The logic of policy-motivated parties

Representative democracy is party democracy. Parties helped to create and continue to shape our democratic institutions, dominating the formation of governments and (most) executive offices around the world. The policies they adopt when in office have significant ramifications, influencing the distribution of resources and laws that regulate behaviour across society. We care about how parties operate, their institutional structures, what their motivations may be, and the implications –if any– if one party takes office rather than another.  


Source: Alex Ellinghausen, The Sydney Morning Herald

An increasingly popular view in both the political science literature and public commentary argues parties are run by office-seekers, whose goal is to win elections; or that they have just become irrelevant to the policy development process.

There are two main ways these claims are made. The first is a supply-side explanation of competition-induced convergence: that party candidates are in a large part motivated by the private benefits of public office (the Downsian model of party competition); including the income, resources and influence that comes with government. The second, demand-side explanation, is one of mostly involuntary constraints on the parties’ abilities to effect policy change. An example of this is the (supposed) decline of economic cleavages. As the social cleavages that underpinned political competition during the formation of the Australian party system declined, so too did the demand for significant economic policy difference.

These claims presume parties are made up of office-seekers for whom interest groups and constituencies are expendable (the interests of the middle class and owners of capital were only of concern to the Coalition parties, so long as they helped them obtain and hold office) or have declined in relevance over time. This argument also depends on voters not holding parties accountable for their previous policy positions. This enables parties to moderate in a mercenary manner for political gain.

These assumptions are not always met in practice (see Groffman 2004), and despite potential electoral benefits offered by policy moderation (Canes-Wrone et al. 2005; Ezrow 2005), politicians, parties and policy outcomes regularly deviate from the median voter’s preferences (Poole and Rosenthal 2000; Scervini 2012).

So why does this happen?

One reason is that important economic differences are still significant influences in the electorate, creating a centrifugal force that drives the parties away from the centre on some issues.

The median voter theorem and similar spatial models all assume parties will moderate to capture the private benefits of public office. However, the actors that comprise the parties — including candidates, activists, donors and voters — do not gain the same private benefits from their party winning office, if they receive any at all. Instead, it is the public benefits they can obtain from supporting a party that aggregates their policy preferences.

Rather than purely opportunistic office-seekers, we can view Australia’s major parties as interest aggregators representing this diverse electoral alliances united by these shared policy preferences. This includes economic policy.

We have many reasons to believe important economic interests continue to exist and are represented by the Coalition and Labor Party. Despite claims that the significance of economic cleavages have declined, social mobility may be lower than was previously thought (Chetty et al. 2014; Clark 2014), and growing inequality has been attributed to policy decisions (Atkinson and Leigh 2013).

Shared interests on labour laws and public spending can unite lower income employees and public servants to support the Labor Party, and those with higher incomes, and the owners and managers of capital, around the Liberal-National Coalition. Rather than electoral success being an outcome sought for itself, winning office is the means to gain these desired public goods.

These differences are represented in the interests of donors, parliamentarians and voters of the parties. Despite the growing number of career politicians, the parties continue to attract parliamentary representatives with distinct occupational backgrounds (Miragliotta and Errington 2012). Sources of financial contributions match these observations, with the Coalition receiving far more financial support from corporate interests than Labor (McMenamin 2008). Voters with higher incomes, own assets and hold managerial occupations, are on average more supportive of the Coalition. This is not the case for Labor. Here it is employees in blue-collar jobs and with lower incomes were more likely to vote.

Figure 1 shows there has been very little change in these patterns over 35 years. As Figure 2 shows, there are also significant differences in the economic preferences of Labor and Coalition voters and their candidates for federal parliament. Yet, this pattern isn’t as strong for social issues, which are driven by concerns over immigration. Again, there was little change in these patterns over multiple decades.


Figure 1. Trends in the partisan support of high- and low-income voters by key occupations, 1979– 2013. Curves are the patterns in predicted two-party vote choice estimated from the multilevel model fit to Australian National Political Attitudes Survey and Australian Election Study (AES) data documented in the supplementary material.

SR fig 2

Figure 2. Patterns in the issue preferences of voters and candidates, 1987–2010. Each party is measured by the difference between the preferences of their median voter and candidate, and the electorate’s median voter. Preferences were estimated using the IRT model fit to respondents’ answers to the issue questions included in the AES and Australian Candidate Study (ACS) data. The solid lines represent voter preferences and the dashed lines estimated candidate preferences

If the parties are asking citizen for their vote, donors for their money and activists for their time, these actors that comprise the parties’ electoral alliances require something in return. The theoretical framework outlined above suggests party actors comprising the Coalition – activists and candidates motivated to reduce the size of the welfare state, and the corporations who donate money for a more business-friendly government – have no incentive to win an election by moving to the left of Labor on policy. For voters, activists and donors to continue to support the Coalition if it did move to the left would result in a government that provides them with poorer results. Even limited moderation on important issues potentially reducing the value of effective support by activists and donors.

These assumptions create a limited set of possible policy movements available to the parties on key issues. This ensures that despite possible electoral pressure to moderate, parties will retain distinctive policy platforms to obtain utility from election wins. In this framework, the Coalition will generally be to the right on highly salient issues. To increase its chances of winning the election, it may sacrifice its preference on an issue that is relatively unimportant to its policy-motivated factions, but that is important to the electorate. However, it will stop at trading off policy goals integral to its electoral alliance.

My research supports this theoretical framework. Examining data from the Manifesto Project covering federal elections from 1946 to 2013, shown in figure 3, I find the Coalition never adopted policy position to the left of Labor and in general they have been significantly to its right (averaged across all the issue domains measured by the Manifesto Project).

SR_Fig 3

Figure 3. Left–right position of party policy promises during Australian federal elections, 1946–2013. Each point represents the left–right score of Coalition and Labor election promises (from –100 to the left to +100 to the right) for each Australian federal election. Curves represent the trends of party positions over this time, smoothed using a LOESS regression, with the dashed blue curve the Coalition and the dotted red curve Labor.

I find similar results after an examination of a sample of 147 legislative votes (divisions) held in the Australian parliament from 2006 to 2015, sorted into three issue domains. Shown in Figure 4, these divisions indicate that the parties have behaved very differently in parliament in recent years. This is particularly so on issues concerning the influence of trade unions in the workplace. In both chambers, Coalition parliamentarians voted to reduce the power of trade unions nearly 100 per cent of the time. By contrast Labor parliamentarians voted against without exception.

Although the gap between the parties was smaller on taxation and social spending, a consistent difference remained, reflective of the interests the parties represent. With the electoral alliances of both parties shown above to be less clearly differentiated on social issues (driven by matters concerning immigration and asylum seekers, as shown in figure 2), the parties were less likely to take different stances on these issues in parliament. This is exactly what we observe.


Figure 4. Parliamentary votes by party, chamber and issue. Each closed circle represents the aggregate support for legislation in each of the three policy domains. As these represent samples of divisions taken from the whole possible universe of votes, 95 per cent confidence intervals are represented by the horizontal bars to indicate uncertainty in these estimates. These data were scraped from the database hosted by the They Vote for You website.

These findings suggest that viewing parties as primarily or solely office-motivated deserves a second thought, and an interest-based model of Australian parties needs updating, not abandoning. The major difference between parties characterised in this way, and those viewed as convergent types, is that interest-aggregating parties are policy-motivated and will sacrifice a certain amount of electoral success, rather than give up key policy goals to win office.

This is not to suggest that the parties provide alternative policy offerings on all issues and at all points in time. Additionally, I am not arguing that the parties have remained completely unchanged. Australian society is not the same in 2016 as it was 50 years ago. However, the results do indicate there are important constants in the behaviour of the major political parties based around the interests they were established to represent. Understanding the policy outcomes parties pursue and the explanations for these goals can allow political scientists and commentators to better interpret the preferences and behaviours of these parties and the actors that comprise them.

Still interested? You can read the full version of this article ‘Interest aggregators, not office chasers: evidence for party convergence and divergence in Australia’ which was published in the Australian Journal of Political Science

Shaun Ratcliff United States Studies Centre, University of Sydney

Information on data and methods can be found in Ratcliff (2017), Interest aggregators, not office chasers: evidence for party convergence and divergence in Australia, Australian Journal of Political Science, 52:2, 236-256, and its supplementary materials, available here.