As mentioned earlier in the chapter, there exists a widespread notion in both the popular media and government and business reports that economies of scale exist in the provision of goods and services within the level of local government, such that economic benefits can be derived from the elimination of numerous small local government authorities through a process of amalgamation (forced or voluntary). It is also widely assumed that attempts to redress the poor financial position of this level should await resolution of this structural problem (Business Council of Australia 2006; Department of Transport and Regional Services 2005; House of Representatives Standing Committee 2003; State Chamber of Commerce (NSW) 2005; Sydney Morning Herald 2006).

In addition, many of the more radical proposals for reform of the federal system involve the elimination of the current state and local levels of government and their replacement by a single regional level of government (e.g. Hall 1998; Soorley 2004). This in turn implies both the devolution of most, if not all, current functions allocated to the States to the new regional level, and the amalgamation of a large number of the current local government authorities into larger regional units. It is of considerable interest then to examine the empirical evidence available relating to the economics of local government.

Overseas empirical evidence suggests that while total costs do vary with population size, it is not the case that a decreasing cost curve prevails for all population sizes (Byrnes and Dollery 2002; Dollery and Fleming 2005). That is, while there is evidence that costs do decrease up to a certain population size, after this point these costs generally begin to rise again – with suggestions that for large metropolitan areas the best solution is for a body charged with ensuring co-ordination of policy and implementation across key areas (such as transportation systems) but that this body be complemented by a series of smaller local government units responsible for ongoing service delivery (Sancton 2005). In addition, the results of various studies point to the population range associated with the lowest per capita costs being different depending on the type of service being investigated. This in turn highlights the fact that much of the overseas research in this area is of limited relevance to the Australian context, since local governments in our federal system have responsibility for a different range of services than those in other countries – for example, police and schools in the United States and public housing in the United Kingdom.

Empirical evidence from Australia is extremely limited in terms of volume, and mixed in terms of its results. However the majority of studies conducted to date suggest an ‘optimum size’ from an economic viewpoint as in 30-80,000 population range (Byrnes and Dollery 2002; Marshall 1998; Soul and Dollery 2000; Soul 2000) – a size exceeding many of our current local government jurisdictions. At the same time, the results obtained from these empirical studies have been criticised (Allen 2003; Dollery and Crase 2004; Dollery, Crase et al 2005; Woodbury et al 2002) for ignoring the variable nature of local government services and the presence of differentiated economies of scale – a fact confirmed from a detailed reading of the reports of the various state grants commissions referred to above as well as by studies focussing on particular local government services (for example, Worthington and Dollery 2000; Woodbury et al 2002). Labour intensive customer-oriented services would generally be expected to generate lesser economies of scale than capital intensive services (such as water and sewerage) where benefits can be obtained from spreading fixed costs across a larger number of service points (Dollery and Fleming 2005).

Many of the empirical studies conducted based on the Australian experience have also been criticised for employing single- rather than multi-variate analysis with omitted variables (demographic/geographic characteristics) such that the true impact of population size has, in general, been overestimated (Byrnes and Dollery 2002). It is not surprising then that local government grants commissions introduce variables such as population density, remoteness, ethnicity, indigenous status, age structure, non-resident service provision, climate and terrain into their analysis of the appropriate size of their equalisation grants, rather than distributing them on a strict per capita basis (Department of Transport and Regional Services 2005). This in turn suggests that the optimal size of a local government jurisdiction may differ for rural and remote communities than those based around provincial centres and metropolitan areas (Soul and Dollery 2000), but investigation of this hypothesis in a rigorous manner has yet to be undertaken.

Given the mixed results and the methodological flaws discussed above, it is not surprising that evaluations conducted post-amalgamation have almost invariably reported that the economic benefits actually experienced have been considerably lower than those estimated by the proponents of the change prior to amalgamation (Allen 2003; Dollery, Crase et al 2005) – and that this experience has been replicated overseas as well (Dollery, Keogh et al 2005; Frontier Centre for Public Policy 2003; Rouse and Putterill 2005). This in turn has led to some commentators suggesting that alternative means exist for achieving economies of scale in key functions without amalgamation (eg. ROCs, strategic alliances for joint provision, tendering) (Dollery 2005a; Dollery and Crase 2004; Dollery, Crase et al 2005; Dollery and Johnson 2005). However, little if any evaluation has been conducted with respect to the cost effectiveness of these alternative arrangements.