0.01)
Edges
-3.74 (0.51)
-4.11 (0.51)
Seniority in
-0.03 (0.00)
-0.03 (0.00)
Gender homophily
0.17 (0.03)
0.18 (0.03)
Gender (Male)
-0.14 (0.02)
-0.13 (0.02)
GWESP
0.58 (0.29)
0.74 (0.29)
AIC
40823.63
40536.39
BIC
40929.13
40721.02
Values in bold indicate significant parameters (p-value of 0.05 or lower).
34
Table 2. IG Domain Type
IG Domain Type
Party family homophily
1.52 (0.03)
Party family (Left)
0.30 (0.02)
Party family (Far right)
0.31 (0.02)
Joint committees
0.29 (0.02)
Region homophily
0.69 (0.03)
Region (German speaking)
-0.15 (0.02)
Reciprocity
1.60 (0.04)
H2 : Difference IG economic issues
-0.03 (0.00)
IG economic issues out
0.00 (0.00)
IG economic issues in
0.04 (0.00)
H2 : Difference IG law&order issues
-0.07 (0.01)
IG law&order out
0.06 (0.01)
IG law&order in
0.07 (0.01)
H2 : Difference IG social issues
-0.00 (0.00)
IG social issues out
0.01 (0.00)
IG social issues in
-0.01 (0.00)
H2: Difference IG foreign policy issues
-0.02 (0.02)
IG foreign policy issues out
0.09 (0.02)
IG foreign policy issues in
-0.21 (0.02)
Edges
-4.23 (0.51)
Seniority in
-0.03 (0.00)
Gender homophily
0.17 (0.03)
Gender (Male)
-0.16 (0.02)
GWESP
0.88 (0.30)
AIC
40450.56
BIC
40661.56
Values in bold indicate significant parameters (p-value of 0.05 or lower).
35
Figure captions
Figure 1: Network of co-sponsorship among Swiss MPs
36
Figure 2: Marginal effect plots
37
Appendix
Figure A1: Goodness of fit, Basic Model
Figure A2: MCMC diagnostics, Basic Model
Figure A3: Goodness of fit, IG Type Model
Figure A4: MCMC diagnostics, IG Type Model
Figure A5 Goodness of fit, IG Domain Model
Figure A6: MCMC diagnostics, IG Domain Model
