Economic conditions and forecasting of Australian elections revisited

In a recent paper in this journal, Jackman and Marks (1994) examine the relationship between economic conditions and election outcomes in Australia using data from 1949 onwards. Their analysis, however, is in my opinion technically inadequate because they seem to have overlooked the possible existence of autocorrelation in their models. In this paper, I present a re‐examination of some of their models which takes autocorrelation into account and which results in improvements in both the fit of the models and of their overall predictive accuracy. Inflation is shown to be a more definite influence on vote shares than are changes in unemployment rates. Also, I show that Jackman and Marks’ conclusion that the 1993 election was an exceptional one in the way in which economic conditions affected aggregate vote shares is incorrect, and that the election result was actually quite well within the normal range in the extent to which it was accurately predicted by economic conditions.     

Three Tools for Forecasting Federal Elections: Lessons from 2001

What is the best way to predict Australian federal election results? This article analyses three forecasting tools: opinion polls, economic models, and betting odds. Historically, we find that opinion polls taken close to the election are quite accurate, while economic models provide better medium-run forecasts. The November 2001 federal election largely follows this pattern, although the economic models provided more accurate projections than recorded through the 1990s. Against these, we compare betting odds, analysing a rich data source from one of Australia's largest bookmakers, Centrebet. The betting market not only correctly forecast the election outcome, but also provided very precise estimates of outcomes in a host of individual electorates. Betting fluctuations present an intriguing quantitative record of the shifting fortunes of the campaign. Particularly in marginal seats, the press may have better served its readers by reporting betting odds than by conducting polls. We conclude that the results of these three models can help determine how important the events of August and September 2001 were in deciding the outcome of the election.     

National Uniformity,and State and Local Effects on Australian Voting: A Multilevel Approach

The existence and extent of influences arising within spatial contexts is an important issue in the study of voting behaviour. This paper extends previous Australian research by using the relatively new technique of multilevel analysis to draw together individual survey data from the 1993 Australian Election Study and ecological census data to investigate the question. The results show that, once individual voter characteristics are taken into account, influences on first preference voting for the ALP at the 1993 election were quite uniform nationally, with relatively small spatial variations. Moreover, those spatial variations which were present were at the divisional, not the state, level and can be almost completely explained by a very small number of sociotropic factors, especially a local economic prosperity influence and the well-known rural-urban cleavage. As far as influences on voting at the 1993 election at the level of individual voters are concerned, these multilevel analyses provide some new insights, as well as confirming some previous results.     

Spatial Variations,Contextual and Social Structural Influences on Voting for the ALP at the 1996 Federal Election: Conclusions from Multilevel Analyses

Multilevel modelling of individual survey data from the 1996 Australian Election Study and aggregate contextual Census data is used to investigate the extent of spatial variations in voting at the 1996 election, and to examine potential explanations for the variations. The size of the variations was about the same as at the 1993 election, and was again mainly at the divisional rather than the state level. However, unlike in 1993, only a small amount of the variations could be explained by differences in the individual sociodemographic compositions of the areas, suggesting that local influences were more important in 1996 than 1993. The most significant change in the effect of the individual-level variables between the two elections was that the influence of being personally unemployed completely changed direction (working against the ALP in 1996, whereas it had favoured it in 1993), although the estimated strengths of several others also changed. I suggest that the concentration of the 1993 election on the 'Fightback' package might be the main cause of many of these differences. The nature of the main contextual influences is similar to those at the 1993 election, involving an urban-rural effect, a local economic effect and an ethnic effect, confirming that these are more than short-term factors. However, there also remain somewhat larger unexplained local influences than in 1993. Detailed investigation of the ethnic contextual effect suggests the presence of a specifically anti-Asian influence.     

Spatially Disaggregated Modelling of Voting Outcomes and Socio‐Economic Characteristics at the 2001 Australian Federal Election

This study uses GIS and spatial modelling to relate voting outcomes at the 2001 federal election for polling booths across Australia with the socio‐economic characteristics of polling booth catchment areas. The data and analysis used are more detailed and comprehensive than previous studies. It is conducted at a fine level of spatial disaggregation across the whole nation to examine voting outcomes for both major and minor political parties. Because the aim of the paper is to distinguish voting outcomes between political parties rather than to predict voting outcomes for particular political parties, a discriminant analysis is used rather than regression analysis. The statistical discriminant analysis identifies two main socio‐economic dimensions that are able to predict polling booth outcomes with a relatively high degree of accuracy. That analysis shows how, at the 2001 federal election, the middle ground, in terms of socio‐economic characteristics, was being claimed by the Liberal Party, Country Liberal Party, The Greens, and, to a lesser extent, by the Australian Labor Party. However, the Australian Democrats, National Party and One Nation had more distinctive constituencies, with the National Party and One Nation Party competing for areas with similar socio‐economic characteristics. Using GIS mapping tools, examples of actual and predicted polling booth voting outcomes are given, along with selected socio‐economic characteristics of booth catchments.     

Interest rates and the 2004 Australian Election

During the October 2004 Australian federal election campaign the expected or possible effect of the election outcome on interest rates was a key point of differentiation between the Australian Labor Party and the Liberal–National Party coalition. The purpose of this paper is twofold. First, we examine whether this effect was a significant factor in the election outcome, as measured by the percentage swing towards the coalition in each electorate. Second, we use standard methodology from financial economics to examine whether the election outcome had an effect on interest rates. Contrary to media coverage of the campaign, we find that the election result did have an effect on interest rates but that the possibility of interest rate changes was not a dominant factor in the election result.     

Internal migration and elector turnover in Australia

This paper argues that internal migration should be given more attention in Australia as a factor to be taken into account in studying voting patterns. After briefly outlining some facts about the recent extent and nature of Australian internal migration, the paper describes and analyses some new information about the amount of elector turnover in the 14 Western Australian Commonwealth electoral divisions between the 1990 and 1993 elections. Even in divisions which had little change in overall enrolment levels, the analyses show such a high degree of turnover that, in almost all of the divisions, between a quarter and a third of the 1993 electors had not been enrolled in the same division at the 1990 election. The uses and limitations of this type of data are also discussed.     

Concentration Effects,Spatial Mismatch,or Neighborhood Selection? Exploring Labor Market and Neighborhood Variations in Male Unemployment Risk Using Census Microdata from Great Britain

A classification of census enumeration districts has recently been added to the 1991 British Census 2% Sample of Anonymised Records (SAR). Through the use of multilevel modeling techniques, the area classification information on the SAR is used to investigate geographical differences in unemployment. Previous research has indicated that where a person lives can affect their propensity to unemployment. However, the understanding of these relationships are confounded by the reciprocal nature of the relationship between unemployment, housing, and geographical location. This paper examines the relative importance of the individual, the type of neighborhood of residence, and the local labor market in which one lives in explaining variations in unemployment risk. It also examines the role of housing tenure at an individual and contextual level in mediating this relationship. Two competing hypotheses are evaluated. The first is that local concentrations of unemployment are the result of the process of neighborhood selection. The other suggests there are contextual affects on unemployment risk, which may include access to job opportunities and “concentration effects.” The paper concludes that most neighborhood level variation in unemployment is due to housing market effects, particularly neighborhood selection. As well as offering insights into the relationship between unemployment and geographical location, this paper aims to demonstrate methodological innovations in the analysis of census microdata. In particular it shows how area classifications can be used in conjunction with microdata in a multilevel modeling framework, to get a better understanding of the role of individual and contextual factors in social processes.     

Postmaterialist and Economic Voting in Australia, 1990-98

Recent research in Australian sociology and political science has debated the extent to which postmaterialist values and economic self-interest shape voting in federal elections. Some researchers have argued that postmaterialist values have partly displaced materialist concerns with physical security and economic well-being in Australian public life. This displacement, coupled with the adoption by major political parties of postmaterialist 'quality of life' issues such as the environment, has meant that voting in Australia has come to be more dependent on postmaterialist values than on perceptions of economic interest. Other research, however, has found no relationship between postmaterialist values and voting behaviour, while economic evaluations remain a strong determinant of voting behaviour. Part of the disagreement reflects methodological differences in the research. But different methodological problems compromise each of the previous studies. In this paper we use data from the 1990, 1993, 1996 and 1998 Australian Election Studies to investigate postmaterialist and economic voting in the Commonwealth House of Representatives and the Senate. Using various statistical methods, we first explore bivariate relationships between key variables and then use multivariate models of postmaterialist and economic voting to adjudicate between the contending positions.     

The problem with counting preferences: the example of one nation party preferences at the 2001 Western Australian legislative assembly election

A growing number of lower‐house seats at Australian State and federal elections rely on a distribution of preferences from Independent and small‐party candidates before seats can be awarded. Actors have attempted to gain political capital from this situation by claiming that the preferences of particular small parties have affected election outcomes. This paper uses the events of the Western Australian State election held in 2001 to explore the validity of such claims. More specifically, it investigates the widely propagated contention that the One Nation Party's anti‐sitting‐member preference strategy was a key determinant of the Coalition's electoral defeat. It concludes that the increased number of candidates contesting elections makes it difficult to assess whether the second and subsequent preferences of any particular small party were critical to the outcome.     

Party commitment,vote switching and liberal decline in Australia∗

Electoral outcomes are determined in part by voters who switch their political allegiance between elections. Though the subject of extensive study in other countries, this floating vote has received comparatively little attention in Australia. This paper uses 1967 and 1979 survey data to analyse vote switching in Australian elections. We argue, firstly, that turnover tables based on recalled voting are, as research in other countries has shown, inaccurate reflections of the election result. As an alternative, we propose a measure of ‘potential’ vote switching based on the concept of lifetime voting. Secondly, using this concept of lifetime voting to derive a measure of party commitment allows the construction of a three‐fold typology of Australian voters — partisans, marginals and switchers. Finally, a longitudinal analysis of the voting patterns of these three groups indicates a gradual erosion of electoral commitment to the Liberal party, whereas Labor has largely retained its partisan base.     

Ballots and billions: Clive Palmer's personal party

Founded and led by the billionaire businessman, Clive Palmer, the Palmer United Party (PUP) achieved what was arguably the best debut result of recent decades at the 2013 Australian federal election. This article examines PUP's ideology, organisation and campaigning strategies along with the implications of its experiences to date for Australian party politics. Based on an analysis of original party documents, policies, media communications and semi-structured interviews with PUP candidates and key figures across Australia, we find evidence of a party which is utterly dominated by its leader, which was deliberately never built to last, whose ideology cannot be easily classified and whose campaigning was well-funded but extremely disorganised. We conclude that its experience shows how more professionalised new personal parties in Australia should be able to do even better electorally in the future.     

Are Spells of Unemployment Longer in Nonmetropolitan Areas? Nonparametric and Semiparametric Evidence

I use nonparametric and semiparametric proportional hazard models to examine whether individuals resident in nonmetropolitan areas experience lower per period rates of exit from unemployment following job loss than metropolitan area residents. Results show that between 1989 and 1993 per period cumulative rates of exit from unemployment were slightly higher in nonmetropolitan areas, mainly due to nonmetropolitan–metropolitan differences in individual characteristics and local economic conditions. Employment density is found to have a positive association with rates of exit of unemployment in metropolitan areas but not in nonmetropolitan areas.     

Some approaches to election night forecasting in Australia

We outline the history of election night forecasting in Australia. We first identify the major structural features of the Australian voting system which determine how the problem must be approached: single‐member constituencies; preferential voting; the counting on election night, in polling booths, of first preference votes only: and the existence of a two‐party system. We note that the major statistical difficulty associated with election night forecasting in Australia is that progressive counts in a seat, far from being random samples of the votes cast, are non‐random and systematically biased samples. We finally outline the increasingly sophisticated computer models which have been used to correct for this bias, culminating in the ‘matched polling places’ model implemented by the Australian Electoral Commission at the 1990 election, which has effectively eliminated the problem of bias.     

Australia–Japan economic relations: ten years after the NARA Treaty

The NARA Treaty looked to a strengthened Australia–Japan relationship, which foreign ministers of both countries celebrated this year. Events of the mid 1980s, however, demonstrated how an adverse turn in economic circumstances can stress even a strong bilateral relationship. The two countries were then at opposite points in their economic cycles, Japan almost too successful, with the high yen putting pressure on Japan's export industries, steel in particular. Seeking to reduce costs, the steel mills transferred pressure to their Australian suppliers of coal and iron ore, through reductions in prices and volumes. This was unwelcome to an Australian government coping with a recession and worried about balance of payments and unemployment, and led to concerns about equal treatment. Although the minerals trade returned to normal relatively quickly, ambitious investment projects aimed at enhancing the economic relationship did not materialise.     

A Right-of-Centre Triumph: The 2004 Australian Half-Senate Election

The Australian Senate is a significant part of the Australian parliamentary system and the electoral contest for the Senate results in representational outcomes that will influence the way the Senate will perform. This paper argues that the 2004 half-Senate election result was significant because it resulted in the Liberal–National Coalition obtaining a majority in the upper house. It accounts for this outcome by examining the contest by way of inter- and intra-party bloc contests. It finds that a particularly strong Right-of-Centre performance in Queensland, to which voters voting for Ms Pauline Hanson made a major contribution, delivered the Senate majority to the Howard government. The significance of the result also lies in the way it confounded previously held views that the combination of proportional representation used for the Senate with the need to elect six senators from each State would make it unlikely for either Labor or the Liberal and National Parties to ever win an upper house majority in the future.     

Children and families: Relationships between family attitudes,household composition and voting in Australia

In the light of ongoing demographic trends (such as low fertility rates and growth of single-person households), some of the features of the 2004 Australian federal election outcomes and campaign raise the possibility that we might have seen the beginnings of a divide in voting behaviour based on family structure, particularly those aspects related to the presence of children. Relevant data from the 2004 Australian Election Study (AES) are quite limited, so I use data from both the 2004 AES and the 2003 Australian Survey of Social Attitudes to explore the relationships between federal voting and family structure, and attitudes towards children, parenting and families of different types. The results show effects on voting that apparently result from financial considerations related to the presence of children and others that are related to differences of values.     

Policies and performance in the 2019 Australian federal election

ABSTRACT

The Liberal–National Coalition success in the 2019 Australian federal election surprised many observers, with the opinion polls consistently predicting a Labor victory. The election was notable for Labor’s proposals for wide-ranging tax changes, a historically unpopular Labor leader, and the fourth change of prime minister outside an election since 2010. Using the 2019 Australian Election Study, we test the influence of two models of voting, the first based on ideology together with an emphasis on policy, and the second on performance. The results show that performance was the dominant explanation for the result, with evaluations of party competence and leader popularity playing a major role in explaining voting behaviour in the election, both of which benefitted the Coalition.     

The South Australian Election of 18 March 2006

The South Australian election on 18 March 2006 saw the Labor Party; led by Mike Rann, storm out from the shadows of minority government to record a decisive victory which has the potential to set a platform from which the party could govern for at least the next eight years. Labor won 28 of the 47 seats in the House of Assembly gaining 45.2% of first preference votes. However, the decisiveness of its victory in the Lower House was not matched by the result in the Legislative Council. Voters turned away from both major parties giving 4 of the 11 seats contested to minor parties and independents. This continued a trend that has been evident in South Australian elections for the last two decades, although on this occasion its dramatic acceleration was explained by the extraordinary result of the ‘No Pokies’ independent Nick Xenophon who won 23.5% of the vote and easily secured two quotas. The election campaign itself marked new heights in the trend towards presidential style campaigning on the part of the Labor Party which used extensive television advertising to make the most of its popular leader. The Liberals, on the other hand, lacked both a leader to counter Rann and the funds to match Labor's advertising budget. The election also saw the confirmation of Family First as a significant player in South Australian politics, and may mark the beginning of the end of the Australian Democrats. The Democrats failed to make any impact and were effectively replaced by the Greens, who in winning a seat in the Legislative Council enjoyed their first South Australian electoral success.     

Persuasion,Partisanship and Evaluations of National Conditions and Governmental Impact: An Experimental Study

The current study considers the circumstances under which opinions about national issue conditions and government impact on those conditions might be influenced. Using an experimental framework, we examine the effectiveness of messages on a variety of policy issues important to the 2015 Canadian Federal election campaign. The messages consist of a statement about the country’s performance on a variety of issues, accompanied by a relevant image, which together convey either a positive or negative message about issue conditions in Canada. We then evaluate the extent to which messages affect opinions about policy conditions in Canada as well as evaluations of government impact on these policy issues. Finally, we consider whether the effect of the messages is moderated by the partisan leanings of individual voters. We focus upon five policy areas that, we argue, were among the most germane to the 2015 election: the economy, health care, the environment, immigrants and minorities, and national security. Our findings have implications for the literature on issue-ownership, agenda-setting, opinion persuasion, and the attribution of responsibility for national performance.