Australian Politics Papers A-K

Malcolm Alexander
Griffith University

The 'Small World' of Australian company directors

Abstract:

Networks of acquaintanceship and personal contact among the managers and directors of large companies have the potential to foster community identity and cohesion among the ‘corporate elite’, national ‘business leaders’ or ‘capitalist class'. Such identity and cohesion, or its lack, can be an important factor in the development of business-government relations. This paper examines the network of personal contacts, created by interlocking directorates, among all board members of the Top 500 publicly listed Australian companies. In the past it has been difficulty to analyse such sparse networks, however recent advances in our understanding of the 'small world' phenomenon and the application of random graph theory to large networks provide new tools to measure the connectivity and density of large networks of this type. The paper analyses the 'small world' of Australian company directors in 1976 and 1996 with some reference to the U.S. network of 1996. These comparisons suggest that the personal network of directors is less prominent in Australia than the US but remains as closely connected in 1996 as in 1976 despite an increase in the numbers of people drawn into the network.

Introduction:

Business elite studies, inter-corporate networks and the network of personal contacts among company directors

The study of business-government relations involves an imbalance of information. On the government side we know a great deal about the organisation of policy advising machineries, within both state bureaucracies and parties. Also, avenues of discussion and decision-making are, to a large extent, public. On the business side there are no comparable structures. Policy advice is generated by a range of industry and ‘peak employer’ bodies. There are few mechanisms for coherent and decisive debate about issues nor clear avenues of discussion and decision-making. Individual firms, industry associations and employer associations all lobby for their own interests on any given issue and attempt to ‘get the ear’ of the most relevant decision makers.

Despite the lack of overt organisation most theorists hold that it possible to speak of a ‘business’ view on economic and social issues. For some it arises from the embedded structural power of business to which the state responds despite the lack of overt business organisation . In this view the state is drawn into providing de facto forums of business class organisation. Alternate theories recognise various degree of business class or business community agency. Perhaps the most explicit recognition of overt agency is power elite theory. Power elite theories, still usefully exemplified by C.W. Mills classic work, claim that consensus among business, government and other elites arises from the homogeneity of social origins, professional training and career networks among the members of these elites .

Australian sociologists and political scientists have studied business from this range of perspectives. In the 1960s Sol Encel provided a good account of the power elite of that era . In the 1970s Bob Connell provided a more sophisticated theory that stressed the closeness of political and big business leaderships but also gave a prominent role to the media using Gramscian notions of organic intellectuals and hegemony . Higley and others carried out an extensive survey of Australian elites in 1974 but their theoretical defence of elites troubled Australian audiences and there was virtually no subsequent work of this kind. At the end of the 1980s McEachern outlined the inner circles of business leaders who were influential in the Hawke years. An account of media glorification of business leaders in the 1990s that continues the Gramscian aspects of Connell’s approach is contained in Graeme Turner’s book, Making it National . However there has been little Australian work in the direct tradition of C.W. Mills and William Domhoff.

A general sympathy with structuralist approaches in sociology in this country means that analyses of business power in Australia emphasise the size and power of the large corporations more than the internal organisation and agency of the business elite. This tendency toward structuralist theory has important implications for the conceptualisation and handling of information about corporate elites and interlocking directorships. From the structuralist perspective, research studies of company managers and directors look at multiple directorship holders just as ‘interlockers’; people who create linkages between corporations. Seen primarily as ‘bearers’ of these structural relations, analysis of personal attributes or personal relations is a secondary exercise. The analysis of interlocking is, in the first instance, simply about networks among corporations, or groups of corporations .

However, elite studies of business leaders and network studies of interlocking directorates begin from the same information base; lists of the directors and managers of major corporations. Both sets of researchers select a population of large companies and then make a census of all persons on the boards of these companies. That they analyse this information in different ways creates an impression of distinct research traditions. On the one side, elite studies focus on the population of persons, what is the average age and gender, what is their background, what are the principal clubs or associations to which they belong etc.? Network studies of interlocking directorates disregard most of this personal information. They shift the material to find those directors who have two or more positions. They then reconstitute this basic information as a set of linkages, a network, between the companies of the original selection set. Corporate power structure research and network based studies do, however, allow for an intermediate form of elite study. High levels of multiple directorship holding, i.e. personal participation in the inter-corporate network, are taken to measure a person’s centrality. The population of ‘big linkers’ (persons with four or more directorships) or ‘network specialists’ is then studied with traditional survey and interview techniques .

A consequence of this methodological division of labour is that the expertise of social network analysis is almost never applied to study the network of contacts among company directors that arises from boardrooms as meeting places. This paper seeks to fill this gap in the research literature. There are, however, practical and theoretical reasons for the division of labour described above. The practical reasons relate to the difficulty of dealing with the very large networks of individuals that these datasets entail. There are, typically, some 2,500 or 3,000 persons in the datasets and it is only recent advances in microcomputer capabilities that permit analysis of such large networks. The theoretical difficulties come from the conceptual differences between studying small, relatively complete networks, the principal terrain of social network analysis, and the large, sparse networks typical of interlocking directorate data. The computing difficulties need no further comment however the theoretical sections of this paper deal with the theoretical and conceptual issues.

The small world of company directors: Network connectivity and structure

The networks of personal contacts among company directors generated by board memberships are very sparse. Traditional social network analysis has developed mostly in small group settings. The directors’ networks of personal contact are so sparse that the techniques and theory of social network analysis have little immediate application to these networks. Thus, for instance, most social network theorists see relative density, the number of ties made between persons as a proportion of all possible ties, as a fundamental feature of networks. In the case of directorship networks this measure is very small, usually only a fraction of one percent. With indicators of such extreme values, insights from traditional social network analysis seem to have little relevance to these directorship networks.

Recent discussions in network analysis have opened new approaches to the conceptualisation and measurement of large, sparse networks. These discussions arose partly from the need to study very large networks such as the internet. Other innovations were stimulated by graph theory, the mathematical basis of network analysis, particularly advances in the theory of random graphs. Another impetus has been a renewed interest in understanding the dynamics of the ‘small world phenomena’.

The impact of this new thinking is to shift the conceptual idea of networks away from a concern with density towards a concern with connectivity. Connectivity is an attribute of the whole network. It reflects its sparseness or density, the basic concept of social network analysis. However, whereas the traditional concept of density sets as its benchmark the situation where every person in the population is directly connected to every other person in that population, connectivity simply requires that every person have some point of connection to the network. Network connectivity and density are only loosely related. They converge only at the extremes. The maximally dense network where all persons are directly connected to every other person is, of course, fully connected. At the other extreme there are only two possible forms of the minimally dense, fully connected network. In between these extremes however connectivity and density can come in many combinations.

Its two minimal forms, the tree and the star illustrate the idea of connectivity. Any population (of size N) requires a minimum of N-1 ties to connect all its members. The connections can be made in two configurations.

Network Example 1. Network Example 2.
4 Points, Star Configuration 4 Points, Tree Configuration

Both these populations are fully connected. Every point is drawn into the network. They achieve full connectivity with only three lines (N-1) compared to the six lines (N(N-1)/2) that are possible among four points. Thus we have connectivity although these are not dense networks.

Connectivity is an attribute of the network itself. It has not an attribute of the points. In order to measure the impact of connectivity at the level of the point (the person) we use the idea of average geodesic distance. The geodesic distance between any two points in a network is the shortest number of steps needed to get from one point to another. For each point we count the number of steps between it and every other point. We then average this total by the number of points reached from that point. This provides the average geodesic distance for that point. The average distance for the network is the average of these point averages. Calculations of average distance for Examples 1 and 2 are given in Table One.

Network Example 1. Star Configuration Example 2: Tree Configuration

The tree and the star also illustrate a key point about network structure, its degree of centralization. Although both networks have the same number of points and lines, the star is more centralised structure. In the star configuration no point is more than two steps away from any other point. In the tree configuration, by contrast, the points at the end of the tree are two steps away from each other. Because of this difference the average distance in the star configuration is less than the average distance in the tree configuration. In the star configuration the three outer points are at distance two from each other but each of them is one step away from the central point. The central point is only one step away from the other three points. The average distance in the star network is 1.5 whereas the average distance in the tree network is 1.66.

A concern with the connectivity of networks, rather than density, allows us to integrate important insights from two other areas of theoretical discussion and mathematical inquiry; random graph theory and discussion of the ‘small world’ phenomena.

Random graph theory takes the number of lines (or connections) in a network, as compared to the number of points (or people), as the key ratio when thinking about connectivity. Random graph theory provides important insights into the random probabilities of connectivity in a network for a given ratio of lines and points. It investigates the patterns of connectivity that can be expected as more lines are added to a network with a given number of points. Random graph theory predicts that increases in connectivity are not simple incremental. There are distinct, although approximate, thresholds. As the number of lines passes half the number of points, a random distribution of these lines will tend toward the creation of one central, large component and a number of isolates. Random graph theory thus provides a way of comparing patterns of connectivity found in real life networks to patterns that can be expected to arise randomly. Nearly all studies of interlocking directorships find inter-corporate networks built around one single large component . Random graph theory suggests that this is the expected pattern given that the number of lines generated by interlocking directorates is usually more than half of the number of corporations selected.

The discussion of networks and the ‘small world phenomena’ comes from sociological and experimental studies. The classic statement of this phenomena was made by Stanley Milgram in the 1960s . He suggested that chains of acquaintanceship would allow a message to reach a particular person through a chain of intermediary steps. He tested this idea by asking a population of randomly selected persons to forward a message to someone who might know the target person of the message. A significant number of messages reached their targets and Milgram calculated that the average number of intermediaries used was around about 6.

The idea that there are just six intermediate steps between any two persons chosen at random from the U.S. population has caught the public imagination through the phrase ‘six degrees of separation’. The phrase was popularised by a Broadway play, now a movie, by John Guare. The basic principles of the small world theory have been further popularised by the movie buffs’ game ‘six degrees of separation of Kevin Bacon’ in which players seek to find the series of co-appearances that link a pair of movie actors. The game takes its name from the fact that Kevin Bacon is the most central person in the network of co-appearances. In a short article in Nature Steven Strogatz, a professor of theoretical applied mechanics and Duncan Watts proposed a mathematical model that explained how the combination of densely connected groups with just a few connecting links can produce a wide network of connections. A more elaborate account of the model is available in Watts .

These popular ideas link the concepts of connectivity (being in the network) and distance (the number of steps between two points). They reduce an immensely complicated and vast structure to a simple and tangible piece of personal experience. What they fail to publicise is the actual chances of making linkages. If one considers that each step in the chain of connections involves one choice out of a large number of possibilities, say 100, then the a successful connection at six degrees of separation is one path out of a possible one trillion paths. From this perspective six degrees of separation is a long, long social distance. Studies of acquaintanceship usually regard links of distance two, ‘friends of friends’, as the practical limit of individuals reachable network. This makes intuitive sense. We can see how this limitation operates in the case of the star and the tree illustrated earlier. While everyone in the star configuration is a friend of a friend (albeit the same friend), the persons on the ends of the tree formation are not friends of a friend.

Is it possible to give an absolute meaning to these measures of distance? We can take six degrees of separation as an outer limit, the average distance we can expect to find among the general population. At the other extreme the average distance within the star configuration represents the minimum distance for connectivity. Although small networks can connect at average distance less than two, as in the two examples, the average distance approximates to 2 for any relatively large network. In absolute terms, therefore, average distance varies between 2 and 6. However, due to the multiplier effects noted above this is a geometric scale, akin to the Richter scale.

The small world theory and the debate and ideas it has generated give social substance to the network concepts of connectivity and distance outlined earlier. The possibility of making a connection to any other person in the world through six or so intermediaries is intriguing despite the extreme probabilities involved. The importance of connections at distance two, friends of friends, suggests a related measure of network connectivity however. In settings where we know, or can estimate, the total number of people in a network, the proportion of people falling within the distance 2 of an actor (their acquaintanceship network) indicates the scope and reach of that person’s network. Thus, average distance and the average acquaintanceship network are, together, good indicators of the connectivity of an inter-personal network.

In summary, the small world of company directors can be understood as the network of personal connections created by direct acquaintanceships in boardrooms facilitated by the potential contacts (‘friends of friends’) available through board interlocks. The basic measures of connectivity in this acquaintanceship network are the average distance between persons and the proportion of persons falling within distance two of the average member of the network. But these are crude aggregate measures. Just as the same statistical mean may have various skewed or even bimodal distributions behind it, there may be different network structures behind these aggregate measures. In this paper we will also investigate one of these structural measures, the centralization of the network.

Small Worlds: Australian company directors 1976 and 1996

In this section we report findings from a comparison of the personal networks, created by board memberships, among the directors of Australia’s top 250 companies in 1976 and 1996. The population for analysis was initially taken as the directors of all 250 companies. When we analyse the connections among them, however, the population in the network reduces to those persons sitting on a board with a director who interlocks with another board in the main component of the network. The effective population of the small world comparison is, thus, a sub-population of the initial population. We present information about the small world of connections among this sub-population. We count the number of direct and second degree (‘friends of friends’) contacts of each person and the average distance between a person and any other person in the network.

In summary, the changes between 1976 and 1996 show that the small world of Australian company directors has become more inclusive, there are a greater number of people drawn into the network in 1996 as compared with the earlier year. Despite the increase in population the ‘small world’ of 1996 is even ‘smaller’ (ie more connected) than that of 1976 when viewed from the perspective of an individual participant. The average distance between people is less and the number of contacts at distance one and two increases both absolutely and as a proportion of the total contacts available in the network.

The population of companies for each year comprises the top 200 non-financial companies, determined by revenue and the top 50 financial companies, determined by assets. These populations were identified using the criteria used in international studies of the 1970s . Information for the companies of 1996 was taken from the annual listing of the Top 1,000 Corporates made by Business Review Weekly. This list covers privately held companies, (including subsidiaries of foreign companies), cooperatives and government-owned enterprises as well as the publicly listed companies for whom information is freely available through stock exchange sources. The initial population of companies for 1976 covered the same range of companies identified from a variety of sources including stock exchange yearbooks, (particularly Jobsons Yearbooks), and a range of business directories.

The directors of companies were identified from annual reports and business directories. The principal source of information for 1996 was the Business Who’s Who of Australia, now compiled and published by Dunn and Bradstreet. The principal source of directors’ information for 1976 was Jobsons Yearbook and, for unlisted companies, the Key Business Directory, also published by Dunn and Bradstreet. Additional reference was made to the Kompass business directory of that year. For both samples the names of all directors were cross-checked and any identical names were investigated more closely to confirm whether or not it was the same person sitting on two or more boards.

The lists of positions were read into the social network analysis program, UCINET . The affiliations command produced the matrix of connections between persons. The distance command then calculated the shortest distance (the geodesic) between all persons of the initial population. The total number of persons that any individual connects to is the size of the component they are in so that all persons in the main component had the same number of total connections. This sub-population of persons in the main component was then analysed for average distance and number of contacts. We were also able to calculate the centralization of the main component.

The first finding was that the sub-population of persons included in the main component was substantially bigger in 1996 than in 1976. The main component of connections is, effectively, the network. Everyone in the main component is, potentially, reachable by anyone else in that component. In 1976 this network covered a population of 982 directors. In 1996 it covered a population of 1162 directors, an increase of 18.3%.

If there are more people to be connected, one might assume that it is harder to make links between them and, as a consequence, the ‘small world’ or each participant, their connectedness to other persons in the network would be less. That is to say, the average distance between persons would increase. In fact, we find a decrease in this measure. The average distance between persons in 1976 was 4.78 steps. In 1996 this has decreased to 4.34 steps. This is a decrease of 9.3%. The small world phenomenon is stronger although the population it covers is larger.

A similar finding applies to the scope of person’s contacts. On average participants in the network in 1976 had 9.76 direct contacts. In 1996 the average number of direct contacts increases to 10.31, an increase of 5.6%. The greatest change occurs however in the number of indirect contacts for the average participant. In 1976 the combined number of direct and second degree contacts was 50.85. In 1996 it was 63.34, an increase of 24.6%.

A feature of a network is the most central person in the network. Who are the Kevin Bacons of the Australian corporate world in 1976 and 1996? The usual measure of centrality for this purpose is the person with the largest number of direct contacts. In social network terms this is the point with the highest degree. Table Two gives all persons with degree greater than 40 in both years. There are fewer people in 1976 but the degree of the top person, Richard Law-Smith is significantly greater than that of the top person, Nick Greiner, in 1996.

The final step of analysis is to consider is the centralization of each network. Centralization is a characteristic of the network itself, not of points in the network. The measurement of network centralization proposed by Freeman and implemented in UCINET measures the extent to which the network configuration resembles the star configuration described earlier. The higerh the index the more closely the network under examination approximates a star. Using this measure of network centralization we obtain a value of 5.5% for 1976 and 3.7% for 1996.

This is a surprising result. The earlier indicators show that the ‘small world’ phenomenon is more evident in 1996 than 1976. Both the average degree and the average number of contacts at first and second degree of separation have increased. Commensurate with these facts the average distance between persons lessened. One would expect that these greater ‘efficiencies’ in the connectivity of the larger network would be a result of greater centralization. This result suggests that there are other forms of structural efficiency beside network centralization available in these networks of boardroom contacts.

Summary of findings and discussion

This paper has examined the network of personal connections among company directors in Australia that arises through their position as members of a company board. The presence of interlocking directors on boards means that even those directors who do not have multiple directorships have the potential to contact directors on other boards as ‘friends of friends’. The ‘small world’ created by these contacts can be measured through the average distance between persons in the network and the scope of contacts they have.

The network among company directors in Australia in 1976 and 1996 has grown in size. It draws in a larger population of directors in 1996 than it did in 1976. At the same time, however, the ‘small world’ of connectivity within that network has become smaller. The average distance between persons has decreased and the number of contacts at distance one and distance two has increased. Surprisingly, however, this increase in connectivity has not been associated with greater network centralization. On the contrary, the network centralization in 1996 is less than in 1976.

The Australian figures need also to be compared with overseas networks. At this stage we do not have comparisons on all the measures presented here. However, a preliminary analysis of the personal networks in the U.S. suggests that the average distance in this network is much less than in Australia. It is only 2.68 in 1996 as compared to the average distance of 4.34 in Australia. This is very significant difference given the multiplier effect associated with average distance. Much of the difference may relate to the way our populations are identified. The top 250 companies in the U.S. are a much smaller part of the business world in that country than are the top 250 companies in Australia. On the other hand the difference is consistent with previous studies of interlocking. It is likely, therefore, that networking among company directors in the U.S. is stronger than in Australia and that the ‘small world’ of connections among them is significantly denser.

The interest in examining these ‘small worlds’ of connectivity among company directors was motivated by a concern to understand the dynamics of the business side of the equation involved in government-business relations in Australia. This paper has described ways to understand the networks of connectivity that are created within sparse networks. These sparse networks can contribute to the feeling of a ‘small world’ even among fairly large populations. We would suggest that there is a tangible and meaningful ‘small world’ created by the networks of interlocking and contacts among Australia’s company directors. That world has remained in place in the 1990s even though the population of people it includes has increased significantly. But the degree of connection within the Australian ‘small world’ still appears to be fairly loose. In comparative terms it is significantly less than the connectivity of the U.S. network.

This difference suggests that the internal mechanisms for generating business consensus in Australia will differ from those of the U.S. In the U.S. ‘peak organisations’ such as the Business Roundtable (the U.S. precursor of the Business Council of Australia) operate in parallel to the boardroom networks of business leaders. In Australia, by contrast, they may be the only effective game in town.

Jeff Archer

UNE

Defining Rural Australia: Sources of Political Backlash in Contemporary Australian Politics

Abstract

This paper considers the use of the term ‘rural’ in Australian popular and political discourse. It is important to see how stereotypical and ill-informed views about the nature of the rural can influence public discourse on vital questions of regional policy and service delivery. The electoral backlash from rural and regional Australia, along with the rise of rural independents and the One Nation party, can also be analysed through this perspective.

Australian debates about rural identity and national identity are longstanding and unresolved. Rural stereotypes include such figures as the rodeo cowboy, the country music star and the outback adventurer. Inland and coastal images are considered, along with views of the rural as agricultural, as pastoral, as remote, as old economy, and as heritage. In particular, the creation of rural identity and rural stereotype is analysed through an examination of exhibits in the Stockman’s Hall of Fame in Longreach.

Introduction

One aspect of the current debate about the future of rural Australia is the way that decision-makers, policy-makers, and regulation makers, especially in government, perceive rural Australia (Archer, 2000). This paper considers some of the uses of the term ‘rural’ in Australian popular and political discourse. In particular, the creation of rural identity and rural stereotype is analysed through the presentation of a popular rural stereotype in the exhibits in the Stockman’s Hall of Fame in Longreach. It is important to consider if stereotypical and ill informed views about the nature of the rural can influence public discourse on vital questions of regional policy and service delivery. The electoral backlash from rural and regional Australia, along with the rise of rural independents and the One Nation Party, can also be analysed through this perspective.

Australian debates about rural identity and national identity are longstanding and unresolved. The stereotypical rural person is depicted in either positive or negative ways, often in oppositional ideological frameworks. To their admirers, and in the self-perceptions of people living outside the major cities, emphasis is placed on the noble and moral community values or countrymindeness that they often believe to be the hallmarks of an authentic Australia. To their urban-based detractors they are more likely to be seen as unsophisticated, racist, environmentally unsound, rustic rednecks. The nostalgia for the bush as a positive stereotype has its alternative reading by those who see rodeos, country music and line dancing as an aberration, imported from the cowboy country of the USA, and regard them as deeply politically regressive activities. These two stereotypes portray the inhabitants of non-metropolitan Australia as either as the salt of the earth or as the agent of salination. These stereotypes paint a very simplistic picture. They are actually the same stereotype viewed from different ends of an ideological spectrum.

This division is very similar to the division in the two main readings of European settlement. One version of Australian national identity perceives the global intruder as the enemy. In this interpretation of Australian history, the culture of the migrant oppressor is embedded in the history of invasion. Alternatively, in a different reading of migration, the pioneer tradition of settlement is drawn on in the politics of settler nationalism, for example in the populism of Pauline Hanson. In particular, Hanson’s world view draws on one of the most potent foundation myths of white settler 'democracy' in Australia, in which notions of equality, representation, rights and justice are all based on exclusions of other peoples (Kane, 1997, 117-131, Thompson, 1994, 22-47).

Consider a few examples of this rural stereotype. Rural people are frequently presented as slow thinking and moving, but open-hearted and naïve when compared with their city cousins (Astley, 1999, 241). Rural stereotypes include such figures as the rodeo cowboy, the country music star and the outback adventurer. Inland Australia is rural in this view. Coastal Australia has a different story associated with tourism and retirement. Unlike most of inland Australia, the coastal strip has been seen as having overwhelmingly positive and desirable stereotypical qualities. Inland Australian rural is more often depicted as agricultural, pastoral, remote, and obsolete. It is therefore seen as part of the old economy.

Media institutions such as the ABC and News Limited habitually use rural as a euphemism for agricultural. The ABC, for example, through its long-running regional radio broadcast, Australia All Over, or ‘Maca on a Sunday Morning’, constructs a version of rural Australia that is nostalgic, remote and wild, and one that claims to evoke the real or authentic Australian identity. On 15^th July, 2001, for example, the host of this show played country music and the ballads of Banjo Patterson, laced together with stories about such topics as wild birds in wild places, the rodeo at Fitzroy Crossing (where lost traditions are regretted because the bull riders now wear crash helmets), outback tourism, wartime knitting, and bush recipes. This is a heritage version of rural inland Australia. Its most famous depiction is found in Longreach. The Australian Stockman’s Hall of Fame and Outback Heritage Centre at Longreach demonstrates what Umberto Eco calls the hyperreality (that is, the quality of being even more authentic than the merely authentic) of an allegedly authentic rural Australia (Eco, 1987).

Longreach

Longreach in western Queensland is 700 kilometres from the coast on the Tropic of Capricorn. Many travellers to this remote town use the Spirit of the Outback train, with its folksy Captain Starlight’s Lounge, the Stockman’s Bar, and the Tucker Box Dining Car. Travellers may also visit the iconic Black Stump at Blackall or stop at Barcaldine to see the Tree of Knowledge Monument celebrating the rural workers of the west who helped to found the Australian Labor movement in 1891. Others fly to Longreach airport where the Qantas Founders’ Outback museum celebrates the birth of Qantas airlines. Coach and car travellers to the north-west of Longreach also take in the so-called Matilda country around Winton. Here tourism is based on tales of explorers and the origins of Australia’s national song, Waltzing Matilda. Around Longreach tourists can visit Oakley Homestead to witness authentic Brahman cattle and Merino sheep, Boree and Coolibah trees.

According to the Hall of Fame Fact Sheet, (October, 1999) the Australian Stockman’s Hall of Fame and Outback Heritage Centre was inspired by the American Cowboy Hall of Fame in Oklahoma City. The air-conditioned three thousand square metres of display area was designed by Feiko Bouman using timber, stone and corrugated iron. It resembles a fallen series of silos or gigantic water tanks. The founders believed that, while the heroes of the "wild west" shaped the "fortitude of the American character" … , " the exploits of the quiet Australian bushman" on the Australian "great and colourful frontier" have not been sufficient eulogised. The initial project was conceived by stockman and artist, Hugh Sawrey. It was to be a memorial to the pioneers of the Australian outback and a record of the heritage of rural Australia. Apart from a major museum and an art collection it now aims to maintain "a research library of outback literature and artistic expression pertaining to inland Australia’s past present and future and encourage the reprinting of lost Australian works".

The Hall of Fame was developed by a private company supported by National Party politician, Bob Katter Senior, art patron, author and grazier, Dame Mary Durack, riding boots and clothing magnate, R.M Williams, Longreach grazier and local government leader, James Walker and many others. Sites including Albury, Rockhampton and Cloncurry were considered before Walker supplied the land for the site and steered the project to his own town. The Queen opened the museum in 1988. It cost $12.5 million dollars to build. The Hall has been a major tourist attraction ever since, attracting a total of around a million visitors. According to the Chief Executive Officer, Peter Andrews (interview with author 29.10.99), most of the tourists are mature aged Australians, mostly self-driven, with less than one-fifth arriving by air. More than half of the visitors come from Queensland, and over a third come from NSW, with only 2.5% of the total coming from overseas. Of the overseas tourists, most come from either the UK or from the settler societies of Canada, USA and New Zealand. Only a handful of tourists come from Asian countries, and these are mostly journalists on working trips. One visitor came overland by camel across the Tropic from Western Australia. Given the huge and varied Queensland coastal tourist industry, the bulk of the visitors to the Hall of Fame come from a very homogenous cultural background.

The museum celebrates the rural as predominantly the pastoral, although the frontier experiences of mining are also accepted alongside the stories of outback sheep and cattle droving. Viewing areas glimpse the grounds where Comet and Southern Cross windmills stand erect. Inside, a major exhibit rejoices in the commercial opportunities provided in arid country by the bores of the Great Artesian Basin. The world of Lawson and Patterson is accepted as if they agreed on the splendour of all things outback, and as if it could all be eulogised as a story of struggle, manhood and sacrifice amidst sensational sunsets. True, pioneer women are honoured for their key supporting and nurturing role, but country towns are largely absent from the story, apart from the stories of agricultural shows (including the popular freak shows) and rodeos.

Manhood is central to this story. It is celebrated as a product of the harsh environment. One caption reads:

"The history and the environment of the Australian male have put into his bloodstream a feeling for games and competitive sport which seems to be necessary for his fulfilment. When Australia was a pioneering community involved in conflict with a land and a climate where victories were never certain and seldom easy, games or contests in which a bloke had a chance to win were … compensation for the dispiriting setbacks and failures … "

The first exhibits introduce the visitor to the ancient continent and to the first Europeans. These form a preamble to displays depicting the nineteenth and twentieth century pioneer experience of stockmen, weddings, funerals and drought. Another caption tells us:

Museum exhibits include the items that make up the assumed rural heritage. These include: the explorers; a hawker’s wagon; the drover’s camp; life in the saddle; a poetry recital; mustering; bushcraft; the saddler’s shop; the blacksmith’s forge; the golden fleece (an idealised shearing shed); gold miners; and advances in communication (including camels, paddle-steamers, posts and telegraphs and aviation). In other displays bushfires, floods, droughts, dust storms and cyclones, along with plagues of mice, locusts, and rabbits are all represented. The library has many volumes dealing with the explorers, bushrangers, station life, the beef empires, and even the bush tucker man. The viewing finishes at the gift shop with its Akubras, branded clothing, and outback reading material. Souvenirs of the Hall of Fame are sold alongside videos of kangaroos, military history, Outback Magazine (with its stories on campdrafts, horses, and 4-wheel drives), audios of Mary Durack’s fiction, James Walker’s 1999 autobiography (un-ironically titled, A Rewarding Life), and John Williamson’s songs.

Representations of Indigenous Australians are not absent from the museum, but they appear from a narrow non-Indigenous perspective as emblematic stockmen (Matt Savage, Aboriginal Boss Drover), or on the fringes of the mid-nineteenth century settlement (Nellie from the Black Camp), and there are some central desert paintings (Bush Banana Dreaming by Eunice Napangardi, for instance). Aboriginal artefacts are evident alongside exhibits of explorers and convicts. The message in most of these representations is that the heritage to be celebrated by the tourists has no place for Indigenous Australians, except as either part of an unchanging and primitive pre-European settlement or as a minor part of a harsh but cooperative pioneer settlement. Surprisingly, one volume is available in the Gift Shop that opens up a very different narrative. Pamela Lukin Watson’s (1998) volume tells stories of what Henry Reynolds (1990) has called the other side of the frontier, looking at Aboriginal dispossession and decimation in the area around Longreach in the nineteenth and twentieth centuries. Presumably the word legend in the book’s title was sufficiently distracting and the content was therefore not registered.

One map in the Hall of Fame tells the story of the northward and westward movement of the frontier, with its self-conscious gesture to the equivalent American version of the assumed taming of a continent. This map has a near mirror image in a map exhibit at the Rockhampton Dreamtime Museum. The Rockhampton map, at what is claimed to be Australia’s largest Aboriginal cultural centre, is located far to the east of Longreach. It shows, from an entirely oppositional perspective, the northward and westward movement of invasion and dispossession that opened up the great Queensland pastoral stations.

Political Implications

The 2001 election may well be decided in regional and rural Australia. The future of the National Party in fending off challenges from One Nation and rural independents will be of great interest to political analysts. The Nationals are now attempting to run local election campaigns to distance themselves from their Coalition partners and to fend off the challenges of independents such as renegade Bob Katter Junior, in the seat of Kennedy (Sydney Morning Herald, 14-15.7.01, 3). Party leader John Anderson, however, still seems to consider that the challenges to traditional party loyalty in rural Australia can be explained by the breakdown in family values (Sunday, Channel 9 TV, 22.7.01).

The problems of regional and rural Australia are well documented. The debate about alleged rural and regional disadvantage has received enormous publicity in the period leading up to the 2001 Federal Election. The 1999 Regional Australia Summit advocated 247 "strategies" and a 2000 parliamentary committee made 91 recommendations (National Farmers Federation, 2001) to overcome distortions arising from government regulation or policy. These include: progressive income tax, given the volatile income patterns of many rural areas; capital gains tax exemptions, on residential property for example; service level disparities; and many externalities in areas such as the environment. While, arguably, these may cause such distortions, so do grants commission payments, microeconomic reform and competition policy. The funding of superannuation may also be seen to have caused an exodus of savings from rural Australia.

The outcomes of the growing divide are also clear. Australians continue to flock to the suburbs of the largest cities and high growth coastal zones such as the Sunshine Coast and the Gold Coast, while inland Australia grows more slowly or depopulates (Australian, 4.7.01. p.6). In some areas, Broken Hill for example, the rate of depopulation is dramatic. One should not confuse depopulation with disadvantage. In many outer suburban areas of Brisbane and Sydney for example, there is rapid population increase accompanied by significant patterns of disadvantage. However, in many rural and regional areas disadvantage and depopulation are joined together in a vicious circle of decline. In Australia the rural economy has changed so much in the past twenty-five years that it is now almost unrecognisable. Competition policy and privatisation of state services, alongside the decline of commercial banking and other services, have transformed the rural economy. ^{Most notably, the economic and trading base of Australia has changed greatly over the past century and this has impacted on the structure of agriculture. Since the early 1970s, the policy framework for dealing with regional Australia has} changed greatly. Competition policy, for example, may have lead to some efficiencies, and it may even have reduced some prices, but many have been excluded from its benefits, especially in rural areas where it does not necessarily improve the quality of life for all that are affected by it.

There is no easy way to fix this problem. The short term remedies considered by competing party policies in 2001 are unlikely to have any sustained effect, and other remedies, such as new tax zones, will be of advantage to the rich in deprived areas, but will not necessarily halt the disadvantage to entire regions. Enterprise zones or funded regional economic initiatives have the bad publicity of ‘picking losers’, but may repay more serious scrutiny using models developed in the European Union and The United States. In France, for example, the value of agricultural production is justified partly in terms of heritage protection of traditional ways of life. Regional policy is thus directly linked with heritage. In Australia the link is not so clear, and the Commonwealth Government Regional Policy, if it exists at all, is much harder to understand. Arguably Australian regional policy consists only of a series of programs that may be applied for, and a series of expenditure items in areas such as health, education roads and railways that are not coordinated in any clear way. Indeed, there are no agreed definitions of regions (apart from Tasmania and the Northern Territory) and there are certainly no clearly established links between policy outcomes and defined Australian regions.

In this wider debate it is useful to distinguish between the politics of the pork barrel, and the serious systemic debate about redressing rural disadvantage, and promoting regional growth. Ironically, it is the loss of the former public regulatory role of government in areas such as finance, tariffs, and public sector management that has provoked ad hoc use of short-term financial incentives by government in many areas such as the Centenary of Federation Fund and the National Heritage Trust. Arguably, these have little long-term or widespread contribution to make to the disadvantage of rural and regional Australia. Indeed, this is an example of the wider crucial distinction between the planned, sustained, long-term commitment to recurrent expenditure in public services and the short-term one-off expenditure that undermines equitable public service provision. This distinction allows us to consider the failure of elected politicians to deal with the long-term policy debate, and it further presents us with the means to identify reasons for widespread, unprecedented, electoral volatility that arises from disaffection with politicians, and with the activity of politics.

While perceptions of a democratic deficit are responsible for a problem of legitimacy in many contemporary liberal-democratic states, in Australia stereotypical perceptions of the rural may further alienate some voters from all the major parties and from the National Party in particular. How is the rural represented in current Australian political discourse? Let me briefly discuss some examples. Since 1987 the Australian Government’s Rural Book has purported to provide a guide to Commonwealth Government services and programs for people who live away from the capital cities. Most of the information summarises Federal government services for all Australians in health, education, transport, support for migrants. Yet when the term rural is addressed, as in rural industries, it is placed in the context of regional and remote, and overwhelmingly it privileges the agricultural sector.

Even the Commonwealth Government’s Programme for Services to Rural and Regional Communities tends to ignore any town with a population of more than 3000. It provides some Centrelink access points to small and remote places where all service provision is a major problem. Similarly, the Commonwealth Government’s Rural Plan concentrates on the development of rural industries, narrowly defined. This view of the rural even more clearly marked in the use of the term rural in publications such as the Rural Vision Magazine, put out by the Commonwealth Department of Agriculture, Fisheries and Forestry (2001).

This emphasis on the rural that concentrates on its remote and agricultural aspects is found, for instance, in the Ministerial Budget Statement (2001-2002) by National Party leader and Minister of Transport and Regional Services and National Party leader, John Anderson (2001), and in the associated ‘Budget Papers. Highlights for 2001-02’. These papers relate to a range of departmental initiatives. It includes items that are to be provided for all Australians, but some initiatives are addressed to the rural population specifically. My edited list of these projects includes the following:

Agriculture - Advancing Australia Package; the Regional Health Strategy; the Networking the Nation Programme; the Regional Equalisation Plan; the Stronger Families and Communities Strategy; the Regional Solutions and Rural Transaction Centre Programmes; the region-specific structural adjustment measures such as the dairy package, the Centenary of Federation Fund; the Working Credit initiative for rural workers in seasonal jobs such as harvest work; the Agricultural Development Partnerships Programme for structural adjustment targeted to specific agricultural industries and regions; the industry diversification and employment generation for the South West Forests of Western Australia Structural Adjustment Package; the Rural Financial Counselling Service "to assist businesses in rural Australia - particularly farm businesses - to deal with a range of financial pressures and adjustment issues; the new Federal Flood Mitigation Programme; mobile phone coverage, internet services, payphone, telecommunications projects in the education and health services; the improvement in the quality of health workforce in rural and regional Australia; the extension of the Natural Heritage Trust to address environmental challenges; the National Action Plan for Salinity and Water Quality; the joint Federal, New South Wales and Victorian Government effort to improve the Murray River; a major increase in road quality; support for the Alice Springs to Darwin railway; measures for quarantine strengthening against farm diseases; changes to the treatment of private trusts and private companies to assist 'succession planning' for farms; extension to the Rural Financial Counselling Service; and extra undergraduate student places for regional higher education institutions.

Despite some important exceptions, the great part of this list gives a rather limited view of rural and regional Australia. It confirms some of the rural stereotyping discussed in this paper.

Conclusion

Museums maintain iconic views of identity as heritage. War memorials, for example, reveal militaristic senses of national identity based on sacrifice, as well as many variations based on civic identity (Inglis, 1998). Donald Horne (1984, 1997) has used the concept of a fictional museum to denote all the intellectual currents in contemporary Australia. Museums are the keepers of our heritage values. One positive stereotype of rural Australian heritage is expressed most clearly in Longreach. A trail of self-funded retirees make pilgrimages to the shrine of this version of rural Australian authentic identity. Its influence is intense. A heritage, we should recall, is a folk version of a tradition. As Hobsbawm and Ranger (1983) have famously observed, it is often invented to serve the political needs of current political activists. It is "a retrospective analytical construction which produces a rationalized version of the past" (Gunnell, 1978).

The use of the term ‘rural’ in most current political debates fails to make clear that most rural residents do not live in remote or isolated settings, maintaining pioneer heritage ways of life, occupied in mining and agricultural or pastoral production. Governments do not always appear to recognise that the populations of large country towns have very different problems of comparative disadvantage with the metropolitan centres from the inhabitants of the smaller owns. These in turn are very different from the problems affecting people in very remote areas. The various stereotypes hide such distinctions. From the perspective of the city, all the places of inland Australia (apart from wineries, snow resorts and national parks) are relatively unlikely places for city dwellers to spend much time. They are beyond the pale. There is therefore some reliance on the stereotypical views of the rural.

Australian depictions of the rural has a complex history where myths and realities seldom coincide (Ward, 1965). It should be stressed that an acceptance of the rural as depicted at Longreach (as well as in the other places discussed in this paper) is often internalised by rural residents, including those in large country towns with occupations unconnected with agriculture. It is also clear that many residents of small towns are positively fearful of city life, even if they simultaneously draw succour and support from stories of settler and pioneer survivalism. The mystique of the bush legend is very powerful, and it is no surprise that it is swallowed whole by many rural residents.

Stereotypical views of the rural being presented at Longreach and on radio shows such as Australia All Over, are watered down, yet still evident, in some of the policy outcomes of governments. This is not to claim that there are not be other reasons for the agricultural emphasis in the use of the term rural in public discourse. One important factor here is likely to be the influence of the agricultural lobby, particularly the Nation Farmers’ Federation, in capturing the attention of the National Party and other powerful parties and decision-makers in the capital cities. The cultural context in which such institutional politics is embedded should not be neglected, however. This cultural context includes popular discourses such as those that draw on the stereotypical views of the rural discussed in this paper.

These cultural influences may be subtle. Of course it is not the case that Canberra-based politicians expect the typical voter of Eden-Monaro, for example, to look like the Man from Snowy River. It is true, however, that this sort of stereotypical image can be a useful political tool for an ambitious politician such as former National and now independent candidate, Peter Cochrane (Sydney Morning Herald, 7-8.7.01, 34). It is also true that such stereotypes, in both their positive and negative forms, draw attention away from demographic, cultural and economic factors that more accurately represent rural inland Australia. This is how all such stereotypes work, whether they be based on gender race, class, nation or region. Governments continue to fail to fully understand the ‘rural’ in rural Australia. If we are to move to a more inclusive discussion about regional policy in Australia, this should be addressed.

Anderson, J., and I. MacDonald, 2001, ‘Regional Australia, Partners in Growth’, Ministerial Statement, Department of Transport and Regional Services, Canberra, 22.5.01, http://www.dotrs.gov.au/budget/regional/index.htm, downloaded 25.7.01, plus extra information in Budget initiatives for Regional Australia, ‘Budget Papers. Highlights for 2001-02’, Department of Transport and Regional Services, Canberra,

http://www.dotrs.gov.au/budget/regional/2001_2002/highlights.htm downloaded 25.7.01.

Astley, T., 1999, Drylands, Ringwood: Penguin.

Department of Agriculture, Fisheries and Forestry, 2001, ‘Rural Vision’, Canberra. http://www.affa.gov.au/docs/innovation/ruralvision/index.html downloaded 24.7.01.

Eco, U., 1987, Travels in Hyperreality, London: Picador.

Gunnell J., 1978, ‘The Myth of Tradition’, American Political Science Review, 72, 1, pp. 122-134.

Hobsbawm, E. and T. Ranger, 1983, The Invention of Tradition, Cambridge: Cambridge University Press.

Horne, D., 1984, The Great Museum, London: Pluto.

Horne, D., 1997, The Avenue of the Fair Go, Pymble: Harper Collins.

Inglis, K.R., 1998, Sacred Places: War Memorials and the Australian Landscape, Carlton: Miegunyah Press.

Kane, J., 1997, 'Radicalism and Democracy' in G. Stokes (ed.), The Politics of Identity in Australia, Cambridge, New York, and Melbourne: Cambridge University Press.

National Farmers Federation, 2001, ‘Taxation Zones and the Country-City Divide’, A Discussion Paper, May, http://www.nff.org.au/ downloaded 4.7.01.

Reynolds, H, 1990, The Other Side of the Frontier: Aboriginal Resistance to the European Invasion of Australia, Ringwood: Penguin.

Thompson E., 1994, Fair Enough: Egalitarianism in Australia, Sydney: UNSW Press.

Walker J., 1999, A Rewarding Life, Longreach: Santa Publishing.

Ward, R., 1965, The Australian Legend, 2^nd ed, Melbourne: Oxford University Press.

Watson, P.L., 1998, Frontier Lands and Pioneer Legends; How Pastoralists Gained Karuwali Land, Sydney: Allen and Unwin.

Clive Bean

Queensland University of Technology

Patterns of Australian Mass Opinion on Issues of National Sovereignty

Abstract

This paper considers data on two dimensions related to the concept of national sovereignty, namely public attitudes concerning the maintenance of economic boundaries and attitudes related to the social, economic and cultural impact of immigration. The data suggest that in general Australians are inclined to favour economic protection rather than openness but at the same time they tend to emphasise the benefits rather than the costs of immigration. The paper tests a number of theoretical propositions concerning the social and attitudinal predictors of economic and immigration-related sovereignty attitudes. Despite some differences between the two dimensions, the most important variables prove to be occupational status, birthplace and education, plus attitudes towards multi-national companies and general feelings towards immigrants. These results imply that, other things being equal, public opinion on both the immigration and economic dimensions are likely to become more open over time.

Like many economies around the world, during the last two decades Australia has attempted to open up its local markets to external competition, in line with the increasing worldwide emphasis on globalisation. The Australian government has also pursued policies over a much longer period of time aimed at strengthening its economic and cultural foundations by broadening the ethnic mix of its population through immigration. Citizen attitudes to these policies and their consequences are likely to be interactive with the direction of the policies themselves. To some extent politicians will be mindful of perceived public attitudes when adopting such policies, which in turn are likely to help shape public views, which in turn again may have an influence on the path that future policy-making follows. At very least it is important for decision-making elites to have some understanding of ‘the mood of the people’ on relatively sensitive public issues such as these, which have at times proven politically explosive in Australia. This paper explores public attitudes on these issues and tests various theoretical accounts of their origins.

A common thread that these two policy areas have is that in some sense they both have implications for national sovereignty, or national independence (Warhurst 1993a). The notion of sovereignty in this context refers to the ability of nation-states to control their destinies free of external influences and relates also to the more restricted concept of boundary maintenance. ‘External sovereignty’, as it is sometimes termed, ‘is generally considered to be synonymous with political independence and brings with it the right of exclusive jurisdiction over a territory and its population ...’ (Alston 1995, p. 17). Boundary maintenance can be considered to be an aspect of sovereignty and may relate to more formal or more symbolic boundaries (Phillips 1996).

Ideas connected with boundary maintenance range across a variety of issues, including in particular economic, defence, cultural and ethnic identity questions. This analysis focuses on two dimensions of this concept, namely attitudes related to the maintenance of economic borders and attitudes related to the impact of immigration and immigrants on Australia’s national sovereignty, especially the social, economic and cultural costs and benefits of immigration. The interest here is not only in the structure and distribution of these views across the community but also in how they are shaped by social background and by social and political attitudes. An additional point of interest is that the data were collected during the period preceding the advent of the right wing populist politician, Pauline Hanson and her One Nation Party, whose political agenda has issues of immigration and economic protection at the forefront (Grant 1997; Abbott et al. 1998; Leach et al. 2000). On these key dimensions, the data in this paper thus provide something of a baseline for those interested in judging the impact of One Nation.

The distribution and structure of sovereignty attitudes

Let us look first at a number of questions which address the issues of economic and immigration-related sovereignty. The data for this analysis come from the 1995-96 Australian National Social Science Survey (Kelley et al. 1998). The data file consists of 2438 cases and the questions on attitudes towards national sovereignty come from the International Social Survey Programme National Identity module, which was included in the survey. Table 1 shows responses to three questions relating to economic sovereignty and four questions concerning immigration-related sovereignty, all based on five-point Likert type scales.

Table 1 here

Looking initially at the questions on economic borders, we see that the distribution of opinion among the Australian public is consistently in favour of limiting access to external elements and protecting the national economy. For example, over three-quarters of the sample agree with the proposition that Australia should limit the import of foreign products in order to protect the national economy and only one respondent in ten disagrees. These findings are consistent with similarly strong endorsements of economic protectionism found in other studies (for example, Bean 1995). On the question of whether foreigners should be allowed to buy land in Australia, the balance is more even but more people remain against allowing foreign ownership (43 per cent) than in favour of it (36 per cent). The third question asks whether preferential treatment should be given to Australian films and programmes on television and a clear majority say yes (58 per cent), while only 21 per cent say no.

From the questions related to the costs and benefits of immigration, a rather different picture emerges. Whereas there is a strong emphasis on preserving Australia’s national sovereignty with respect to economic competition, respondents do not display the same concern in regard to immigration. On each item the balance of opinion emphasises the positive with respect to the impact of immigration. Thus, 43 per cent of respondents disagree that immigrants increase the crime rate, while 32 per cent agree. Sixty-three per cent agree that immigrants are generally good for Australia’s economy compared with only 15 per cent who disagree. A narrower difference results from the proposition that immigrants take jobs away from people who were born in Australia, but the balance of opinion is still a favourable outcome for immigrants: 42 per cent disagree that they take jobs away while 36 per cent agree. Most strikingly of all, fully 87 per cent agree that immigrants make Australia more open to new ideas and cultures and only 5 per cent demur.

It is worth noting, however, that these responses suggesting that the public regards the beneficial effects that immigrants bring to Australia as outweighing the negative effects, do not coincide with an equally positive view of immigration itself. Views on these issues are clearly very complex, as revealed by the fact that, in answer to a further question, over 60 per cent of the same sample believe that immigration into Australia should be reduced and only 11 per cent believe that it should be increased, a balance of opinion on the direct question about levels of immigration that is consistent with public views as recorded in opinion polls for most of the period since the Second World War (Betts 1988; McAllister 1993).

It is not only the pattern of responses that are of interest but the structure of the data on questions of sovereignty as well. Table 2 contains results from a principal components analysis of the seven items from Table 1. This analysis gives empirical foundation to the assumption that there are indeed two separate dimensions of attitudes towards sovereignty, with the four immigration-related items loading on the first dimension and the three economic items loading on the second. For the analysis to follow, the three economic items are thus combined into a simple additive scale of attitudes towards economic sovereignty and the four immigration items are likewise combined into a single scale (with the two items that load negatively on the dimension suitably reversed). These two scales then form the dependent variables for the analysis. A certain amount of caution is, however, required in the interpretation of some of the individual scale items, since past analyses of similar data have revealed problems in their utilisation (Goot 1993).

Table 2 here

Theoretical expectations

Theoretical considerations would lead us to predict that a number of social structural variables might influence the attitudes a person holds towards economic and immigration-related sovereignty. In particular, education has been identified as a potentially crucial factor in shaping views along these dimensions. Those who have experienced higher education have been exposed to a wider set of perspectives likely to facilitate a more cosmopolitan, less parochial view of the world. The highly educated may also be encouraged to adopt these more open views because they regard a cosmopolitan world-view as being something of a status symbol (Betts 1988; 1993; Birrell 1995; Bean 1995). To the extent that cosmopolitanism relates more to issues of culture and population than to those of economic borders, it is also possible that education would in turn be more influential in shaping immigration-related sovereignty attitudes than sovereignty attitudes based on economic considerations.

The reverse is likely to be true for occupational status, another key social structural factor likely to shape views on sovereignty. Persons of higher occupational status arguably have an interest in promoting more open economic boundaries in particular, because they increasingly work in an international market place in which their services and products can be traded worldwide and in which national borders are thus becoming much less relevant (Reich 1991; Bean 1995). There is less reason to assume, however, that occupational position per se would be directly related to immigration sovereignty.

We would also expect immigrant status itself to have a substantial impact on the formation of attitudes on questions of economic and immigration-related sovereignty. People born overseas could be expected to be more likely than the Australian-born to emphasise the benefits of ethnic and cultural diversity (Graetz and McAllister 1994; Markus 1993), since they embody that diversity within the Australian population themselves. The overseas-born would presumably be especially inclined to highlight the benefits associated with immigration, but they could also be expected to see virtues in more open economic borders, partly because they might see benefits from such policies for the economies of their original countries and indeed for themselves.

Membership of a trade union is another factor which may influence a person’s position on issues of economic and immigration-related sovereignty, though not necessarily in the same direction on the two issues. On the issue of economic sovereignty, the union movement has been an opponent of opening up Australia’s borders on the grounds that its members’ jobs will be threatened if local industries do not receive sufficient protection, while on the question of the benefits of immigration the union movement has tended to be somewhat equivocal, there being a tension between the potential threat to the jobs of the Australian-born from competition by immigrants and the potential benefits of economic growth that immigration may bring (Warhurst 1993b). Officially, however, the union movement has generally supported policies which have encouraged immigration. We would thus expect members of trade unions to oppose economic openness but if anything to be more inclined to see the benefits than the costs of immigration.

Several other aspects of social structure should be included in any analysis of sovereignty issues, partly because they are related to those we have just discussed (and their exclusion would therefore create the potential for incorrect conclusions to be drawn from the analysis) and in some cases because there are grounds to believe that they may have an impact of their own on the attitudinal dimensions in question (Evans 1995). These include income, residence in an urban or rural location, age and sex. Among these, age and place of residence may have particular relevance. Younger adults may favour an open stance to a greater extent than older people because they have been socialised into a world in which the dominant discourse among political elites has emphasised the virtues of openness. With respect to place of residence, people who reside in rural areas may be more suspicious than their urban counterparts of both immigrants and open economic borders because they have less experience of the former and consider themselves to be especially vulnerable to the latter.

In addition to having origins in social structure, views on specific aspects of sovereignty may also come about partly through the possession of broader political and social attitudes. Central dimensions of political ideology in Australia include attitudes towards the free market and attitudes towards trade unions (Kelley 1988; Kelley, Bean and Headey 1990). The former in itself has a number of separate strands and in the context of boundary issues the most salient of these is attitudes towards the operation of multi-national corporations, which are a powerful symbol of the global free market. We could expect, for example, that individuals who have a positive view of multi-nationals would be inclined to favour the opening up of Australia’s economic borders, in particular, while, based on the above discussion on the stances of the union movement, those holding positive attitudes towards trade unions should be more likely to be against opening up economic borders while at the same time they could be expected to be more likely to see benefits rather than costs in immigration.

As well as these economic dimensions of political ideology, general feelings towards immigrants from different countries are also likely to play a major role in shaping attitudes towards national sovereignty. We could expect broad feelings towards immigrants to exert a powerful influence upon specific evaluations of the costs and benefits of immigration and to a lesser degree on the issue of economic boundaries, with those who generally felt positive towards immigrants more likely to see the benefits of openness.

Finally, party politics may play a significant role. Despite the fact that the major political parties have tended to avoid taking issue with each other over immigration policy (McAllister 1993), they do have clearly differentiated stances (Newman 1995), with the Labor Party being supportive of higher levels of immigration and of multiculturalism compared with the Liberal and National parties which favour lower levels and a greater emphasis on cultural assimilation. Labor supporters may thus be more inclined than coalition supporters to perceive benefits from the presence of immigrants. On the other hand, despite the policies of the recent Labor government aimed at opening up Australian markets to global competition, Labor has traditionally been an advocate of economic protection and its followers among the public may still be more likely than coalition followers to support this view.

Results

Having outlined theoretical expectations as to the likely influences on attitudes towards our two dimensions of sovereignty, it is now time to put these expectations to the test. This is done by way of ordinary least squares multiple regression analyses which show the impact of each variable in the model net of the effects of the others (details of the variables in the analysis are in the Appendix). Table 3 contains an analysis of attitudes towards economic sovereignty and Table 4 contains a parallel analysis of attitudes towards immigration-related sovereignty. In each case, the first column of figures presents the simple Pearson product moment correlation between each independent variable and the dependent variable and then the table introduces a number of different models, starting with the social structural variables on their own and successively adding attitudinal variables to each subsequent model.

Table 3 here

Concentrating initially on the first column in Table 3, we see that all but two variables have statistically significant zero-order correlations, in the predicted direction, with attitudes towards economic openness. Most of the correlations are modest in size, however, although attitudes towards multi-national companies correlate a good deal more strongly than any of the other variables. When we turn to examine the standardised regression coefficients (betas) for the social structural variables in model 1 only some of our expectations are confirmed. Higher occupational status and being born overseas lead to a more open economic stance, as predicted, and so does having a higher income. Age also has the expected relationship, with older people being more protectionist than the young.

The most surprising result is that, once other aspects of social structure are controlled, education has no independent impact on attitudes towards economic sovereignty. Our prediction was that education may be less influential in shaping economic than immigration-related sovereignty attitudes, but not that it would be entirely uninfluential. Although education is significantly correlated with economic openness, the analysis suggests that this association occurs not because of characteristics to do with education itself but rather because education is related to variables such as occupation, income and age, which do significantly influence economic sovereignty attitudes. Trade union membership, place of residence and gender also fail to exert a significant influence and the overall amount of variance explained by social structure is small (5 per cent).

Model 2 adds attitudes towards multi-nationals and trade unions to the equation. As anticipated, views about multi-national corporations have a strong impact. Positive evaluations of multi-nationals lead to more open attitudes towards economic boundaries (beta = .28). Attitudes towards trade unions, however, have no net impact on economic sovereignty attitudes in this equation. In addition, income is no longer significant and the coefficients for both age and occupational status have declined somewhat from model 1, indicating that to some extent these factors work through attitudes towards multi-nationals in shaping views on economic sovereignty. Birthplace, however, retains its influence at the same level, suggesting that immigrant status as such, and not its connection with other attitudes, predisposes those born overseas to support more open economic borders.

In model 3, feelings towards European immigrants on the one hand and Eastern/Asian immigrants on the other, are included. Feelings towards the latter group have a substantial impact (beta = .21), with those who rate Eastern/Asian migrants more positively more likely to favour open boundaries, as predicted. The effect for ratings of European migrants, by contrast, is rather surprising. While the coefficient is the same size as the zero-order correlation, the sign has changed from positive to negative. In other words, once attitudes towards Eastern/Asian immigrants have been taken into account (plus the other variables in the equation), it appears that positive ratings of European immigrants influence people to be slightly less supportive of open economic boundaries. It is tempting to speculate that in contrast to migrants from Eastern and Asian countries, European immigrants, perhaps by virtue of their generally longer time in Australia and greater cultural similarities, may in some sense not be viewed as ‘foreign’ and thus attitudes towards them coincide with protectionist sentiments to some degree. Given that Australia’s moves towards economic openness have occurred alongside efforts to enhance relations with Asian countries and the growth of Asian immigration into Australia, it may also be that economic openness is seen as being connected with openness towards Asia.

Attitudes towards multi-nationals remain the pre-eminent influence in model 3, while attitudes towards trade unions are now marginally significant in the expected direction: favourable attitudes towards unions are associated with a more protectionist stance. The presence of the immigrant rating variables has little impact on the effects of the social structural factors. The final model, which includes political party identification, has virtually no additional impact. Presumably reflecting the current bipartisan stance of the major parties on the issue of economic sovereignty, party identification itself has no effect. Even this full model does not explain a great deal of the variance (only 15 per cent) in attitudes towards economic sovereignty.

Table 4 repeats this analysis with attitudes towards immigration-related sovereignty issues as the dependent variable. The size of the raw correlations in the first column signal that this dimension of sovereignty attitudes is more closely related to the variables in our model. Only union membership, age and sex do not register significant correlations. Model 1, concentrating again on the social structural predictors indicates that this time our prior expectations about education are verified. Education in fact has a very substantial impact on attitudes to immigration-related sovereignty, with higher levels of education leading to greater emphasis on the benefits of immigration, as anticipated. While this finding provides general endorsement for our theoretical explanations for the expected impact of education - the cosmopolitan and status symbol arguments - it does not of course help distinguish between the two.

Table 4 here

Birthplace again has a considerable effect, as predicted, not much smaller than that of education and a larger one than it had on the economic dimension. Occupational status has a significant, but much reduced effect compared to its raw correlation, while income loses its statistical significance altogether. On this dimension the tables thus appear to be turned, with education dominating occupation and income as the key factor among the indicators of socio-economic status. Contrary to expectations, but perhaps reflecting the equivocal nature of the union movement’s positions on these issues over the years, union membership has no impact on either dimension of sovereignty attitudes. As expected, urban residents are somewhat more positive about the consequences of immigration than rural folk. Age also has a significant impact net of other elements of social structure, but it is older people, not younger, who are more ready to emphasise the benefits rather than the costs of immigration.

Although their introduction makes little difference to the pattern of social structural influences, attitudes towards both multi-nationals and trade unions have significant, though relatively modest, effects (model 2). In both cases, despite their being negatively correlated with each other, favourable views of multi-nationals on the one hand and of trade unions on the other lead to a more open stance on the question of immigration-related sovereignty.

The big impact, however, comes from the general feelings towards immigrants introduced in model 3. Ratings of both European immigrants and Eastern/Asian immigrants show large zero-order correlations, but only the latter retains its impact in the multivariate analysis. The impact of Eastern/Asian migrant ratings is very large indeed (beta = .51), while the European dimension loses its significance altogether, perhaps for reasons along the lines of those canvassed above, that (net of feelings towards Eastern/Asian migrants) Europeans are not so much who people have in mind when they think of immigrants. The dominant influence of ratings of Eastern/Asian immigrants is such that the effects of both multi-national and trade union attitudes (the more so the latter) are reduced somewhat in model 3.

The introduction of the immigrant rating variables also coincides with the disappearance of the occupation effect and, more interestingly still, with a marked reduction in the impact of education (even though it remains clearly influential in its own right). In other words, to some degree education influences immigration-related sovereignty issues through feelings towards migrants. To the extent that ratings of migrants from a variety of different countries can be viewed as an indicator of cosmopolitan attitudes, this evidence can be taken as partial endorsement of the theory that it is indeed a cosmopolitan perspective derived from higher education that leads to well-educated people being more likely to emphasise the benefits of immigration rather than its costs.

Model 4 introduces party identification into the equation and although this does not add to the already substantial R-squared of 38 per cent, party support is significant in its own right, with Liberal-National supporters less likely than Labor supporters to emphasise the benefits of immigration. Furthermore, attitudes towards trade unions are now no longer significant. It would appear that the impact of this variable works partly through feelings towards immigrants and partly through its association with party identification. Finally, a comparison of the variance explained at the bottom of Tables 3 and 4 indicates that the model accounts much better for immigrant-related attitudes towards sovereignty than it does for the economic sovereignty dimension.

Summary and Discussion

Many but not all of the theoretical expectations outlined earlier in this paper have been confirmed by the analysis. The most notable surprise was the lack of an independent effect for education on attitudes towards economic sovereignty, although for a variety of reasons the rejection of this relationship must remain a qualified one. Furthermore, education was shown to exert a powerful influence on the other dimension of sovereignty, relating to the effects of immigration. This second analysis also provided evidence for the argument that the quality that education imparts to its recipients that leads to their adopting a more open stance is a cosmopolitan view of the world. Occupational status also had some influence on both dimensions and the most consistent social structural variable was immigrant status as measured by birthplace.

In thinking about the implications the results in this paper may have for the future shape of Australia’s economic and population policy, it is necessary to be mindful of the fact that over time the levels of formal education, the occupational status of the workforce and the proportion of overseas born in the population – the three most important social structural variables in the analysis – are all increasing. This is also true for many other societies and the conclusions in this paper are thus bound to have significance beyond Australia. Each of these structural factors is positively associated with greater openness on sovereignty issues. Other things being equal, we are therefore likely to see a gradual trend towards greater public acceptance of open borders of both the economic and immigration varieties. Based on the data in this paper, national sovereignty appears to be more of a concern to the public in its economic manifestation than in its ethnic or cultural manifestation. As we move into and through the 21st century, it may gradually become less of a concern on either front.

This is not to say, however, that it will be an easy political issue to manage from a policy-maker’s perspective. In general, the public may indeed be likely to move progressively in the direction of a more open stance. Yet, the rise of the One Nation Party in Australia plus similar parties in other countries, together with the increasing incidence of anti-globalisation protests around the world, suggest that there are significant numbers of citizens for whom moves to reduce immigration and economic boundaries remain anathema. This direct linking of sovereignty to the globalisation debate makes it an issue of considerable delicacy for policy-makers in democratically elected governments who are doubtless aware of the adverse political consequences that lack of sensitivity to the apprehensions of such people could bring.

Appendix

Appendix Table 1 provides the basis for the multiple item scales which comprise the attitudinal independent variables in the model, showing items that make up each scale and the principal components analysis demonstrating the distinctiveness of each dimension. The table shows there to be two separate dimensions of feelings towards immigrants, one of which can be labelled Eastern/Asian immigrants and the other of which relates to European immigrants. The first is based on feeling thermometer ratings of Vietnamese, Lebanese, Chinese and Indian migrants and the other is based on thermometer ratings of Italian, British and Greek migrants. In addition the three items addressing attitudes towards trade unions load strongly together, as do the three multi-national items. As was done for the dependent variables, the relevant items from each dimension are combined into additive scales (with items that load negatively reversed), representing each of the four dimensions. Cronbach’s alphas for these scales are: .93 for feelings towards Asian immigrants; .87 for European immigrants; .79 for attitudes towards trade unions; and .73 for attitudes towards multi-nationals. The two immigrant scales are thus highly reliable, despite some cross-loadings evident in the principal components analysis. Appendix Table 2 contains summary information on all variables in the model, showing how they are scored and their means and standard deviations.

Appendix Tables 1 and 2 here

Endnotes

References

Table 1. Attitudes towards Economic and Immigration-Related Sovereignty Issues (percentages)

Source: Australian National Social Science Survey, 1995-96 (n=2438)

Table 2. Principal Components Analysis of Items Measuring Immigration and Economic Sovereignty Attitudes (varimax rotation)

Source: Australian National Social Science Survey, 1995-96 (n=2438)

Table 3. Multiple Regression Analyses of Variables Predicting Attitudes towards Economic Sovereignty Issues (standardised regression coefficients)

* p < .05; ** p < .01

Source: Australian National Social Science Survey, 1995-96 (n=2438)

Table 4. Multiple Regression Analyses of Variables Predicting Attitudes towards Immigration-Related Sovereignty Issues (standardised regression coefficients)