![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
Studies of policy attention find only mixed support for a partisan impact, instead showing that policy attention reacts more to world events. Yet, a rigorous examination of the ways in which change in the partisan composition of government matters for the distribution of policies across issues has yet to be completed in a cross-national framework. Combining data on policy output from the Comparative Agendas Project, the authors present a detailed investigation of parties’ effects on agenda stability in six advanced industrial democracies over time. The authors consider parties as dynamic organizations by arguing that parties’ organizational characteristics and goals interact with their electoral context to determine their impact on policy attention. The results show that parties’ influence on the policy agenda depends on economic conditions, the type of government, the government’s seat share, and the number of parties in the governing cabinet, particularly following a major transition in government.
Acknowledgements
Portions of this cross-project collaborative work (Projects C1 and C2) were supported by the Collaborative Research Center 884 Political Economy of Reforms, funded by the German Research Foundation. We presented a draft of this study at the September, 2014, Meeting of the Midwest Political Science Association in Chicago, IL, USA, and the 2014 Comparative Agendas Project conference in Konstanz, Germany. We are grateful for comments from Will Jennings of Southampton University and Ian Budge of the University of Essex. We would like to thank the editors and our anonymous reviewers for their insightful comments and support.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on contributors
Shaun Bevan is Lecturer in Quantitative Political Science at the University of Edinburgh.
Zachary Greene is a Chancellor’s Fellow and Lecturer in Political Science at the University of Strathclyde.
Notes
1 We refer to the policy agenda as the distribution of issues addressed in parties’ manifestos or government policy (Bevan et al. Citation2011). We refer to agenda-setting as the broad process of constraining which issues are attended to by government. This usage contrasts the more constraining definition used in studies of legislative behavior, which refers to the agenda as the structure and order of voting used to manipulate voting outcomes (Döring Citation2001).
2 We define issue attention as the distribution of policies passed in a year across issues.
3 Parties fear retribution for a weak economy. Governments appearing unconcerned with economic decline likely face even greater punishment (Green and Jennings Citation2012a, Citation2012b). A weak economy, therefore, leads to greater agenda stability because the government focuses policy attention on topics it believes improve the economy.
4 This perspective complements a veto-players approach. While coalitions may contain multiple partisan agenda-setters (Tsebelis Citation2002), more parties increase the likelihood of ideologically ‘captured’ parties, but not the total number of effective partisan veto players. More parties, however, increase the number of demands across issues and allow the prime minister multiple routes to achieve policy.
5 *CAP: www.comprativeagendas.net *Canada: www.snsoroka.com/data.html.
6 A pooled model (not presented here) only including the two longest series, the UK and Denmark, produces the same inferences.
7 Our interest in agenda stability allows us to be neutral about the types of issues parties address to look for systematic change in the aggregate agenda. However, this work is related to research on diversity of that agenda and the tradeoff between issues using Shannon’s H (Jennings et al. Citation2011; Boydstun et al. Citation2014; Greene Citation2016a). Nevertheless, the measures only weakly correlate (0.39).
8 As a robustness check the analyses were also run using a fractional logit model to account for agenda stability’s bounding between 0 and 100. These results produced the same inferences.
9 We use Shannon’s H to construct ENCP. Our results are similar using a measure based on the Herfindahl index, although predictably it does not perform as well as the Shannon’s H. The coefficients are in the same direction, although they barely drop below standard levels of significance for the Herfindahl. The overall model fit decreases as indicated by the adjusted R squared.
10 In rare cases with multiple transitions in the same year, we use the value for the longest serving cabinet in that year.
11 Analyses of the effect of both pre- and post-election dummy variables as possible alternative measures led to poorer fitting models.
12 We hold each of the primary independent variables at their mean values in , varying GDP growth from its minimum to its maximum observed values for observations following transitions. Ninety-five per cent Confidence Intervals are simulated from 1000 draws of the variance covariance matrix based on Model 3. The dashes across the bottom show the distribution of observations. This same approach is used for and .
13 Wald tests of the coefficients’ joint significance indicates that the effect of coalition size is statistically different from zero at the 95 per cent level during transitions and 99 per cent level during non-transition years.
14 Models excluding the single-party government variable with a parabolic ENCP variable lead to similar results. Both single-party governments and coalitions with more parties have decreased stability relative to coalitions with only a small number of parties.