< - Contents < - Preliminary Pages  : Conclusion - >

CHAPTER FOUR: Difference or Convergence? A Longer Term Analysis of the Parties Ideological Directions and Relationships

Are the convergences indicated in the previous 2004 election case-study part of a longer-term trend? There is a considerable body of literature arguing that the basic differences of ideology or principle that once existed between the ALP and the Liberal Party have receded. However, as Murray Goot notes in a recent review of the literature, there are differing views on when this argued convergence started to take place.⁽¹⁾ Many commentators identify the 1980s, during the Hawke Government, as the crucial period. It has been claimed that during that time the Labor Party became more like the Liberal Party in accepting a free market and the importance of a private sector, and the Liberals more like Labor in accepting significant levels of state welfare, and limited state intervention in the market.⁽²⁾ One commentator has remarked that, in the name of competition and greater efficiency, the Hawke and Keating Governments deregulated financial markets, presided over the end of centralised wage-fixing (and through the accord, a decline in real wages), sold revenue earning public assets, and abolished the alleged guarantor of equality of opportunity, free tertiary education.⁽³⁾ Other analysts and commentators, however, argue that the parties convergence manifested much earlier than the 1980s. Goot notes Hugh Mackay as declaring that from the post-Whitlam years, the two party system had lost its way, or, perhaps, lost its point .⁽⁴⁾ Brian Head and Allan Patience went even further, to suggest that policy directions in both parties had been converging since the 1940s.⁽⁵⁾

More recently, there has also been a perception among surveyed voters that the two major parties had become increasingly similar between 1993 and 2001, as reflected in Chart 8 below. Interestingly, this perception changed for the 2004 election, where there was a notable increase in the proportion of surveyed voters who saw the two parties as showing a good deal of difference, and a notable decline in those who saw little or no difference.

Chart 8: Percentage of Surveyed Voters and Perceptions of Party Difference (1993 2004)

Source: Australian Election Studies 1993 2004

Measurement of Party Difference

While not denying the importance of these commentator insights and analyses, nor the reality of perceived party convergence, there is still a lot that all of this does not tell us. Are those insights and public perceptions accurate? Without some systematic criteria or scale for comparison, and for answering the questions the same or different in exactly what respect? , and how much the same, how much different? , the comparatives same and different can begin to lose their meaning in this context. Moreover, if the two parties have converged in some way, is this convergence due to one party converging on the other, both converging on each other, or both converging on some further ideological focal point that of the typical voter, perhaps. Each of these is importantly different, and it would be useful to have some way of measuring ideological positions and movements that could throw light on these.

The question now is what the best systematic measures for this purpose might be? Ideally, the measures of party difference to be preferred would be ones that (i) are as precise, quantified and fine-grained as possible; (ii) are repeatable, and depend on the subjective judgements of the measurers, as little as possible; (iii) allow for systematic comparisons over time between parties, (iii) measure the ideological positions of parties when they are, as much as possible, subject to similar background influences, and (iv) identify when parties are of the left or of the right as validly and accurately as possible. A number of candidate measures of party ideological difference might fit these desiderata to varying extents. For example, one such candidate measure examines the fiscal behaviour of parties when they are in government the relative amounts they spend on different policy areas, and the means by which they raise revenue. There is plausibility to the view that there are ideologically based differences in this behaviour, with the ideologically left being disposed to greater state-intervention, bigger government and greater expenditure (e.g. in welfare areas); and the ideological right to smaller government, less expenditure and less reliance on taxation for revenue.⁽⁶⁾ Here, there is a reliable (i.e. repeatable) quantifiable measure (outlays and revenue) that allows for inter-party comparisons, is plausibly related to the general understanding of leftness and rightness, and allows to some extent for comparisons over time.

While an exhaustive analysis of party convergence and difference would take account of the outcomes of a range of best measures (i.e. provide a meta-analysis), the discussion to follow here will make use of a measure of ideological position that is particularly powerful, and rates very highly on the above desiderata (though, of course, not perfectly). This measure will, again, rely on the textual analysis of the parties election policy launch speeches. This time, however, the analysis will have a quantitative and statistical orientation, both in its methodological basis and in its application. Given the weight that will be placed on this manifesto-based measurement of ideological position, the next section provides a detailed account of its methodological basis, rationale and defence. After that, the measure will be applied to compare and systematically map over time, the relative ideological positions, and their various changes, for the Australian Labor Party and the Liberal Party of Australia.

Manifesto-based Analysis of Left-Right Ideological Position Rationale and Methodology

The approach to be applied here to measuring the ideological positions and distances of political parties is an application of the methods employed by the Manifesto Research Group (MRG) of the European Consortium of Political Research.⁽⁷⁾ This group has conducted systematic content analyses of the election manifestos of significant political parties competing in elections in twenty-five democracies over a period of more than fifty years (1945 1998).⁽⁸⁾ The manifesto documents that were analysed ranged from party election platforms to party leaders election policy speeches. The group s analyses encompassed democracies from Australasia, North America and Western Europe, including party-systems that had often been overlooked in other comparative studies (e.g. Japan and Israel). The specific countries investigated were Australia, Austria, Belgium, Canada, Denmark, France, Finland, Germany, Greece, Iceland, Ireland, Israel, Italy, Japan, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States. These countries (and the considerable period over which they were examined) reflect a significant degree of cultural, socio-economic and institutional variation.⁽⁹⁾

Analysis of an early sample of manifestos identified distinct sets of policy references or categories of policy emphasis in the texts which were systematically related to whether the political party referring to them was of the ideological left or right.⁽¹⁰⁾ Those categories were applied by the MRG to define a Left-Right Scale. Because of the statistical correlation between the categories and a party being of the left or of the right, it was argued by the MRG that the degree to which an election manifesto or policy speech places emphasis on these categories can be taken as a measure of the degree to which the manifesto (and the party presenting it) is ideologically of the left or of the right. Degrees of emphasis were measured by the MRG in terms of the frequency with which the categories were mentioned in the text, where mentions were counted as references in sentences or meaningful quasi-sentence clauses. The categories that form the MRG s Left-Right Scale are as follows⁽¹¹⁾:

MRG Left-Right Scale

Left	Right
Military, negative	Military, positive
Peace	Freedom
Internationalism, positive	Constitutionalism
Decolonisation	Government Effectiveness and Authority
Nationalisation	Free Enterprise
Regulation of Capitalism	Economic Incentives
Economic Planning	Economic Orthodoxy and Efficiency
Protectionism, positive	Protectionism, negative
Social Welfare Expansion, positive	Social Welfare Expansion, negative
Democracy	National Way of Life
Controlled Economy	Traditional Morality
Education Expansion, positive	Law and Order
Labour Groups, positive	Social Harmony

On the basis of this scale, the MRG plotted the left-right position of a party s election manifesto by determining the frequency (percentage) of all statements in the text that are left category statements and the frequency of all statements that are right category statements, and subtracting the former from the latter. This was taken as giving a net ideological score on a scale between -100 and +100 (left to right), with 0 as centre or left-right neutral. The MRG s simple subtractive computation can be represented as follows:

ID Party = (%IDRight %IDLeft)

So, a policy speech that makes a total of 200 statements, with 40 (20 percent) mentioning Right categories, and 100 (or 50 percent) mentioning Left categories, will receive an ideological score of -30 (i.e. 20 minus 50), which puts it on the ideological left. Such scores can be used to spatially plot the ideological positions of parties in graphs.

Within the MRG s left-right scale, a distinction can be made between categories of policy emphasis that are within the economic domain, and categories within a broadly social or non-economic domain (keeping in mind, of course, that there is not always a black and white distinction). This separation between left and right economic and non-economic indicators allows for an even more fine-grained mapping of parties ideological positions. The left-right economic categories identified by the MRG, and used as a left-right scale for economic policy, are as follows:

MRG Economic Policy Left-Right Scale

Left	Right
Nationalisation	Free Enterprise
Regulated Capitalism	Economic Incentives
Economic Planning	Economic Orthodoxy and Efficiency
Protectionism, positive	Protectionism, negative
Controlled Economy

There will be differences of view as to what should be counted as social policy. The MRG, for example, makes a distinction between welfare/quality of life policy categories, and categories to do with external relations, freedom and democracy, the fabric of society and the nature of government. Rather than entering a debate over what precisely counts as social policy, this monograph takes the remaining non-economic indicators on the MRG s left-right scale as defining a left right scale for non-economic policy. Scores can be calculated for the economic and non-economic leftness and rightness of a party in terms of these scales, and those scores plotted graphically. In this monograph, the economic and non-economic ideological positions of parties are measured as follows:

EcoID Party = (%EcoIDRight %EcoIDLeft)

Non-EcoID Party = (%NonEcoIDRight %NonEcoIDLeft)

The terms %EcoIDRight and %EcoIDLeft refer respectively to the percentage of economic policy references in the text that are within the ideologically right economic indicator categories and the percentage within the left economic categories, specified above. %NonEcoIDRight and %NonEcoIDLeft refer respectively to the percentage of other than economic policy references in the text that are within the ideologically right non-economic indicator categories and the percentage within the left non-economic categories.

One of the most obvious criticisms of this whole approach to mapping parties ideological positions is its reliance on what parties say, not what they do, and especially what they say at election time when their overriding interest is winning elections. It is true that what parties say and what they do after gaining office are not always the same, and that election statements are designed to win votes. Both of these facts might call into question just how accurately manifestos and policy speeches reflect parties real degree of leftness or rightness. After all, it could be argued that the real ideological leaning of a party is reflected in what it actually does.

While these concerns are genuine ones, they are not as significant as they might first seem. What a party does while in government, or what it says it will do while in opposition between elections, is still subject to compromise and competing interests, often electoral ones. The assumption that election policy statements are infected by pragmatic electoral interests, but that real policy actions are not, is false. Moreover, even if there are differences between election statements and subsequent actions, it is unclear how consistent and ubiquitous those differences are, and unclear how often those differences involve ideological shifts. The perception that governments or parties do not keep their promises can often be anecdotally based. But when it comes to systematic evidence for inconsistencies between platforms and practise, the jury is still out. Some studies suggest a generally good correspondence between these,⁽¹²⁾ while others do not.⁽¹³⁾

So, while there will be possible drawbacks of focusing on policy speeches, they are less substantial than they appear. There are also some definite advantages in adopting this focus for the purposes of comparative measurement of party ideology over time. While the average voter is unlikely to read policy statements at election time, those statements are nonetheless a focal point for political argument and debate during an election campaign.⁽¹⁴⁾ They stand as the only clear statement of party policy available to the electorate and directly attributable to the party.⁽¹⁵⁾ Usually, statements of election policy have the authority of the leadership, or are products of debate within the party, and to that extent they have a considerable claim to being authoritative indicators of what parties present themselves as standing for.⁽¹⁶⁾ Election policy speeches are also specifically dated sources of data, which makes them particularly useful for temporal comparisons. They also have a further important advantage. When defining or measuring the ideological position of political parties in competition, it is important to be able to measure this as much as possible in abstraction from other contextual influences that may differentially affect the policy behaviour of parties. Parties will behave differently in government than when in opposition, for example. Parties in government compete with other parties from a position of incumbency. An election context, while it does not eliminate all background differences in party competition, at least provides a similar context for competition: no party is in full power, and all are competing for election to government and present their policy and platforms in that same circumstance.

All of these considerations, concerns and counter-arguments ultimately go to the question of whether the left-right scale developed by the MRG validly measures ideological position. As was said, there is a significant statistical correlation among the policy categories on each side of that scale they do seem to group together, which suggests that they indicate some underlying factor. But are the underlying factors really ideological leftness and ideological rightness? Are there any ways of checking the validity of the claim that they are? One immediate check is to see if parties that have been identified via the MRG scales as of the left or right are consistently of that ideology over time. If the ideological positions of parties as measured by the MRG scales were to constantly leap-frog each other and constantly cross between left and right ideological spaces, that would strongly suggest that the those scales do not measure a real ideological distinction. The twenty year MRG data, however, does not show significant leap-frogging, and shows that parties generally remain within their left or right spaces.⁽¹⁷⁾

There is further systematic evidence that the MRG scales are a valid measure of ideological position compared to leading alternative measures. Besides the MRG left-right scales, the leading approach to measuring parties ideological positions is via the expert judgements made by party specialists.⁽¹⁸⁾ A systematic statistical analysis comparing the outcomes of this measure with party positions plotted over 20 years on the MRG scale, indicates a correlation in the ideological positions provided by both measures (with the exception of a small number of specific parties, under particular circumstances).⁽¹⁹⁾ As well as this, the MRG scale has certain advantages over expert judgements. The scale is more reliable in the sense that it is a less subjective, and easily repeatable, measure. It is also more fine-grained in being able to measure degrees of ideological movement over time. Expert judgements place parties in more static ideological locations, and are consequently less suited to investigating party dynamics. These considerations suggest that manifesto-based measures of ideological position are not only as good as the leading alternative measure, but are better in key ways.⁽²⁰⁾

The MRG left-right scales, as they have been refined over the years in the light of extensive comparative empirical data, have a strong claim to being a sound and useful measure of parties ideological positions over time. The scales are not without their limitations something true of most methodological tools and scales but they are as valid and reliable as alternatives, and arguably considerably more so.⁽²¹⁾

The Ideological Positions of the Two Major Parties Changes and Relationships

The remainder of this chapter will outline and spatially map some outcomes of the analysis of ALP and Liberal Party election policy launch speeches. It will begin by mapping historical ideological trends from existing policy speech analyses for the period 1946 to 1990, and then turn to analyses conducted by the author for the purposes of this monograph for the period 1993 to 2004. The policy launch speeches that are analysed for this second period are as follows over page:

Election	Liberal Party	Labor Party
1993	Liberal Party Campaign Launch Address by John Hewson MP Sydney, March 1, 1993	Advancing Australia; Building on Strength The Hon Paul Keating MP Bankstown, February 24, 1993
1996	Election Campaign Launch Speech The Hon John Howard MP Ryde Civic Centre, February 18, 1996	Speech by the Prime Minister: ALP Campaign Launch The Hon P. J. Keating MP Melbourne, February 14, 1996
1998	A Stronger Australia The Hon John Howard MP September 20, 1998	Kim Beazley s Plan for the Nation Jobs, Security and Opportunity The Hon Kim Beazley MP Brisbane, September 23, 1998
2001	Address at the Federal Liberal Party Campaign Launch The Hon John Howard MP Sydney, October 28, 2001	ALP Campaign Launch The Hon Kim Beazley MP Hurstville, October 31, 2001
2004	Address at the Coalition Campaign Launch The Hon John Howard MP Brisbane, September 26, 2004	Taking the Pressure off Families Mark Latham MP Brisbane, September 29, 2004

Historical Trends in Ideological Position

To give an idea of the longer-term historical trends, Chart 9 below maps ideological positions that have been identified for the Liberal Party and the Labor Party at all elections between 1946 and 1990.⁽²²⁾ As can be seen, the two parties have a history of converging ideologically as well as diverging at various times. However, during this 45 year period the Labor Party has generally remained on the ideological left (with a mean position of 13.4 points on the left), while the Liberal Party has been predominantly positioned on the ideological right (with a mean position of 20.8 points on the right). There have been a number of points at which Labor has been positioned on the ideological right (in 1966, 1980, 1983 and 1984), and the Liberal Party on the ideological left (in 1963 and 1966), where at the 1966 election the Liberal Party was marginally further left than the Labor Party. (Budge, Klingemann and Volkens, et. al., 2001 provide a number of possible causes and explanations for these cross-overs.). Importantly, it can also be seen that on some occasions the parties positions have converged as a consequence of one party moving closer to the other one, which remains relatively stable in position (e.g. as in 1963 and 1966 elections when the Labor Party was relatively stable in position, but the Liberal Party moved considerably toward Labor); and on other occasions, the parties positions have converged as a result of both moving closer to each other (e.g. in 1974, 1975, and 1977). Furthermore, the various convergences of the parties have not necessarily been convergences toward the ideological centre between left and right. In the 1960s, the parties converged toward the ideological left, and in the 1980s they converged toward the right.

Chart 9: Party Ideological Positions (1946 1990)

Source: Budge, Klingemann, Volkens, et. al., 2001

Recent Trends in Ideological Position 1993 2004

Analysis and mapping for the period between 1993 and 2004 conducted for this monograph show that similar patterns apply. Chart 10 indicates that in this period, the Labor Party remained on the ideological left and the Liberal Party on the right. Between 1993 and 1996, there was a slight move by Labor to the right, probably as a reflection of Keating s strong emphasis on economic issues in the 1996 election, particularly the drive to increased productivity and strengthened economic relations with South East Asia. In 1998, however, Beazley renewed the emphasis on health, education and employment in Labor s stated policy priorities and the Labor Party moved left again, to the position it occupied in 1993. In 2001, Labor turned slightly right again (probably due to the salience of security and border control issues in Labor s 2001 policy priorities). In 2004, Mark Latham brought Labor s election focus again to health and education, and the Labor Party returned to the left position it had occupied in 1998.

Chart 10: Labor Party and Liberal Party Ideological Positions 1993 2004

Source: 1993 2004 Policy Speech Coding and Spatial Analysis

The Liberal Party showed a consistent trend to the right between 1993 and 1998. It then moved in a distinctly left direction in 2001 and maintained that position at the 2004 election. Moving further to the left is not new for the Liberals (after all the 2001 analysis of Budge, Klingemann, Volkens, et.al., showed the Liberals to be further left than Labor in 1966), and Liberal Party ideological positions in 2001 and 2004 were only marginally further left than the position occupied in 1993. Nonetheless, the movement in 2001 is perhaps a little unexpected in view of the Liberal Party s then strong pitch for security, sovereignty and border control emphases that are of the right on the MRG scale, and ones that also helped bring Labor further to the right in 2001. However, even though military and security issues were salient in the Liberal Party s election policy speech priorities in 2001, there was also a strong emphasis on MRG scale left issues, like the expansion of education and welfare provision in various social areas. That emphasis was maintained in the 2004 election, with its very strong focus on Medicare.

The ideological stability of the parties can be judged in terms of deviations from their party s mean ideological position in the 1993 to 2004 period. Despite having had three leaders taking the Labor Party to elections in this period, that party has been more ideologically stable than the Liberal Party. Labor has experienced a smaller average deviation from its mean ideological position (3.9 chart points average deviation), than the Liberal Party has from its mean position (6.7 average deviation). Again, judged in terms of their mean ideological positions, the Liberal Party has on average been closer to the ideological centre position than the Labor Party over this period.

The Parties Economic vs Non-economic Policy Ideologies

The ideological positions mapped in Chart 10 are based on the parties entire range of left-right policy categories, both economic and non-economic. Once those two policy streams are measured separately for the ideological positions they reflect, some interesting differences emerge.

With the exception of 1998, both parties placed greater emphasis on non-economic policy issues than economic ones in their stated election policy priorities between 1993 and 2004. In 1998, the GST was a major issue for both parties, and both also gave considerable emphasis to issues of productivity and economic management. The greatest emphasis placed on non-economic policy issues compared to economic ones by the Labor Party occurred in Mark Latham s 2004 speech. There were five times more non-economic than economic policy references. The closest to this was in 1993 under Keating, where there were nearly twice as many non-economic policy references as economic ones. The greatest emphasis given by the Liberal Party to non-economic compared to economic issues occurred in 2001, where there were over twice as many of the former. Along with education, priority issues in this election for the Liberals were border control, security, social harmony and family. These contributed to the greater overall focus being on non-economic policy.

Proponents of the view that the Labor Party has become more like the Liberal Party sometimes cite the example of economic policy to support this. Chart 11 maps the parties positions for economic and non-economic policy streams. It does reveal some economic similarities between the two parties, not so much in their ideological convergence on economic policy, as much as the fact that Labor, together with the Liberal Party, has consistently been economically right of centre between 1993 and 2004. The exception was in 1998, when economically Labor was ideologically centre. Despite both being predominantly right on economic ideology, there was still significant ideological distance between the parties for most of the 1993 to 2004 period. Between 1993 and 1998, the parties underwent a period of divergence in economic ideology, with Labor showing a greater movement away (toward the left) from the Liberal Party s ideological positions. In 1998, Labor was economically its furthest left from the Liberals, most likely a reflection of the Labor Party s policy stance on the GST.

Chart 11: Labor Party and Liberal Party ECONOMIC and NON-ECONOMIC Ideological Positions 1993 2004

Source: 1993 2004 Policy Speech Coding and Spatial Analysis

The parties divergence, and the Labor Party s movement left, changed markedly between 1998 and 2004. During this time Labor moved consistently to the right in its economic policy ideology, converging on the positions of the Liberal Party, even as the Liberal Party moved further right itself economically. In 2004, the Labor Party was economically the closest it had been to the Liberal Party within the 1993 to 2004 period, and both parties were economically the furthest right they had been in that period. Even though Mark Latham, at the 2004 election, focused predominantly on non-economic policy issues in his election policy launch, the economic policy focus that there was, emphasised ideologically right categories, such as economic incentives (e.g. income tax cuts, raising the threshold for top marginal tax rate) and efficiency and economic orthodoxy (e.g. budget savings from cuts to state bureaucracy, downward pressure on interest rates, responsible economic management). John Howard also emphasised these economic policy categories, but placed more emphasis on small business and economic enterprise. Measured in terms of the parties average deviations from their respective mean economic ideological positions for the period between 1993 and 2004, the Liberal Party has been more stable in its economic ideology than the Labor Party (average deviation of 6.7 chart points, compared to 15.1 for Labor).

When it comes to the parties ideological positions on non-economic policy, a different picture emerges. As Chart 11 shows, Labor was on the ideological left in its non-economic policy between 1993 and 2004. Perhaps unexpectedly, so too was the Liberal Party, with the exception of 1996 when it occupied a position marginally to the right of centre (by 2 chart points). This left positioning can be accounted for by the fact that an emphasis on the expansion of education and training, and expansion of support for health, has been salient in the Liberals election policy launches (the left categories of social welfare expansion positive, and education expansion, respectively). In fact, as a proportion of non-economic policy, these references have steadily increased in those launches (from 26 per cent in 1993 to 58 per cent in 2004), with the exception of 1996. In 1996, although John Howard expressed a commitment to Medicare and undertook to expand training, he also placed significant emphasis on freedom of choice and family values (right categories), and also expressed a strong focus on environmental issues (a left-right neutral category).⁽²³⁾ These together brought the Liberals non-economic policy position in 1996 to the right of 1993 and subsequently. In 2004, the Liberal Party moved significantly further left in its non-economic policy position, largely due to the dominance of the Medicare Plus proposals to strengthen bulk-billing and reduce out-of-pocket medical costs. There was also a significant emphasis in the 2004 launch on expansion of school education and training (for the national skills shortage), expansion of childcare support, and support for older Australians. Between 1998 and 2004 there was a convergence between the two parties in their ideological positions on non-economic policy. In moving right between 1998 and 2001, the Labor Party converged on the Liberals position in 2001, which had remained virtually the same since 1998. However, in 2004, while Labor had moved slightly left of its 2001 position, the Liberal Party also moved left to strongly converge on Labor in 2004.

The Labor Party, as stated, was further left in each of its non-economic ideological positions than the Liberal Party between 1993 and 2004. With the exception of the 2004 election, the Labor Party placed significantly more emphasis on social welfare and education expansion than did the Liberals. In 1998, in the context of the GST issue, Kim Beazley focused more on economic than non-economic policy in his election policy launch (60 per cent compared to 40 per cent of policy references). However, within the non-economic policy, there was a very strong emphasis by Labor on increased health funding, and funding for school education and post-school education and training. As well as this, in 1998, Labor made little reference to policy issues within MRG right or left-right neutral categories.

Measured in terms of the parties average deviations from their respective mean non-economic ideological positions for the period between 1993 and 2004, the Liberal Party has been very slightly more stable in its non-economic ideology than the Labor Party (average deviation of 9.6 chart points compared to 9.7 for Labor).

One of the upshots of this spatial analysis of party positions is that while the last five elections have shown the parties to have ideologically diverged as well as converged at various points, the analysis does strengthen one of the general views on how the parties have converged more enduringly. It was noted earlier that some analysts had speculated that the two parties had become much more similar in virtue of the Liberals coming to accept state provision of welfare, and Labor a free and efficient market economy. The spatial analysis above to some extent confirms this. Throughout the 1993 to 2004 period analysed the Liberal Party has strongly emphasised welfare and education expansion (placing them more left and closer to Labor), while the Labor Party has consistently emphasised economic incentives, efficiency and orthodoxy, along with the importance of enterprise in economic policy (placing them more right, and closer to the Liberal Party).

It is significant also, that the observations of the previous chapter on the relative philosophical directions reflected in the parties flagship policies in the 2004 election conform with the spatial analysis in the present chapter. It was observed earlier that the ladder of opportunity theme in Mark Latham s flagship economic policy, and the emphasis that policy placed on individual and civic responsibility, all echoed key elements of the devolutionism in John Howard s Liberalism. This accords with the spatial analysis of party positions just conducted. Labor, it was shown, moved significantly to the right economically in 2004 to converge on the economic ideological position of the Liberal Party. It was also observed in the earlier case study that the Liberals approach to Medicare concessional benefits suggested a downplaying of their usual user-pays principle, and increased emphasis on collective responsibility for people s needs something more typical of Labor Party social policy values. The ideological mappings for non-economic policy clearly show the Liberal Party moving significantly to the left in 2004, to converge on the position held by Labor.

One further upshot of the spatial analysis of party positions relates to certain views about the Liberals recent ideological attitude to social policy. A prominent view, and the avowed position of John Howard, is that the Liberal Party is currently socially conservative . It was noted earlier that there will be differences on what counts as social policy, and exactly what the social in John Howard s social conservatism is meant to include and exclude. This means that the non-economic ideological positions analysed and mapped above may not exactly speak to social policy as it is conceived by Howard. Nonetheless, non-economic policy as it has been categorised here for the ideological mappings, does encompass major social policy categories. While it is plausible to think that many social policy measures pursued by John Howard s Liberal Party particularly in social values , the fabric of society and security areas are ideologically of the right, the evidence from the above analysis suggests that the situation may be more complex with social policy more broadly conceived, especially to include social welfare and quality of life domains. The Liberal Party s ostensive approach to social values might not tell the whole story about that party s more comprehensive ideological dispositions in the non-economic sphere.

Party Positions and Voters Positions

The preceding discussion mapped out some of the recent ideological relationships between the parties, and the movements in these positions over time. A key issue of interest for this monograph, however, is whether the parties are abandoning or compromising their usual ideological stances in order to pragmatically target voters. An important question, then, is how the parties ideological positions relate to Australian voters ideological positions? Do the changes that have been observed above in the parties positions track those of voters? Do the positions of the parties converge on the ideological positions of the median voter, as Hotelling s median voter theorem predicts?⁽²⁴⁾ Do the parties ideological positions track those of swinging voters? How well matched are the ideological positions of the parties with the positions of their respective constituencies?

How easily these questions can be answered will depend on how validly these voter groups ideological positions can be identified and measured. It has been argued that the MRG party ideology scale, with its statistical basis and validity testing, is a comparatively strong measure of party ideological position. But there does not appear to be a similarly rigorous (and independent) measure for voter ideological position.⁽²⁵⁾ However, there are some indicators of this that are available, namely, surveyed voters own direct and indirect estimations of their left-right ideological positions. Although, as measures, these self reports might not inspire the same degree of confidence as more rigorously tested ideological scales, they can still be useful in providing a preliminary analysis of ideological relationships between parties and voters. The following makes use of Australian Election Study data on self-reports of surveyed voters left-right positions, along with self-reports on some of those surveys other ideology relevant variables, to chart some ideological relationships between parties and voters. It is important to keep in mind when viewing the charts that, although they are represented on the same -100 to +100 left-right scale, the positions of voters, and positions of parties have been arrived at via different measures.

Ideological Positions of Voters, Major Party Supporters and Swinging Voters

Chart 12: Party Ideological Positions and Median Surveyed Voter Ideological Positions 1993 2004

Source: 1993 2004 Policy Speech Coding/ Spatial Analysis and Australian Election Studies 1993 2004

Chart 12 maps the median ideological positions of all surveyed voters, voters who identify with the Labor Party, and voters who identify with the Liberal Party for the period 1993 to 2004.⁽²⁶⁾ Chart 13 maps the median ideological positions of those voters who sometimes vote Labor and sometimes Liberal (i.e. swinging voters). These median voter positions are plotted against the background of the already identified ideological positions for the Labor Party and Liberal Party (shaded red and blue continuous lines). As can be seen from Chart 12, the positions of the aggregate All Surveyed Voters (green segmented line) were consistently on the ideological right for this period, although they were only marginally to the right of centre. On average, the Liberal Party was closer to these surveyed voters ideologically than the Labor Party, the mean all-voter position being 2.7 on the right, and Liberal Party and Labor Party mean positions being 15.6 on the right and -29.4 on the left, respectively. Ideologically, the Liberal Party has been very close to surveyed voters at the last two elections, and Labor was at its most distant in 2004.

Chart 13: Party Ideological Positions and Median Surveyed Swinging Voter Ideological Positions 1993 2004

Source: 1993 2004 Policy Speech Coding/ Spatial Analysis and Australian Election Studies 1993 2004

The disaggregated election survey data shows that Labor Party identifiers have been ideologically just left of centre, and consistently so, while Liberal Party identifiers have occupied positions considerably on the right. On average, the Liberal Party has been ideologically closer to its constituency than the Labor Party has been to its constituency. The mean ideological position for Liberal identifiers is 25.5 on the right compared to 15.6 right for the Liberal Party, and for Labor identifiers -6.8 left compared to -29.4 left for the Labor Party. Indeed, on average the Liberal Party is ideologically very slightly closer to the mean ideological position of Labor supporters than is the Labor Party (22.4 chart-points compared to 22.6). Interestingly, in 1998 when the GST was the dominant election issue, the Liberal Party was slightly right of its constituency, but the closest it has been to them in the 1993 to 2004 period.

As might be expected, the ideological positions of surveyed voters who report sometimes voting Labor and sometimes Liberal, are very close to centre. Although on average this swinging voter group was slightly to the right of general voters (with mean ideological positions of 3.7 right and 2.7 right respectively), swinging voters were closer to centre at the 2001 and 2004 elections. Again, the Liberal Party is on average closer to the mean ideological position of swinging voters than the Labor Party.

Swinging Voters and Optimal Party Positions

It was noted in the introduction to this monograph that, in the context of partisan de-alignment, competing parties may be more inclined to target the preferences of non-committed voters. If this were the case, what would be the optimal ideological position(s) for parties to occupy in relation to the positions occupied by swinging voters? And what light, if any, might the party and voter ideological maps charted above throw on whether the parties have situated themselves optimally?

It is plausible to assume that in general the ideologically closer a political party is to a voter s ideological views, the more attractive that party will be to the voter. It is also plausible to assume that a party, in seeking to attract non-committed voters, will not want to alienate and lose the vote of its already committed voters (i.e. its party identifiers or supporters). These two assumptions suggest that the optimal ideological position for a party will be between the positions of committed and non-committed voters. Just whereabouts in between will depend in part on how strong the existing support is from the party s committed voters. The stronger the commitment from committed voters, the further away the party can afford to be from them ideologically (and the closer to non-committed voters) without losing their votes.

Based on the ideological positions mapped in Charts 12 and 13, the Liberal Party has indeed been optimally positioned between 1993 and 2004 to attract swinging voters without alienating its existing support base. The median ideological position for swinging voters was 3.7 chart-points on the right, for Liberal identifiers was 25.5 points on the right, and for the Liberal Party, between these at 15.6 points. In only one year (1998) was the Liberal Party not between the two, and only very slightly so, to the right of its existing supporters.

The Labor Party, on the other hand, does not appear to have been optimally placed in relation to swinging voters. Its ideological positions between 1993 and 2004 have been consistently to the left of both swingers and its existing supporters. In line with the assumptions and observations made here, it appears that the Labor Party would have been better placed to attract non-committed voters and to maintain its existing support base if it had been positioned more at the ideological centre.

Voters Economic Ideological Positions

According to the earlier analysis and mappings of ideological positions in economic policy between 1993 and 2004, both parties were clearly on the right economically, and since 1998 both became more so. How does this compare with voters economic ideological views? While the Australian Election Studies did not include questions specifically about voters perceptions of their leftness or rightness in economic matters, it is possible to construct an index of their left-right economic ideology, based on responses to certain other questions in the studies.⁽²⁷⁾ The index used here is based on the view that a person s economic leftness or rightness will be revealed by their attitudes to who should hold economic power in society and how economic resources should be distributed. Accordingly, the index is based on responses to questions relating to trade union power and big business power, and income redistribution and social spending versus individual tax cuts.⁽²⁸⁾ Median surveyed voters responses to these questions are averaged to form a left-right economic ideology score for each survey year. The index is scaled to have a range of minus-100 (left) to plus-100 (right) in conformity with the spatial scales for the manifesto analysis.

Chart 14 maps voters economic ideological positions derived from applying this index, again against the background of the already plotted economic ideological positions for the two parties, on the same scale. While both the parties have been right of centre economically between 1993 and 2004, and indeed, have been converging even further right between 2001 and 2004, voters appear to have been moving leftward in their left-right economic attitudes. Measures for the all surveyed voters group show its position to have moved from right to left of centre since 1993. Between 1996 and 2001 there was a notable shift to the left in this group economically, after which it remained in the same position. Surveyed voters who identify with the Labor Party show a virtually parallel pattern of movement in these periods, although this group has remained consistently on the left economically. Surveyed voters who are Liberal supporters have been on the right economically since 1993, with the exception of 2004 when they were positioned at the centre. Again, since 1996, those Liberal supporters have moved steadily leftward in their economic attitudes, with that leftward movement slowing slightly between 2001 and 2004.

Chart 14: Party Economic Ideological Positions and Median Surveyed Voter Economic Ideological Positions 1993 2004

Source: 1993 2004 Policy Speech Coding/ Spatial Analysis and Australian Election Studies 1993 2004

The parallel in direction of ideological movement on economic matters between all these distinct voter groups is quite notable. All show either a movement to the right or a static ideological position between 1993 and 1996. But between 1996 and 2001, all show a distinct movement to the left, and then either a rightward movement or a decline in leftward movement between 2001 and 2004. These same ideological movements are not mirrored by either of the two parties, with the exception of both parties pronounced movement to the right between 2001 and 2004.

In economic ideological position, the Labor Party is on average closer than the Liberal Party to the mean positions of the all-voters group, the Liberal supporters and Labor supporters (i.e. all represented groups). The Liberal Party is also closer in economic ideology to Liberal supporters than the Labor Party is to Labor supporters (a chart-point difference of 37.2 compared to 49.6 between mean ideological positions). However, the Labor Party is considerably closer to Liberal supporters economically, than is the Liberal Party (a chart-point difference of 7.6 and 37.2 respectively).

In general, there seems to have been a divergence in the economic ideological positions between the parties (which were on the right and generally moving rightward) and surveyed voters and party supporters (who were often on the left, and generally moving leftward).⁽²⁹⁾ All told, because the Labor Party was further left than the Liberal Party, it was on average closer to all these voter groups economic ideological positions.

Ideological Difference and Electoral Outcomes A Systematic Relationship?

Harold Hotellings median voter theorem implies that parties will maximise their electoral success in majoritarian democracies the closer their policies are to the ideological position of the median voter. According to the theorem, success for a party means converging on the ideological position of the median voter. Similarly, it has been proposed that, in the context of partisan de-alignment, parties are more likely to target the preferences of non-committed voters in order to increase their prospect of success. However, in tension with these views it is sometimes suggested that a party will increase its electoral success if its policies or ideological perspective diverge from those of its opponents. Is a party s seeking to differentiate ideologically from its opponents a strategy that maximises its prospect of electoral success? Will a party in opposition have more chance of winning office if it makes itself significantly ideologically different from the incumbent at election time? If it offers something significantly different, in other words? Information based on the MRG ideological mapping techniques do not support an affirmative answer to either of these questions.

Chart 15 pictorially represents the relationship between the parties proportions of the primary vote between 1993 and 2004, and the ideological distances between the parties in these years (as analysed and measured via MRG methods). The bars on the bottom of the chart are the total primary vote for each party, and the horizontal area maps the difference in the ideological positions of both parties, the wider the area, the greater the difference. It turns out that there has been no significant positive or negative correlation between the parties ideological distances and their proportions of the primary vote for the 1993 to 2004 period. There are no statistically significant relationships that suggest either (i) that the difference in the primary vote between the parties has gone up or down in relation to increases or decreases in the ideological distance between the parties; or (ii) that the primary vote for a party has gone up or down in relation to the leftness or rightness of that party. An analysis of ideological distance and primary vote in elections between 1949 and 1990, show the same to be the case. It is worth noting, finally, that the Labor Party has won office when it has been ideologically close to its major party opponent as well as distant. The same is true for the Liberal Party.

Chart 15: Ideological Difference and Proportion of Primary Vote 1993 2004

Source: 1993 2004 Policy Speech Analysis Australian Election Studies 1993 2004

What does all this suggest about left-right ideological differentiation as something that a party should seek? If a party seeks it for strategic reasons, to enhance its electoral success, then it is not necessarily a winning strategy according to the measures here. There is also a view that parties in opposition tend to gain office with an agenda for change. To the extent that an agenda for change is something that would be reflected in the left-right ideological distance of that party s election policy program from its opponent s, then this view needs to be reassessed as well.

Endnotes

1. Goot, 2002.
2. Jaensch, 1989.
3. Cassin, 2003.
4. Mackay, 1993, p. 169.
5. Head and Patience, 1979.
6. See, for example, Gruen 1985. An element of this measure would very likely be comparative levels of public sector employment under different governments. On this, recent data show that the absolute number of Commonwealth public sector employees to have fallen steadily by 41 per cent between 1984 and 2005. See Australian Bureau of Statistics, 2005.
7. This MRG commenced its research under the direction of Ian Budge in 1979, and has published a series of major reports on its ongoing findings since then. These reports primarily include Budge, Robertson and Hearl, 1987; Laver & Budge 1992; Klingemann, Hofferbert, Budge, et. al., 1994; and Budge, Klingemann, Volkens, et. al., 2001.
8. However, manifesto data for Greece, Japan, Portugal and Spain was not examined for the full fifty three year period.
9. Budge, 1987. As of late 2002, the Comparative Manifestos Project (as it has now become) had analysed 2,347 manifestos of 632 significant parties in 52 countries, (the bulk of the newer countries being from Eastern Europe and Latin America). See Budge, 2002.
10. An early version of those categories was first identified and reported in the seminal work of David Robertson on party competition. See Robertson, 1976. The MRG refined these through further empirical work. The left-right categories are a subclass of a broader set of major policy areas that have been recurrently emphasised in party election manifestos between 1945 and 2001.
11. Sources, Budge, Robertson and Hearl, 1987; Laver & Budge 1992; Budge, Klingemann, Volkens, et. al., 2001. The coding framework for the content analysis supplied in these sources, specifies which policy subjects or expressions are to be counted under each of these left-right scale categories.
12. For example, Budge and Hofferbert 1990; Budge and Hofferbert 1992; Petry, F. 1995.
13. See, for example, Imbeau, Petry, and Lamari, 2001. For a discussion of the methodological issues involved in measuring a correspondence, see Chapter 2 of Budge, Klingemann, Volkens, et. al., 2001.
14. McAllister and Moore, 1991.
15. Robertson, 1976.
16. Budge, 1987.
17. Budge, 1994.
18. For more details of this approach, and some of its drawbacks, see Budge, 2000.
19. See McDonald and Mendes, 2001.
20. It is also of note that expert judgements of party ideology encompasses not just parties election statements, but also their policy actions and later party behaviour. The fact that the MRG scales based on election statements - provide ideological positions that match expert judgements, further weakens the criticism that the MRG scales are inaccurate measures of ideology because they are confined to policy speeches and do not encompass what parties do.
21. Notwithstanding this, it should be kept in mind that the mappings produced in this current text are the result of codings performed by one person (albeit in conformity with the prescribed Manifesto Coding Instructions[1], and with each speech coded twice). Coding of the same speech material by a number of others might result in less person specific frequencies.
22. As identified in Budge, Klingemann and Volkens, et. al., 2001. Other analyses and spatial mappings also exist for many of these elections. See Budge, Robertson and Hearl, 1987; and McAllister and Moore, 1991. The Budge, Klingemann and Volkens, et. al., 2001 analysis mapped here is the most recent and extensive, however.
23. The strongest focus (at 20 per cent of all non-economic policy references) placed on environmental issues by either the Liberals or Labor in policy launch speeches between 1993 and 2004 (with Keating in second place at 15 per cent in 1996).
24. See note 10 above.
25. There is, however, a dependent one based on the MRG measures. See Kim and Fording 1998. For the purposes of expressing the relationship between voters ideology and party ideology (using MRG techniques), an MRG independent measure of voter ideology would methodologically be called for.
26. Based on responses to the Australian Election Studies questions In politics, people talk about left and right . Generally speaking, where would you place your views on this left right scale? ; and Generally speaking, do you usually think of yourself as Liberal, Labor, National or what?
27. This index is largely based on one utilised in Charnock and Ellis, 2004, though there are differences.
28. The questions are: The trade unions in this country have too much power ; Big business has too much power ; Income and wealth should be redistributed towards ordinary working people ; and If the government had a choice between reducing taxes and spending more on social services, which do you think it should do? . The questions all call for a response on a 1 5 scale (the first three on the scale strongly agree strongly disagree , the last on the scale strongly for reducing tax strongly for social services ). On the first and last question, a response of 1 is counted as ideologically right, moving through to 5 as left. On the second and third question 1 is counted as left, moving through to 5 as right.
29. So, within the limits of the measures and findings here, Hotelling s median voter theorem is not confirmed for parties economic positions.

---